Hi-Lite Electrical Led v Wolseley UK Ltd

These proceedings arise out of a fire which occurred in the early hours of the morning of 24 January 2006 in a hair salon and health spa operated by Obsession Hair and Day Spa Limited (“Obsession”) in the Mailbox in Birmingham. It is common ground in these proceedings that the fire originated in a submersible Jet W300 pump (“the Pump”) in a sump in the colour dispensary of the hairdressing salon.

There have already been proceedings between Obsession and the electrical contractor who installed the Pump, Hi-Lite Electrical Limited (“Hi-Lite”), in the Birmingham District Registry (“the Birmingham Proceedings”) leading to a judgment dated 11 December 2008 in which His Honour Judge McKenna, sitting as a Judge of the High Court, found Hi-Lite liable to Obsession in contract for losses which Obsession suffered as a result of the fire. By a further judgment dated 8 November 2010 he assessed those losses at £847,171, inclusive of interest and held that if Hi-Lite had been liable to Obsession in tort the quantum, would have been £2,578,123. On 7 February 2011 Obsession was granted permission to appeal against the dismissal of its claim in negligence and also as to the quantification of the claim.

Hi-Lite have commenced these proceedings against Wolseley (UK) Limited (“Wolseley”) who traded under the name “Pipe Center” and operated stores which sold pipes, valves, fittings and associated plumbing products. Wolseley sold the Pump to Hi-Lite. In turn Wolseley has brought third party proceedings against DAB Pumps SpA, formerly Leader Group Pumps SpA (“Leader”), which is an Italian company which manufactured Jet W300 pumps and supplied the Pump to Wolseley.

In these proceedings Hi-Lite seeks a declaration that Wolseley is liable for the fire and that Wolseley should indemnify Hi-Lite for any sums for which Hi-Lite is liable to Obsession. Wolseley denies liability and, in turn, seeks to pass on any claim to Leader.

A drainage sump and submersible pump were installed in Obsession’s salon in Birmingham in about 2001. Initially a Jung pump was installed to pump the effluent from a number of washbasins and a dishwasher into the main drains. A filter was located upstream of the sump in the pipework leading to the sump to prevent debris from entering the sump.

In July 2005 Mr Christopher Leadbrook, who worked for the company carrying out maintenance at the salon, replaced the Jung pump with a DAB Nova 180 Pump. Hi-Lite was subsequently engaged by Obsession to carry out maintenance work at the salon from about July 2005. At some stage, apparently when carrying out maintenance of the lighting installation, one of Hi-Lite’s employees, Mr Anthony Fletcher, replaced the fixed float arm which had become detached from the Nova 180 pump.

On about 22 November 2005 the Nova 180 pump failed and the sump flooded. Hi-Lite was called out and Mr Fletcher attended and found the pump and sump covered in a mixture of “gunk and hair” and he found rubber gloves, a towel and a knife in the sump. He emptied the sump and concluded that the Nova 180 pump rotor had seized and that pump could not be repaired.

The next morning, 23 November 2005, Mr Fletcher visited Wolseley’s Pipe Center branch at Digbeth with another Hi-Lite employee, Mr Wayne Brown and purchased a Jet W300 pump for some £99.68 plus VAT.

Mr Fletcher returned to the salon and, together with Mr Brown, he set about installing the Pump in the sump at the salon.

The Pump had a mains cable attached to it. There was also a switch which floats on the end of a flexible cable attached to the Pump. This switched the Pump on when the water rose to a particular high level and switched it off when the water has been pumped out to a lower level. The length of the flexible cable could be adjusted to set the depth of water at which pumping should commence and cease. Whilst it is agreed that the fire started in the cable connecting the float switch to the Pump, the cause of the fire in that length of cable is not agreed and forms the essential issue to be determined in this case. The fire took place on 24 January 2006, some 9 weeks after the installation of the Pump.

In the Birmingham Proceedings the central issue, as expressed by Judge McKenna at paragraph 7 of his December 2008 Judgment, was whether Obsession could prove, on the balance of probabilities, that the Pump was the cause of the fire rather than, for example, carelessly discarded smoker’s materials or deliberate ignition by an intruder. In those proceedings, Judge McKenna heard evidence from witnesses called by Obsession who have not been called in these proceedings, in addition to witnesses from Hi-Lite including Mr Fletcher and Mr Brown who were called in these proceedings.

Expert evidence was also called in the Birmingham Proceedings. Obsession relied on the evidence of Mr Stephen Braund, an associate with Hawkins and Associates Limited who had a degree in electrical and electronic engineering and was a member of the Institution of Electrical Engineers. Hi-Lite relied on the evidence of Dr Richard Lipczynski who also had a degree in electrical and electronic engineering and was a member of the Institution of Electrical Engineers. He was a partner in Geoffrey Hunt and Partners.

Mr Braund came to the conclusion that the fire was probably caused by an incendive electrical fault in the Pump. Specifically, he considered the fire most likely to have been caused by the failure of the cable that connected the float switch to the main body of the Pump and he relied on evidence of electrical arcing damage on the relevant wires, 130mm from the point from where the cable was connected to the limit switch. Mr Braund’s view, as set out in paragraph 70 of the December 2008 Judgment, was that “the continual moving up and down of the float in the tank would have led to the fracturing of the wires inside the flex which was likely to have caused electrical in-line arcing to occur inside the flex when the motor was required to operate which in turn was likely to have ignited the insulation of the flex or other nearby combustible/flammable materials.”

Dr Lipczynski disagreed with Mr Braund’s analysis. He considered that the position of the arcing on the wires of the cable float switch strongly suggested that both the live and neutral conductors had sustained arc damage very close to where the float switch’s cable had been clipped to the Pump. He considered that the point where arcing occurred was at the point where the cable would flex. In his opinion radiant heat or fire attack on that cable would be highly likely to destroy cable insulation and cause conductors to touch and arc at the point where the cable insulation was under mechanical stress, that is, at the point he referred to. As a result he believed that the evidence of arcing observed on the live and neutral conductors of the float switch cable had resulted from external fire attack on that cable. In other words, Dr Lipczynski considered that the damage to the cable was caused by an external fire rather than being the cause of the fire itself. Both experts were however agreed that the fire almost certainly started in the colour dispensary. That was where the sump with the Pump was located.

Both experts carried out tests to simulate the behaviour of the float switch rising as water came into the sump and falling as water was then pumped out. On the basis of Mr Braund’s test, in particular test 5, his view was that if the free cable length of the float had been shortened to 3cm, it could lead to fracturing within the relevant time frame between installation of the Pump and the fire. His view, as summarised at paragraph 87 of the December 2008 judgment, was that “failure of a live wire is likely to cause the pump to run intermittently at first, causing the cable insulation to burn and carbonise and the pump then to run continuously. Burning is likely to break through the outer sheath of the cable and the burning of the wires will continue as long as the electricity supply remains live causing nearby combustible materials, including the plastic parts of the pump and cable insulation to ignite as damage spreads along the float cable.”

Dr Lipczynski criticised Mr Braund’s tests as being unrealistic and not being representative of the conditions experienced for a number of reasons, including the fact that the electrical load was grossly in excess of the 300 watts of the Pump and, except for test 5, the cable was different from the cable fitted to the Pump.

Having heard his evidence, Judge McKenna concluded that Mr Fletcher must have shortened the float cable significantly below the 10cm stipulated in the instructions in an attempt to get the Pump to operate without the switch fouling the sides of the sump to such an extent that it would not work reliably. Judge McKenna said that shortening of the float cable to 3cm was consistent with the point of the fracture being no more that 30mm from the edge of the plastic moulded strain relief. It was also consistent with the experts’ original agreement that the point of fracture occurred where the float cable was clipped to the body of the Pump.

At paragraph 106 of the December 2008 Judgment Judge McKenna concluded that a probable cause of the fire was a fault in the Pump and at paragraph 114 said that on the balance of probabilities he concluded that the cause of the fire was the fracturing of the wire in the float cable of the Pump. In doing so he accepted the evidence of Mr Braund and rejected the evidence of Dr Lipczynski.

There was no application by Hi-Lite to join Wolseley into the Birmingham Proceedings. There was some correspondence which indicated that Wolseley was aware of the Birmingham Proceedings at a fairly late stage. The fact that Wolseley were not a party to the Birmingham Proceedings in which Hi-Lite were found to be liable means that there is a risk of inconsistent findings because the evidence that I have heard and, in particular, the expert evidence has been different to the evidence which was put before Judge McKenna.

These proceedings were commenced by a Claim Form issued in the Queen’s Bench Division on 17 March 2009. In June 2009 Hi-Lite sought a determination by the Court that it would be an abuse of process for Wolseley to dispute the findings of fact made by Judge McKenna in the Birmingham Proceedings. After a hearing before Mr Justice Burnett, he held that it was not an abuse of process. The matter was then transferred to the Technology and Construction Court. Directions were given, including directions as to expert evidence, leading to the trial.

By the time of the closing oral submissions the issues between the parties had substantially reduced. Hi-Lite’s case was that the Pump was not of satisfactory quality under section 14(2) of the Sale of Goods Act 1979 because a pump should not catch fire within 9 weeks of installation. Wolseley and Leader contended that the fire was a result of damage to the float switch cable which had been caused either in installation or in service. Hi-Lite responded by saying that Wolseley and Leader had not established that such damage had occurred and that it was more likely that there was damage to the float switch cable in the manufacturing process when the float switch and strain relief was moulded onto that cable.

There was then an issue as to the effect of Hi-Lite not fitting an RCD as part of the installation of the Pump. If the fire was caused by a defect in the float switch cable for which Wolseley and Leader were responsible, they contended that Hi-Lite should have fitted an RCD and that the fire developed in a manner which would have caused the RCD to trip and isolate the cable from the power supply before the cable ignited. Hi-Lite, on the other hand, contended that the fire developed in a manner which meant that the cable would have caught fire without the RCD tripping so that the RCD would not have prevented the fire. On this basis, I shall have to decide how the fire developed. If I am satisfied that the RCD would have tripped I then have to decide what effect this has as a matter of law in relation to causation, novus actus interveniens or apportionment of liability.

Wolseley and Leader also contend that, if the fire was caused by a defect in the float switch cable for which Wolseley and Leader were responsible, damages arising from a fire were too remote because damage by fire was not in the contemplation of the parties as the likely result of the cable being defective. Hi-Lite contends that property damage is sufficient and that damages for fire were recoverable.

It follows that no party now contends that the fire was caused by Hi-Lite shortening the cable so that the length of the cable from the clip on the body of the Pump to the boss on the float switch was less that 10cm. This however was the mechanism of failure which formed the basis of liability in the Birmingham Proceedings. The allegation that the cable had been shortened gave rise to a contention that the Pump was not of satisfactory quality where the design required a particular length of cable to prevent fire but no proper warning was given. This, in turn, gave rise to an issue as to the meaning of the terms of the instruction manual and, in turn, an issue between Wolseley and Leader as to responsibility for the wording of that manual. None of those issues now arise.

Hi-Lite also alleged that the Pump supplied by Wolseley was not fit for a purpose made known by Mr Fletcher to the assistant at the Pipe Center. This led to an investigation at trial of issues which included the number of sinks and other appliances which drained to the sump, a matter relevant to the number of times the Pump operated; what temperature the water was when it reached the sump; what chemicals were in the water and whether the Pump was suitable to handle hair. By the time of the oral closing submissions, Hi-Lite was no longer contending that any of these matters were relevant to the cause of the fire. As a result, evidence of what happened at the Pipe Center and the suitability of the Pump for use at the Salon were no longer relevant.

The case on liability was originally pleaded on a number of grounds: breach of the sale contract; misrepresentations made at the Pipe Center; breach of statutory duty under or failure to comply with the Machinery Directive 98/37/EC and the Electrical equipment (Safety) Regulations 1994 and contribution under the Civil Liability (Contribution) Act 1978. By the time of the oral closing submissions breach of the implied term of the sale contract as to satisfactory quality was the one remaining basis for liability.

By the end of the closing submissions there were no issues between Wolseley and Leader because whether Wolseley sold Hi-Lite a pump which was not suitable and questions as to responsibility for the wording of the Instruction Manual were not relevant to the remaining issues of liability. Leader accepted that if Wolseley were liable to Hi-Lite under the implied term of satisfactory quality then Leader would, in turn, be liable to Wolseley.

It is therefore convenient to deal with these issues, in turn:

Given the narrowing of the issues, much of the witness evidence was no longer relevant. I therefore deal only with the aspects of the evidence which remain.

Hi-Lite called three witnesses. The first witness was Mr Anthony Fletcher who had purchased the Pump from the Pipe Center and installed it in the sump at the salon. He has provided a number of statements. His first was signed on 24 July 2006. He then produced a statement in the Birmingham Proceedings dated 21 December 2007 and has more recently produced a further witness statement dated 11 February 2011. In addition I have seen the transcripts of the evidence which he gave in the Birmingham Proceedings.

His earlier witness statements had been provided in the context of the Birmingham Proceedings in which the issue was whether the fire had been caused by some other cause or whether it had started in the Pump. There was little detail in the statements about the work which was carried out by him in relation to the Pump.

His evidence was in a number of respects unsatisfactory. He had produced his third statement recently and whilst it contained some new material, Mr Fletcher was generally unable to recall any of the detail of what happened in November 2005 and therefore there was no satisfactory basis for his recent recollection.

There were also serious inconsistencies in his evidence. For instance, his recent evidence was that the knife which had been found in the sump was a rounded butter knife, but in his previous statements he had merely referred to it as a knife. When cross-examined in the Birmingham Proceedings he had referred to there being a knife in the sump and said that people “could put their hand in and cut theirselves as well”. In cross-examination he sought to say that he had not meant that they would cut their hands on the knife but on other unspecified sharp objects, obviously in the light of his recent evidence that the knife was a rounded butter knife. I do not accept that evidence. It was quite clear that when he made his previous statements and gave evidence in the Birmingham Proceedings it was on the basis that there was a sharp knife and I do not accept that his recent evidence, over five years after the event, that it was a rounded butter knife.

In relation to the question of whether an RCD should have been fitted, whilst Mr Fletcher said that there was no need to install an RCD, it was clear that he had not applied his mind at the time to the question of how the Pump should be connected to the mains but had just made the same type of connection as had been used for the previous pumps. He said he had read the instruction manual but evidently had not considered what was said about the use of an RCD. Mr Fletcher’s evidence that there was some form of isolation switch was not convincing and fluctuated between there being a separate switch and being a switched fused spur. The evidence, including the experts’ joint statement in the Birmingham Proceedings and a photograph of the spur after the fire, showed that the connection of the Pump was to a fused spur which was not a switched fused spur and so did not have an isolation switch.

Mr Fletcher said that when he went to the salon on 22 November 2005 he could not actually remember how he carried out the work and whether he baled out the water before disconnecting the Pump. He said he recalled recovering debris and putting it in a bin and said that the knife was handed over to the manageress either on that day or the following day. He said that when he saw the installation, the filter seemed fine. It was put to him that he told Mr Braund on 30 January 2006 that the filter had not been working. He said he did not remember that; it was via a telephone conversation.

In relation to adjustment of the float switch cable he said that he slid the cable through the clip but in the Birmingham Proceedings he had said previously “I would personally unclip it, adjust it to the required length and then clip it back in at that length. I wouldn’t pull it in or pull it because it could, obviously, ... tension up the cable.” He said he did not recall whether he slid it in and out or unclipped it; he did not remember either way. He said he couldn’t be 100% confident as to what length of cable he left. It was put to him that in the Birmingham Proceedings he said that he had measured it. He said he did not now remember. He remembered that the 10cm was stated in the manual. He said he adjusted the cable when it was plumbed into the sump. He accepted that the access to the sump was not fantastic but it was accessible and Mr Brown had a torch. He was asked whether he used something like a flat screwdriver to wedge out the cable from the clip. He said “100% no”. He disagreed that he did not take much care about using tools if they would do the job. He was asked in re-examination about the suggestion that he may have used a screwdriver for the purpose of adjustment, he said that “there's no way I'd have put a screwdriver to it. It just doesn't seem practical to me”.

The evidence of Mr Ian Higgins Managing Director of Hi-Lite was interposed during the course of Mr Fletcher’s evidence. He confirmed that apart from the invoice there was no record of the work done on 22 and 23 November 2005 and he had no recollection of any work having been carried out subsequently to the Pump in the sump.

After Mr Fletcher had completed his evidence, Mr Wayne Brown was called. He had accompanied Mr Fletcher to the Pipe Center on 23 November 2005 to purchase the Pump and subsequently assisted Mr Fletcher installing it in the sump at the salon. He said that, contrary to what was in his first witness statement he had not been with Mr Fletcher on 22 November but only on 23 November 2005. He said he had been told by Mr Fletcher what had been taken out of the sump and he remembered seeing the knife, cloth and latex gloves. He remembered saying he did not know how the knife got down the plughole. He said the knife was not a steak knife; it was a usual knife, used for everyday eating, a steel knife, just a normal knife. He said that the old pump had been removed by the time he saw it; the pump was dirty and in a horrible state. He said the tank was dirty with gunk stuck to the sides. He said the knife was sharp enough to cut hair, it was a normal standard knife without a serrated edge.

In relation to the replacement of the pump he said that Mr Fletcher located the Pump in the sump, then carried out the pipework and adjusted the cables and connected the flex. He said he was observing and was holding a torch when Mr Fletcher was connecting the Pump to the spur. He said Mr Fletcher used side cutters to cut the flex; he saw a fused spur, not a switch. He said he could not remember how the Pump was located on the sump. He saw Mr Fletcher adjusting either the float cable or the flex. He said that he would have noticed if the float switch was banging against the side of the sump and he did not notice that. He said that Mr Fletcher read the manual; they checked that the float switch would rise and fall but he did not see what the length of the float switch cable was and there was no conversation about whether an RCD should be fitted.

Wolseley then called four witnesses. The first witness was Mr Peter Whistler who was a product director at Wolseley, and had been involved in discussion with Leader about stocking the Jet W300 pump in the Pipe Center stores. He also dealt with the user manual.

The second was Mr Simon Stringer who was employed at the Pipe Center branch at Digbeth. From the initials on the sales invoice he was the person who sold the Pump to Mr Fletcher. He did not recall the transaction but gave evidence of what he usually did when selling products. The third witness was Mr Fred Wallace who was the deputy manager of the Pipe Center branch at Digbeth where the Pump was purchased. He gave evidence of what staff were told about giving product information to customers.

The last witness called by Wolseley was Mr Stanley Wong. He was called after Leader’s witnesses. He was employed as the Quality Manager for Wolseley in the UK and Europe. He gave evidence as to his involvement in carrying out a Potential Failure Mode and Effects Analysis of the Jet W300 pump after the judgment on liability in the Birmingham Proceedings.

Leader called three witnesses. The first witness was Mr Elio Panesi, the Production Director for Leader. He dealt with the cable testing which had been carried out on the float switch and attached cable. In his witness statement he referred to the fact that the W300 pump was derived from another Leader pump which he referred to as the mother pump. The mother pump had been tested by TUV, a German testing house so as to undertake safety checks in accordance with EC standards.

So far as the float switch and cable was concerned he referred in his witness statement to the fact that Leader had instructed VDE, another independent German testing house, to test that assembly to check that it was constructed in such a way as to be safe for use by a consumer. He exhibited to his witness statement a test certificate for “automatic electrical water level controls of the float type for household and similar applications MATE”. That certificate was dated 25 September 2003 and showed that the float switch had been tested for 50,000 cycles and that the maximum water temperature was 35°C. There was a supplement also dated 25 September 2003 which referred to the place of manufacture being Spanker Industrial Limited (“Spanker”) in China.

Mr Panesi said in his witness statement that this was the company that manufactured the float switch and that Leader sent Spanker the drawing for them to build the tools to manufacture the float switch which Leader approved. Leader, he said, also approved the process by which the float switch was to be manufactured and the float switch was then produced by Spanker in China using Leader’s tools and to Leader’s specifications. In his evidence he said that the cable which is used for the float switch was obtained from Leader’s approved suppliers in Spain. He said that the specified cable had previously been certified to comply with EC Directives on quality, service and reliability. He said that the moulding process for the assembly of the float switch took place at a relatively low temperature (210°C) and that he was not aware of the process ever having damaged the cable, nor did he see how in practice it could do so.

Mr Panesi also referred to the fact that further quality checks were carried out on the float switch and cables. He said that the float switch was checked to ensure it was working and that the cable was not damaged and that it was watertight. He said that Spanker checked to ensure that no electrical part was in contact with the plastic by applying 1500 volts to make sure that there was no residual current on the float switch. He said that he did not remember Leader having to reject any of the float switches or cables after testing and that he did not have any experience of pumps being rejected or returned because of moulding defects. Once the final pump was assembled in Italy he said that it was subject to further testing by Leader for water tightness by using compressed air and for electrical safety. Each pump was tested in that way and in addition a further 2% of the pumps were subjected to wet testing.

He said that, as a further safety check, VDE checked annually to ensure that the suppliers were undertaking the safety tests correctly and that the material Leader was using was fit for purpose. He said that Leader had undertaken a number of quality checks on samples of the cable. He referred to the fact that VDE test the float switch and cable in accordance with the relevant EC standard by bending the cable through 90 degrees for 50,000 cycles. He said that the cycle takes a second to complete and a maximum electrical load was applied to the cable. He stated that the test was carried out over a period of 14 hours and that in normal use the cable would not bend at that pace over such a short period but the test was for accelerated usage and was a dry test with no water or tank being involved. He said that Leader also carry out lifetime tests using a tank. In such tests the cable achieves one million cycles. He said that given that the cable can withstand one million cycles before breaking and that the manual recommends that it should not be started more than 30 times an hour, this would equate to a lifespan of some 4 years working at 30 times an hour for 24 hours a day.

Mr Panesi referred to the fact that the float switch is in normal and accepted use throughout the pump industry and that he estimated that Leader had produced about 7.5 million pumps with cable float switches to the same design concept which was also the same as that used by their competitors in about 90% of their pumps. He said that warranty and other returns comprise approximately 0.2-0.3% of production and are principally associated with water ingress into the pump enclosure.

Mr Panesi was asked questions by Mr Shepherd as to the way in which VDE carried out their tests and was asked whether they carried out the tests using a cable which had a strain relief or a cable without a strain relief. He explained that the float switch was initially produced without the strain relief on the cable and that the first tests failed. He said that test failed because they were carried out with less than 10cm of cable. He said that there were discussions with VDE and VDE suggested that the strain relief should be added and Leader had pointed out that in the manual it clearly stated that the free length of cable had to be no less than 10cm. He said they added the strain relief and carried out the test again with a 10cm cable length and this successfully passed the accelerated test of 50,000 cycles.

Mr Panesi said he had not referred to this failure of the cable and nor had disclosure been given in relation to the failed test and the proposal to introduce a strain relief of the cable. This led to Leader producing further documents and to Mr Panesi being recalled on a number of occasions to explain documents which had been produced.

From the documents now disclosed it appears that testing was carried out to the required standards and in a report dated 8 August 2003 VDE noted the following discrepancies “during the flexing test with 30,000 flexings… the cord is broken after 15,590 and 12,350 flexings. Flexing tests with distance 10cm 30,000 flexings is positive.” This then led to the addition of a strain relief and to the use of the following phrase, in English translation, in the instruction manual: “The cable must have a minimal length of 10cm between float switch and the fixing of cable.”

Leader then called Ms Monica Trasinelli who was employed by Leader as a sales officer and assisted Mr Giancarlo Roggiolani, Leader’s sales director, in relation to sales to Wolseley. She gave evidence about the production of the user instruction manual although Ms Simona Oprea was the person involved at the time but had left Leader in April 2004. Ms Trasinelli has liaised with Wolseley after that time. Leader then called Mr Roggiolani who gave evidence of his involvement with Wolseley in respect of sales of pumps and discussions after the fire.

In addition to the witnesses called at the hearing, I was also referred to the witness statements which had been produced for the purpose of the Birmingham Proceedings. The evidence of a number of those witnesses was referred to in submissions. In particular, on the question of whether there had been damage to the float switch cable in service reference was made to the evidence of Karen Wharton, Claire Robinson, Dennis Raistrick, Charlotte Ridding and Charles Leadbrook.

In his closing submissions Mr Shepherd QC pointed out that no live witnesses had been called to speak to the evidence of damage in service. He referred to the fact that a single email had been sent to Ms Robinson by Leader’s solicitors and that nobody had called Ms Wharton whose address was known. He said that where Wolseley and Leader had the means to prove an intervening act but had elected not to call evidence in support of it, the court should be slow to find that an intervening act had occurred.

Mr Moody QC and Mr Dougherty submitted that it was not open to Hi-Lite to complain about any failure to call the evidence of the witnesses in the previous trial. It was pointed out that Hi-Lite’s solicitors had written on 13 December 2010 concerning the exchange of witness statements to say that “Hi-Lite will rely on the statements and transcripts of evidence of the following witnesses from the liability trial between Obsession and Hi-Lite in June 2008”. The solicitors sought Wolseley’s agreement that the seat of the fire was the Pump but, alternatively, sought confirmation as to which of the witness statements/transcripts could be agreed in order to avoid the need for their attendance at trial. The response to this letter was not produced. The witness statements were therefore served by Hi-Lite but the witnesses were not called. There was no objection to the use of that evidence during the trial.

In those circumstances, I consider that under CPR 32.5(5) Wolseley and Leader were entitled to put the witness statements in as hearsay evidence and that is how the parties proceeded. I therefore treat the witness statements of Karen Wharton, Claire Robinson, Dennis Raistrick, Charlotte Ridding and Charles Leadbrook as hearsay statements. In relation to the witness statement of Claire Robinson Leader, in any event, gave a Civil Evidence Act notice dated 13 December 2010.

In the present case, in considering that evidence there are two matters which affect the weight to be attached to it. First, the witnesses, except for Claire Robinson, were called in the Birmingham Proceedings and there is a transcript of their evidence. That evidence has therefore been subject to cross-examination. In the case of Claire Robinson she was not required to be called. Secondly, the engineering experts, including Hi-Lite’s expert, have referred to and relied on the evidence of Claire Robinson both in the Experts’ Joint Statement and in their expert reports.

In these proceedings Hi-Lite instructed Mr Mark Coates who is a cable engineering consultant with Cobham Technical Service. He has a degree in mechanical engineering and after working in industry joined what became Cobham Technical Services where he initially designed equipment to test the resistance of domestic flexes to continued twisting and bending. He has since worked on a number of projects, including consideration of the mechanical performance of cables and joints, and has written a number of technical publications on the subject.

Wolseley instructed Mr Michael Champion who is a chartered electrical engineer and a member of the Institution of Engineering and Technology and has since April 1994 been an independent electrical safety consultant. For 22 years prior to that he was employed by the Health and Safety Executive latterly as HM Principal Specialist Inspector (Electrical Engineering). From 1978 until 1994 he was the Senior Specialist Electrical Inspector advising HSE Inspectors and local authority Environmental Health Officers and members of the public on all aspects of electrical safety. His work involved the investigation of hundreds of electrical shock and burn accidents and also incidents including fires.

Leader instructed Mr Matthew Bullen an independent Consultant. He has a degree in mechanical engineering is an Associate Member of the Institution of Mechanical Engineers and a Member of the British Academy of Experts. He was an associate and then partner with Dr J.H. Burgoyne and Partners, who investigate fires and explosions. Prior to that he was senior scientific officer at the Fire Research Station. He has investigated some 1300 incidents of fire explosion and mechanical failure and produced a number of publications.

In accordance with the court’s directions, the experts met and produced a joint statement dated 2 December 2010 (“the Experts’ Joint Statement”), following meetings over a number of months. At section 7 of the experts’ joint statement they dealt with the ignition process. They defined the three conductors in the float switch cable as the “go”, “return” and “earth” conductors. The go conductor, which would be live at 230 volts when the float switch was open, connected the pump motor windings to the switch. The return or neutral conductor connected the switch to the neutral of the incoming electricity supply. The third conductor was the earth conductor.

At paragraph 7.1 of the Experts’ Joint Statement the experts agreed the following sequence of failure of the float switch cable as a “plausible supposition”. The process commenced with conductor strands starting to break in one or more conductors due to a fatigue failure leading to arcing which continued until the cable sheath had ruptured and a fire initiated.

Then at paragraph 8.1 of the Experts’ Joint Statement the experts stated as follows: “We agree that failure of the pump float switch cable would not be expected to occur in the two month period (22 November 2005 -24 January 2006) of normal anticipated service operation at Obsessions.” They continued:

“we therefore agree that an unusual or abnormal feature promoted the premature failure. We are aware of several theoretical possibilities:

(a), a defect in manufacture of the float and cable assembly;

(b)

damage during transport/storage

(c)

damage in service

(d)

damage during installation

(e)

erroneous installation, in particular abnormally short float switch cable length.”

Considering each of those possibilities, the experts agreed that a manufacturing defect was a theoretical possibility. They agreed that it was unlikely that a pre-existing cable manufacturer’s manufacturing defect would have been present in the length of cable used for the float switch and happened to coincide with the location of observed electrical damage.

Mr Coates considered that a manufacturing defect “cannot be entirely discounted”. He stated that the point of failure was close to the moulded strain relief at the float end and an error during moulding of the strain relief onto the cable could have damaged the cable at this point, possibly by an over temperature during moulding. He said that such damage may not be revealed by a visual inspection or electrical testing. Mr Champion considered that a manufacturing defect “cannot be entirely ruled out” but assuming that all Jet 300W pumps were tested electrically and mechanically prior to shipment, “a manufacturing defect is unlikely”. Mr Bullen also considered a latent manufacturing defect “to be improbable”. He noted that the pump manufacturers, so far as he was aware, had no history of float cable failures associated with moulding errors. He also suggested that a gross moulding error would be demonstrated by inspection and the 100% high voltage test on completion of the float sub-assembly or full assembly. He was also of the view that such a gross defect should be visually apparent at the time of manufacture, purchase and installation but none had been reported.

The experts all agreed that damage during transportation or storage was “improbable”.

The experts agreed that damage in service might have occurred in view of Hi-Lite’s evidence that a knife was found in the sump, presumably to assist with cleaning the Pump or sump.

The experts agreed that damage occurring during installation “is theoretically possible”.

In relation to whether or not there was a shortened float switch cable, Mr Coates considered that, on the basis that the failure occurred close to the float switch and not at or near the retention slot, the installed length of float switch cable could not be determined from the physical evidence. Mr Champion and Mr Bullen suggested that a shortened float switch cable constituted a clear and demonstrable stress raiser which, alone, would correspondingly shorten the fatigue life of the cable and explain the incident. In Mr Bullen’s and Mr Champion’s opinion the basis of the finding at the previous trial remained: that the premature failure of the Pump was probably principally due to its installation with inadequate float cable length. Mr Coates’ opinion was that the finding at the previous trial was based on the conclusion that the cable had failed at the retention slot and the physical evidence did not support that conclusion.

The experts however agreed that the location of a fatigue failure in a flexible cable might or could be influenced by any mechanical damage to the cable sheath, whether from a manufacturing defect or damage in service or during installation, bearing in mind that the new pump was installed approximately nine weeks before the fire. All experts agreed that a failure of the float switch cable after nine weeks service was premature. In Mr Champion’s and Mr Bullen’s opinion a fatigue failure after only nine weeks service implied an excessive flexing frequency or number of cycles of operation, leading to an excessive stress probably arising from a shortened length of free cable to the float switch or damage to the float switch cable or a combination of some or all of these factors. Mr Coates however was of the opinion that the premature failure was indicative of a defect in or damage to the cable. Mr Bullen considered that mechanical damage to the cable in service could have accelerated its failure and pointed to the evidence of a possible electric shock in the witness Statement of Clare Robinson which would indicate that the inner insulation and outer sheath had failed before the operational integrity of the Pump was lost. He considered that this was evidence of mechanical damage rather than premature fatigue failure in a shortened float switch cable.

The experts then considered Ms Robinson’s evidence that she had suffered an electric shock. Mr Bullen and Mr Champion took no issue with Mr Coates’ comment that he considered that two coincident failures would be unlikely. He said that in his opinion it was more likely that the shock incident was related to the cable failure and that damage to the cable sheath could have led to water ingress into the cable and hence ingress down inside the cable sheath to the inside of the float. He said this would lead to a conductive path from the water in the tank to the live terminals in the float.

The experts then filed their individual expert reports. At section 12 of his expert report Mr Coates concluded that the most likely cause of fire was fatigue failure of the float switch cable close to the strain relief on the float. He considered that it was likely that the fatigue failure occurred because of increased mechanical stress on the conductors due to a combination of shortening of the float switch cable and damage to the cable. In his report at section 11, Mr Champion dealt with the underlying reasons for the float switch cable failure. He said that the most likely cause of fatigue failure was excessive flexing but the number of flexures likely to have occurred in a two month period suggested there may have been other factors involved. He concluded that the most likely cause of fatigue failure was a combination of failures of shortening of float switch cable to less than 10cm and possible damage during installation or use and that the most likely cause was due to shortening of the float switch cable but that damage during use which could accelerate the failure might be an explanation. Mr Bullen in paragraphs 1.3 and 1.4 of his report said that he concurred with the conclusion from the previous technical analysis that the cause was likely to have been associated with shortening of the float switch cable but, taking account of the evidence of Ms Robinson he considered that the failure was likely to have been associated with mechanical damage and the fire was likely to have been caused by the development of an electrical fault as a result of that damage.

When the experts came to give evidence at the trial their views had developed further. This was, in part, because of the further information which had been produced by Leader concerning testing of the cable and the decision to install a strain relief section to the end of the cable after a failure during testing.

In his evidence Mr Coates was asked about this information and he said that the fact that there had been a failure under test and that the strain relief was not part of the original design may have clarified the expert discussions and the information on the number of cycles to failure without a strain relief. However he said he recognized the test did not replicate the condition of service of the float switch cable. Overall he said there was some useful information but it did not dramatically alter his conclusions.

In summarising his view about the failure he said that two conductors in the float switch cable had parted at 13cm from the end where they terminated within the float switch and one conductor was partially broken. He therefore concluded that this was indicative of fatigue failure of the cable. By measuring the cable length on other float switches he said that this placed the position of the failure somewhere between 0.5cm and 1.5cm from the end of the strain relief towards the Pump body. The expert view in the Birmingham Proceedings was that the fatigue failure had occurred at the point where the cable exited from the clip on the side of the Pump. He said that his conclusion, contrary to that view, was that the fatigue failure had actually occurred at the end of the strain relief on the float. He said that where previously the point of failure was taken to indicate that the cable had been shortened that was no longer the position and there was no definitive evidence as to what the length of the float cable was when it was installed.

He said that the experts were now agreed that if the cable had been shortened to a free length of 3cm it could not have operated reliably. This was because the length of the strain relied was itself 2.5cm. As stated in paragraph 5.5 of the Experts’ Joint Statement the float switch would not reliably switch the Pump off if the cable was less than 3cm long, in addition to the 2.5cm of the strain relief.

Mr Coates said that he had great difficulty in seeing how the shortening of the cable alone would lead to a failure in just 9 weeks of operation although shortening of the cable would lead to higher stresses and an earlier fatigue failure than would be expected with a longer cable. He said that the fatigue failure was a function of the magnitude of changing stress and the number of cycles. He did not think there had been an excessive number of cycles because there was no evidence of that on the contact on the micro switch of the Pump. This is a reference to paragraph 8.17 of his report and to paragraph 3.2.7 of Mr Braund’s report in the Birmingham Proceedings. Mr Braund had examined the debris of the Pump which had been recovered after the fire. He had found a contact of the switch which was attached to a wire connected to the Pump. He reported that the electrical contact surface of the switch had not been damaged by electrical arcing. On this basis Mr Coates said that, if the Pump had been stopped and started for sufficient cycles to cause fatigue failure of the float switch cable, he would have expected to see pitting and erosion of the contact surface. This, as he said in answer to questions by Mr Moody QC, showed that there was no evidence that the float switch had switched on and off an excessive number of times in the nine week period which meant that the cause of the fatigue failure of the cable was excessive stress rather than an excessive number of cycles.

In this respect Mr Coates identified three “stress raisers” which would increase the stress from what it would have been with whatever length of cable had been installed. Those stress raisers he said, could first have been damage to the outer sheath or to both the outer sheath and the insulation of a conductor in the float cable. He said that for physical damage during installation or use to have led to a failure occurring where it did would mean that the damage would have to be coincident with the point of higher stress just outside the strain relief.

Initially in his oral evidence he said he thought the damage to that area was less, rather than more, likely because he did not see such damage occurring in that area during day to day use, even with the presence of a knife of some description in the sump or with Mr Fletcher using a screwdriver to prise the cable from the clip. However in his evidence to Mr Dougherty he accepted that there was a short length of float switch cable between the handle and the end of the strain relief and that it would not be surprising for damage to occur in that area if it was assumed that somebody was cleaning that piece of cable with an instrument capable of cutting the sheath. He also accepted that if there was damage to the cable the most significant damage would be any damage close to the change in section of the cable at the end of the strain relief. He accepted that it was a rational explanation for the damage but that it was still a coincidence. He said the experts had all agreed that there had to be some coincidences.

In relation to the possibility that damage was caused by a screwdriver used by Mr Fletcher to prise the cable out of the handle, he thought this was unlikely based on the fact that Mr Panesi had said that the nominal length of cable when the Pump was produced by Leader was 10cm. When asked whether it could happen in adjusting the cable and unclipping and clipping it he said it was something he could not entirely discount. He said that if damage was at the point around the area of maximum stress a “nick” in the outer sheath could open up and develop an outside to inside fault and he accepted that over nine weeks what might have been a very small nick originally might have developed into the failure which caused the fire.

Mr Coates said that the other area of possible damage or other possible cause of the damage to the cable sheath near the float was the manufacturing process. He said he was not entirely clear on the manufacturing process but he thought it possible that the boss and/or strain relief were injection moulded at the same time as the parts of the switch were fused together. He said that the moulding operation was carried out at 210°C which was higher than the manufacturer’s recommendation of 200°C for 5 seconds. He referred to standard EN60335-1 which warned about the possibility of damage to the cable sheath during a moulding operation. It stated “the insulation of the supply cord shall not be damaged when moulding the cord to part of the enclosure. Compliance is checked by inspection.” He said that visual inspection is probably the only thing that can be done but that it is not a “guaranteed, 100% method” for determining that there has been no change in the physical properties of the material of the cable.

Mr Coates was taken in his oral evidence to paragraph 8.41 of the Expert Joint Statement and paragraph 8.16 of his expert report. He accepted that he had not met a failure in a float switch before and had not heard of any history of such failures. He had seen Mr Panesi’s evidence that he could not see how such an error could occur in practice. On this basis he was asked whether he was putting forward a manufacturing defect as a possibility but no more than a possibility. He said “it’s just a possibility. I have no evidence to say it actually occurred. The only thing that I found since writing this report is the reference in the base standard for the pump regarding moulding of material onto cable.” It was suggested to him that the level of certainty he put on it was very, very low and in the Experts’ Joint Statement he had said that it was something that “cannot be entirely discounted”. He said that was correct. It was put to him that this suggested that whilst he was not willing to rule it out completely he was not suggesting that it should be at the forefront of any explanation for what occurred. He said that was correct.

He accepted that the temperature of 210°C for the moulding was only marginally higher than the cable manufacturer specified for the cable. He said he was aware of tests carried out by Mr Bullen by placing the cable in an oven for half an hour at 230°C with no visible difference. It was put to him that it was right to say that a manufacturing defect was no more than Mr Coates had originally put it, a possibility that cannot be entirely discounted. He said that was correct. When asked “but it is no higher than that?”, he said “no”.

As referred to by the experts in the Experts’ Joint Statement, there was evidence of an electric shock felt by Ms Robinson. This is a reference to evidence in a witness statement dated 25 February 2008 produced on behalf of Obsession for the Birmingham Proceedings. In that statement at paragraph 7 she had said:

“I recall a few days before the fire, feeling a small electric current going through my arm as I put my arm into the tank of water to empty the water to gain access to the filter. I felt the shock travel up my arm. It only happened once. I recall this was 2 or 3 days before the fire. I did not mention it to anyone until after the fire, when I mentioned it to Lillie, the salon Manager.”

In the Expert Joint Statement Mr Coates had said that in his opinion it is more likely that the shock incident was related to the cable failure. When he came to give his evidence he said that his thoughts had moved on a little on this subject. He said that if the shock related to the fire then the shock would be indicative of there having been damage to the cable sheath. He said that if there was damage to the cable sheath only and not a cut through the conductor insulation as well to leave a live conductor directly in contact with water, then for Ms Robinson to receive a shock there would have to be a sufficiently conductive path from the point of damage of the cable sheath through to a live part, being the termination of the go conductor in the float switch. He then explained some calculations which he had just done. He said that the conductor path was approximately 13cm and if it were assumed that Ms Robinson received a shock of 1mA, the limit of perception being 0.5mA and if the voltage was 230 Volts, then with water having the conductivity of laundry water he calculated that a sectional area equivalent to 3mm² of water would be needed between the damage to the cable sheath and the switch. He said that he thought that there was hardly room for 3mm² of water within the cable.

Mr Coates said that he thought it unlikely that the shock was caused by a cut to the go conductor because that conductor had only partially failed whereas the other two conductors had fully failed. His conclusion was that this did not mean that the shock could not have been caused by damage to the sheath but it was a factor that was leading him towards the view that it was more likely that the electric shock to Ms Robinson was a coincidence and was not related to any damage that have might have led to the fire.

In his evidence in answer to questions from Mr Dougherty he said he was not saying that his original view was necessarily wrong. He thought that his original view was less likely because he had not considered what was necessary to achieve the electric shock. He said he had not previously done a calculation of the area available within the cable but had inspected it. He accepted that it was rather difficult to determine the available area by looking at the cable. He said the area was not exactly straightforward because the internal diameter of the sheath was not circular as the sheath material protrudes into the gaps in the trefoil between the three cords.

In his oral evidence Mr Coates raised, for the first time, another factor which he said might affect the stress in the cable and that was the temperature of the cable. He said that the warmer the cable the softer the sheath, so that it would provide less support. He said that he could not quantify the effect of a 10°C difference. In answer to questions about this, he said that the ambient temperature would be around 20°C and the water in the sump might be about 30°C. He accepted that the manual for the Pump said it should not be used with fluids over 35°C. Mr Coates was asked whether he relied on temperature as being an explanation of the failure. He said that he did not, it was “just another factor”. He said that the flexing behavior of the cable would change with temperature. He agreed that temperature was just another factor in the equation, a background matter rather than a cause of the failure. Therefore the fact that the cable was operating at those temperatures was not a matter which Mr Coates relied upon as being the cause of the fatigue failure and the consequent fire.

Because the evidence of Mr Champion and Mr. Bullen overlapped to a great extent, it was agreed that they should both be sworn and called to give concurrent evidence, a practice adopted where appropriate in the TCC. In his evidence Mr Bullen confirmed that the breakage of the return conductor was a fatigue based process which, in his view, developed from shortening of the cable which was an accelerator of the damage. He stated that this was an area where his view had developed from the time of the experts’ meeting. He said that, at that time, he considered the shortening was part of the likely process for the fire but felt that the number of cycles which were likely to have been experienced did not explain the incident and therefore that damage was likely to have been material in starting the incident. In relation to the possible cause of the damage Mr Bullen considered that a manufacturing defect was improbable. He said that the float switch was produced as a mass produced item and was quite a sophisticated plastic moulding. He said that as a result the moulding process had to be very carefully controlled to make sure that the molten plastic did not damage the upper or lower shell of the switch. He said he would expect the production process to be very sophisticated in its control, probably by a computer. He referred to his test on a cable being placed in an oven for 30 minutes at 230°C. He said the cable seemed to be unaffected by that process. There was no shrinkage or significant change that he observed. In relation to this Mr Champion said that in his view if anything had gone wrong with the heating time or level of heating there would not just be one failure but there would have been reports of many failures because there would have been a batch of failures rather than a unique single failure arising from the incident.

In relation to damage in installation and service Mr Bullen said that the cable was flexible and was capable of being damaged. He said that if insufficient care was taken, either during the installation or during cleaning, there was the possibility of damage to the cable. Mr Champion considered that the frequency of access played a part in his assessment of the probability of when the damage occurred. He said whilst it was possible for Mr Fletcher to have damaged the cable he was there for two hours making adjustments whereas the Obsession staff, if they carried out cleaning in the vicinity twice a day, as they say they did, had more opportunities to create damage to the float switch cable as and when they found it necessary to open the tank and inspect or remove debris. As a result he thought the most probable source of the damage came from the Obsession staff rather than Hi-Lite.

In relation to the evidence of the electric shock to Ms Robinson, Mr Bullen said that when he saw her evidence it “jumped out of the page” that there had been a failure of the Pump and a few days before the incident Ms Robinson had felt an electric shock. In his view there was a likelihood of a relationship between the electric shock and the fire. He said that her evidence pointed to electricity leaking somehow from the Pump. The most obvious location in his view was damage to the float cable which was electrifying the water. He said that the most likely damage was mechanical damage which affected the outer sheath which had also acted as a focus for the development of fatigue. Mr Bullen considered that the shock could have come either by an electric path to the float switch, if only the sheath was damaged or it could come from an exposed conductor, if both the sheath and the insulation to that conductor were damaged. He said that, in principle, both were possible and there could have been an exposed “go” conductor. In relation to Mr Coates’ calculations it was pointed out that, in his report at paragraph 8.19, Mr Coates had said that it was likely that the shock felt by Ms Robinson was less than 0.5mA but yet based his calculation on 1mA. Mr Bullen said that if 0.5mA were used, the area required in Mr Coates’ calculations would be reduced to 1.5mm².

Mr Champion agreed that there were the two possibilities referred to by Mr Bullen but his view was that to suggest that the damage had been through the “go” conductor was “a leap too far”. His view was that the carbonisation process, which he explained was a slow process, had occurred and the charred insulation was giving a conductive path to the water. He thought that if there had been mechanical damage to the “go” conductor anyone putting their hands into the tank would have received a shock.

In relation to the RCD the experts differed on whether an RCD should have been fitted and also on whether, if an RCD had been fitted, it would have prevented the fire. That led to a difference between the experts on the precise mechanism by which the fire had been started.

The Pump was originally supplied with a mains lead fitted with a conventional moulded 13 amp plug complying with BS1363-A. I am satisfied that Mr Fletcher cut off that plug and connected the mains lead from the Pump directly into a fused spur which was not a switched spur. The supply to that spur was provided from a ring circuit protected by a 32 Amp circuit breaker. Mr Braund, who carried out his inspection after the fire, was unable to identify precisely which ring circuit fed the Pump. Whilst, as all experts agreed, it would have been appropriate under the IEE Wiring Regulations (16th Edition) to provide a switched spur, that was not relevant to the cause of the fire.

The installation instructions for the Pump in the instruction manual included a statement that “Before inserting the plug, check that: (1) The electrical installation is equipped with a life-protector switch (breaker box-I_Δn=30mA)…”. The experts agreed that this is a reference to a residual current device (RCD) which under BS7671 should have a rated residual operating current (I_Δn) not exceeding 30mA and an operating time not exceeding 40 milliseconds at a residual current of 5 I_Δn. The experts also agreed that a competent electrician reading those instructions would have concluded that the manufacturer recommended the installation of an RCD to protect the Pump circuit.

Mr Coates considered that there was no good practice requirement, other than the need to follow the manufacturer’s installation instructions, which required the installation of an RCD in the Pump circuit. Mr Coates disagreed with Mr Champion and Mr Bullen that, in the circumstances of this case, BS7671required an RCD to be fitted in the circuit supplying the Pump. At Clause 471-8-01 of BS7671 it states as follows:

“For installations of increased shock risk such as those in Part 6, additional measures may be required such as:

(i)

Automatic disconnection of supply shall be by means of a residual current device having a rated residual operating current I_Δn not exceeding 30mA.

Part 6 of BS7671 describes the special situations, which include wet situations such as a bath, shower cubicle, swimming pool, garden pond, septic tank and construction sites. It does not refer to immersed sump pumps and Mr Coates did not consider that it came within the definition of “installations of increased shock risk such as those in Part 6”. Mr Champion and Mr Bullen considered that the situation of a submersible pump with a flexible cable gives rise to similar dangerous wet conditions to those in Part 6 of BS7671. Mr Coates supported his conclusion by reference to Clause 601-09-02 of BS7671 which excludes a fixed water heater and a fixed shower pump from the requirement to have an RCD. Mr Champion differentiated between that type of fixed equipment and the Pump in the present case. He referred to Clause 601-08-02 of BS7671 which required that, where a shower cubicle was installed in a room other than a bathroom or shower room, a socket outlet should be protected by an RCD. Mr Coates considered that this clause was irrelevant because it is dealing with a socket outlet not a fused spur and the sump pump was an item of fixed equipment. Mr Champion also relied on Clauses 602-0702 and 604-0803 of BS7671 and Clause 7.12 of BS EN 60335-2-41 as supporting his argument. Mr Coates considered that those regulations were irrelevant because the Pump was not an item of portable equipment, it was not connected to a socket outlet, it was not hand-held equipment and it was not used with a trailing lead on a construction site.

Mr Coates was asked questions by Mr Moody QC on the need to install an RCD, particularly if Mr Fletcher knew that people were putting their hands into the sump which housed the Pump. Mr Coates expressed the view that installing an isolating switch would have been sufficient on its own. Mr Coates however accepted that it was foreseeable that sometimes a user might need to drain down the sump by manually lifting the float cable if they wanted to drain down the sump and that the Pump would still be connected to the power during that operation. Mr Coates agreed that in those circumstances the Pump could not be isolated and that there was a risk of an electric shock arising. When Mr Moody QC put to him that an RCD would be required to protect the user in those circumstances Mr Coates said “In those circumstances an RCD is one option” and “It’s the preferred option”. Mr Coates accepted that if somebody was going to have access to the Pump while it was still connected to the mains an RCD should be provided.

A difficult technical question which is then posed is what would have been the effect of installing an RCD in terms of whether it would have tripped and therefore prevented any fire. The question involved the experts having to consider what sort of electrical fault would have developed and whether it was the type of electrical fault which would have caused the RCD to trip.

At paragraph 11.9.1 of the Experts’ Joint Statement, the experts agree that if an RCD had been fitted in the pump circuit the RCD would have been expected to operate in the event of an earth leakage fault. In this case that would have been either a fault between the “go” and “earth” conductors or, when the float switch was on, a fault between the “return” and “earth” conductors. Their view was that the RCD may or may not operate in the event of an earth leakage fault between the “return” and “earth” conductors when the float switch was off. They agreed that the RCD would not operate in the event of a “go” to “return” short circuit fault. They also agreed that an RCD would not operate in the event of in-line arcing occurring before a “go” to “earth” fault developed.

The general sequence which the experts agreed at paragraph 7.1 of the Experts’ Joint Statement as being a “plausible supposition” was as follows. First, that the conductor strands would start to break in one or more conductors due to fatigue failure. There would then be complete severance of all the strands in either the “return” or the “earth” conductor, but at that stage the cable insulation and sheath would remain intact. When the float switch rose to a high level and started pumping so that it was on, in-line arcing would occur at the break in the “return” conductor which might or might not cause the Pump to operate intermittently at that stage. When the float switch moved to the off position when it was at a low level then in-line arcing in the “return” conductor would cease. Multiple cycles of arcing then lead to carbonisation of the “return” conductor insulation from within forming a resistive path gradually reducing in resistance and, in turn, leading to further heat damage and damage to the insulation of the other conductors and the sheath which would be pyrolysed. The question was then what would happen at that stage?

Mr Coates considered that the degradation of the insulation and sheath materials due to the heat from the arcing would generate gases within the cable sheath. These gases would then pressurise the insulation and cable sheath and the arcing would also heat and soften the insulation and sheath materials. The internal pressure due to the gases combined with the softening of the insulation would then cause the insulation on the conductors to rupture and the degradation process would continue to generate gases and internal pressure which would then cause the sheath to rupture explosively. Mr Coates said that this phenomenon was observed on cables subjected to rapid heating during high current, short time tests. He considered that the insulation on the conductors would not fail due to explosive rupture but would be a more gradual process of thermal degradation. On that basis his view was that the sheath would rupture before an electrical connection was made between two conductors within the cable. Thus his opinion was that the rupturing of the sheath would lead directly to ignition before the slower process of thermal degradation of the insulation of the earth conductor had developed to an extent where there was an earth fault of more than 30mA of earth leakage current to trip the RCD.

Mr Champion did not agree with that mechanism of causing the fire. He was of the opinion, taking into account the close proximity of the internal conductors in the helical formation inside the cable and the thickness of the conductor insulation, that the carbonisation of the insulation to the “go”, “return” and “earth” conductors would cause that insulation to reach a high resistance value leading to earth leakage currents between the “go” and “earth” or between the “return” and “earth” conductors before the sheath ruptured and a fire could be initiated. Mr Champion stated that Mr Coates’s observations regarding an explosion arising from internal gas pressure was based on tests on cable subjected to rapid heating during high current short time tests. Mr Champion stated that in the return conductor the current was limited to a maximum running current of about 1.8 amperes and a starting current of 7.2 amperes for one to two seconds, which did not represent the type of high current conditions which were required for Mr Coates’s mechanism to operate. As a result, Mr Champion’s opinion was that a high resistance, low energy, electrical connection between two of the internal conductors was likely to occur before penetration of the sheath. Mr Bullen endorsed Mr Champion’s views and was of the opinion that the explosive failure process was not supported by tests done by the previous experts and did not seem to be consistent with the sequence of arcing events envisaged.

Mr Coates was asked questions by Mr Dougherty and was referred to pictures in Mr Braund’s report, in particular photographs 6 and 7, relating to Mr Braund’s test 2. Those showed, as Mr Coates accepted, that there was carbonisation of the cable prior to ignition of the cable. Mr Dougherty suggested to Mr Coates that the likely mechanism was that there would be carbonisation before ignition of the cable and that once the sheath has ruptured, the gases would catch light and there would be a heating effect of the arcing on the surrounding material resulting in ignition. Mr Coates accepted that this was “certainly a probability that I have not discounted completely” but he considered the probability of the internal gas pressure rupturing the cable was the higher probability. He said he was not saying that Mr Champion and Mr Bullen were wrong but that he preferred his alternative theory. He accepted that if there was already mechanical damage causing a break in the sheath then the explosive rupture would be limited to the insulation of the conductor.

In his evidence Mr Bullen said that although he appreciated that an explosive failure could occur in high current faults this was case with a relatively small current. He was of the view that any process of explosive eruption would tend to blow away the initial fuel from the fire, based on his experience of cable faults in domestic irons, with the net result that there would be a bang but no sustained fire to follow. Although a fire might have developed in due course following such an event, the initial explosion might well not produce a flame. Mr Champion confirmed that, in his view, the process was a gradual one of carbonisation rather than a “big bang”. He said that with carbonisation it would take time before deterioration took place. He said that the insulation to the cables and the sheath material were hydrocarbon based materials so that when arcing took place there would be a breakdown by way of carbonisation. He did not think that that process would liberate sufficient gases for the explosion relied on by Mr Coates and that, in any case, any gases which were produced would be vented along the cable.

During the course of the cross-examination of Mr Bullen and Mr Champion Mr Shepherd QC put forward a new calculation which Mr Coates had just produced, relating to the heat generated in an arc. This was in response to paragraph 11.12 of the Experts’ Joint Statement where Mr Champion had said he considered that the carbonisation of the conductor insulation from in-line arcing was a slow process leading initially to a high resistance conductive path between the “return” and “earth” conductors or between the go and earth conductors which was likely to occur before the sheath was ruptured. Mr Champion’s view was that if the fault current exceeded 30mA and persisted for 0.2 seconds then the RCD would trip and on that basis the power delivered into the fault would not be sufficient to penetrate the cable sheath. A document subsequently produced by Mr Coates set out his calculation. With an arc across a 0.1mm gap this showed that 3.6 watts would be generated which, as Mr Coates pointed out, would be concentrated over a small area.

In cross-examination by Mr Shepherd QC, Mr Champion said that the process of carbonisation would start with the neutral conductor leading to carbonisation of the other conductors. He said that for the current to leak there would not need to be total carbonisation of the whole of the insulation. He said that all that was needed was a leakage path of sufficient resistance to enable 30mA to flow. On this basis he said he was satisfied that a leakage or tracking path could develop where 30mA would be generated between the neutral or go conductors and the earth conductor so as to cause the RCD to operate. He said he had wide experience of leakage paths within control circuits of less than 30mA which had caused malfunctions to machinery.

The three experts in this case have overlapping areas of expertise. Mr Coates is an expert in cables; Mr Champion is an expert in electrical safety and Mr Bullen is an expert in the investigation of fires which often involve electrical failures. Mr Coates’ evidence has had an important impact on the case. His view that the failure of the cable could not have been caused in nine weeks merely by shortening of the cable and the limited number of cycles, meant that another factor was necessary to explain the failure of the cable and the fire. He saw the significance of the lack of pitting to the switch contact in showing that the Pump had only cycled on and off a limited number of times. He also realised that the failure of the cable could not have occurred at the cable clip of the Pump body.

I have therefore found Mr Coates’ views in the Experts’ Joint Statement and in his report very helpful. When however he came to give evidence he developed a number of new ideas which I formed the view had not been properly thought through and which departed from the more considered views in the Experts’ Joint Statement and his report. Thus he sought to distance himself from the relevance of the electric shock felt by Ms Robinson and sought to develop other ideas which were not focused on the issues. On cross-examination he conceded a number of central aspects of his evidence.

This has led me to prefer the views of Mr Bullen and Mr Champion who initially, with the exception of Mr Bullen’ view on the electric shock felt by Ms Robinson, appeared to approach the case on the basis of the expert evidence given in the Birmingham Proceedings but then accepted the views of Mr Coates, particularly on the need for there to be a further factor to explain the fatigue failure of the cable at an early stage. I found their oral evidence helpful in explaining their views and where they differed from the views of Mr Coates.

There was no issue between the parties that the fire had started in the float switch cable on the Pump. There was however a difference between the parties as to how the fire had been caused in the float switch cable.

As I have stated, by the end of the hearing Hi-Lite’s case was essentially that Wolseley was in breach of its obligations under s.14(2) of the Sale of Goods Act 1979 because there was a manufacturing defect in the float switch cable which had been caused by too much heat being applied during the moulding process or by the heat being applied for too long. Wolseley and Leader contended that there was damage to the float switch cable which had been caused during installation by Hi-Lite or by staff of Obsession whilst the Pump was in service.

The float switch cable came out of the body of the Pump and was then clipped into a slot in the handle of the Pump. There was then a length of cable which ended at the 2.5cm strain relief moulded to the cable before the cable went into the boss which formed part of the float switch. The instruction manual referred to the length of the float switch cable between the slot and the float switch and stated that “Please note; the length cannot be less than 10cm”.

In the Experts’ Joint Statement the experts had defined the free length of the cable as being the distance from the retention slot on the Pump handle to the place where the strain relief started. In fact, as Mr Panesi explained the length included the strain relief so that the experts’ definition had been based on an equivalent of a 7.5cm length, because the 10cm length included the 2.5cm strain relief.

The experts agreed that a distance of 3 to 4.5cm, excluding the strain relief or 5.5 to 7 cm with the strain relief, was necessary for the Pump switch to be able to operate. The cable could not therefore have been shorter than 5.5cm.

In the Birmingham Proceedings the experts had concluded that the cable had broken at the place where it was clipped into the handle. In these proceedings the three experts were unanimous in their view that the cable had broken at between 0.5cm and 1.5cm from the start of the strain relief and that this was not the location where the switch was clipped into the handle. As a result they could not say how long the cable was, other than that it must have been at least 5.5cm to 7cm for the float switch to operate.

However, whilst the experts agreed that the shorter the float switch the greater would be the stress at the point where the break occurred, they were unanimous in their final view that over a period of nine weeks even with a cable of 5.5cm to 7cm there were not a sufficient number of cycles for this to be the cause of the fatigue failure. An important piece of evidence in this respect was that referred to by Mr Coates. He pointed out at paragraph 8.17 of his report that at paragraph 3.2.7 of Mr Braund’s report he reported that he had found a contact surface of the float switch in the fire debris which had not been damaged by electric arcing. As Mr Coates pointed out and I accept, if there had been an excessive number of cycles then he would have expected to see pitting of the contacts. This establishes that an excessive number of cycles was not the cause and tends to show that the cable length was not so short that it had to operate very frequently.

Rather, the experts all agreed in the end that there had to be damage to the cable which acted as a stress raiser and that the failure which occurred and caused the fire would not have occurred without the stress raiser. Thus whilst any shortening of the cable might have applied a higher stress at the point of damage than would have been applied with a longer cable, the operative cause of the cable failure and fire was the damage to the cable. Without that damage the cable would not have failed as it did whatever the length of cable. Equally, it is not suggested that the difference in practical terms between 5.5cm or 7cm and 10cm meant that the cable failed at that shorter length but would not have failed at a length of 10cm.

Tests were performed by Mr Braund and Dr Lipczynski, the experts in the Birmingham Proceedings, to try and simulate the conditions that led to failure. As the experts in this case all agreed, those tests did not use loading or support conditions which replicated what would happen to a float supported in water and subject to a number of filling/emptying cycles in the sump. What they do show is that, even with such aggressive loading and support conditions the cable, without more, would not have failed after nine weeks, a conclusion more than borne out by the tests which Leader carry out which include a test to one million cycles.

The issue between the experts related to the cause of the damage to the float switch cable, which all experts consider led to the fatigue failure of the float switch cable which caused the fire.

Mr Shepherd QC in closing submissions deals with the question of the liability of Wolseley in a number of ways. First, he submits that it is sufficient for Hi-Lite to rely on the fact that there was damage to the float switch cable after nine weeks which raises an inference that this cable was not of satisfactory quality. He submits that it is then for Wolseley and Leader to show what the cause of the failure is. Secondly, he submits that Wolseley and Leader have called almost no evidence about the Pump and, in particular, the manufacturing process for the float switch cable and what checks there were of the moulding process for the cable/float switch assembly. Thirdly, he submits that Wolseley and Leader have not done what is necessary to show that the cable was damaged in installation or in service. They have not carried out tests to show how easily the cable could be damaged or called any evidence from Obsession witnesses as to any damage caused to the cable in service. He submits that Wolseley and Leader have not discharged the evidential burden of showing that the cause was, in fact, damage during installation or in service.

Mr Shepherd QC also criticised the approach taken by Wolseley to the position of Hi-Lite, in particular, in the context of the Birmingham Proceedings and these proceedings. He says that Wolseley had the opportunity to allege that the cause of the fire was damage caused to the cable during installation or in service but they failed to do so. He referred to correspondence from 2006 to 2008 and the way in which Wolseley pleaded its case in these proceedings, including the amendment to the Defence on 22 January 2011 to plead, as follows:

“It is admitted that the Fire originated in the W300. The fire was caused by arcing in the float cable caused by a fatigue failure. The fatigue failure was in turn caused by the fact that the cable was shortened by Mr Fletcher so that it was less than 10cm long, and/or because the cable was damaged in the course of installation and/or service.”

It is therefore submitted by Mr Shepherd QC that the Court should find that the Pump was of unsatisfactory quality and that the lately formulated case by Wolseley and Leader should be rejected.

Mr Moody QC submits that it is for Hi-Lite to establish that the Pump was not of satisfactory quality. It is only insofar as Wolseley seeks to show that it is another cause that Wolseley has an evidential burden in showing that the cause of the damage was not that the Pump was of unsatisfactory quality. He submits that Hi-Lite have not established on the balance of probabilities that there was a manufacturing defect but rather, on the balance of probabilities, the defect was caused by damage to the cable during installation or in service.

So far as the case now being put forward by Wolseley, Mr Moody QC says that Hi-Lite has no grounds for complaint as the cause of the defect arose from the Experts’ Joint Statement which followed meetings from May to December 2010 and led to the agreements in that document. He referred to paragraph 8.15 of the Experts’ Joint Statement where they stated:

“8.15

We agree that a failure of the float switch cable after 9 weeks service was premature. However:

8.15.1

In Mr Champion’s and Mr Bullen’s opinion a fatigue failure after only 9 weeks service implies an excessive flexing frequency (number of cycles) of operation, leading to an excessive stress probably arising from a shortened length of free cable to the float switch, or damage to the float switch cable, or some combination of some or all of these factors;

8.15.2

Mr Coates is of the opinion that the premature failure is indicative of a defect in, or damage to, the cable.”

Mr Moody QC points out that in the Birmingham Proceedings Hi-Lite contended that the Pump was fit for purpose and that the fire was caused or contributed to by the negligence of Obsession.

Mr Moody QC submitted that Wolseley acted reasonably and were not notified of the Birmingham Proceedings until late on and were not brought into those proceedings. As result, no criticism can be made of Wolseley’s conduct.

Mr Dougherty submits that Hi-Lite’s pleaded case had disappeared and the sole basis for now alleging that Wolseley and, in turn, Leader were liable was a manufacturing defect in the moulding of the cable float switch which is only pleaded in the Amended Particulars of Claim as a generalised allegation of a defect of manufacture. He submits that this is not a case where an inference can be made as to the cause of the defect. The parties’ cases had narrowed down to the cause of the damage some 1cm from the strain relief on the float switch cable and whether that was caused by a defect in moulding the cable/float switch or whether it was caused during installation or in service.

Before turning to the evidence in this case, it is necessary to deal with three matters. First, there is the question of the approach to be taken by the court in determining liability when there are a number of possible causes of a defect or damage.

I was referred to a number of decisions where the courts have considered the approach to causation when there are a number of possible causes of a defect or damage.

The starting point for this issue is the decision of the House of Lords in Rhesa Shipping v Edmunds (The “Popi M”) [1985] 1 WLR 948 where a ship sank in calm weather in the Mediterranean and the issue of causation was whether the ship sank because of perils of the sea or the unseaworthiness of the ship. In the leading speech in the House of Lords, Lord Brandon of Oakbrook said this at 951 A to G:

“The question, and the sole question, which Your Lordships have to decide is whether, on the basis of those primary findings of fact, Bingham J. and the Court of Appeal were justified in drawing the inference that the ship was, on a balance of probabilities, lost by perils of the sea.

In approaching this question it is important that two matters should be borne constantly in mind. The first matter is that the burden of proving, on a balance of probabilities, that the ship was lost by perils of the sea, is and remains throughout on the shipowners. Although it is open to underwriters to suggest and seek to prove some other cause of loss, against which the ship was not insured, there is no obligation on them to do so. Moreover, if they chose to do so, there is no obligation on them to prove, even on a balance of probabilities, the truth of their alternative case.

The second matter is that it is always open to a court, even after the kind of prolonged inquiry with a mass of expert evidence which took place in this case, to conclude, at the end of the day, that the proximate cause of the ship's loss, even on a balance of probabilities, remains in doubt, with the consequence that the shipowners have failed to discharge the burden of proof which lay upon them.”

In the more recent cases of Alan Ide v ATB Sales Limited; Lexus Financial Services v Sandra Russell [2008] EWCA Civ 424, Thomas LJ giving the judgment with which the other members of the Court of Appeal agreed, considered whether the trial judges had applied a test which was impermissible after the decision in The Popi M in finding that on the balance of probabilities a defect had caused the relevant damage.

At [3] Thomas LJ summarised the effect of The Popi M as follows:

“Lord Brandon of Oakbrook, giving the only substantive opinion, described the approach of the trial judge as erroneous by reference to the inappropriateness of applying what was described as the dictum of Mr Sherlock Holmes. First, a trial judge was not compelled to choose between two theories, where the evidence was unsatisfactory; he could decide the case on the basis that the claimant had not proved his case. Secondly it was not possible to proceed on the basis of eliminating the impossible and deciding that the remaining explanation, however improbable, must be the cause, unless all the relevant facts were known; that state of affairs did not exist, as the ship had sunk in deep water. Thirdly, the concept of proof on a balance of probabilities had to be applied with common sense. It required a judge, before he found a particular event occurred, to be satisfied on the evidence that it was more likely to have occurred than not. If the judge concluded that the occurrence of an event was extremely improbable, a finding by him that it nevertheless was more likely to have occurred than not did not accord with common sense. He concluded:

“In my opinion Bingham J. adopted an erroneous approach to this case by regarding himself as compelled to choose between two theories, both of which he regarded as extremely improbable, or one of which he regarded as extremely improbable and the other of which he regarded as virtually impossible. He should have borne in mind, and considered carefully in his judgment, the third alternative which was open to him, namely, that the evidence left him in doubt as to the cause of the aperture in the ship's hull, and that, in these circumstances, the shipowners had failed to discharge the burden of proof which was on them.””

At [6] Thomas LJ said this about competing causes:

“As a matter of common sense it will usually be safe for a judge to conclude, where there are two competing theories before him neither of which is improbable, that having rejected one it is logical to accept the other as being the cause on the balance of probabilities. It was accepted in the course of argument on behalf of the appellant that, as a matter of principle, if there were only three possible causes of an event, then it was permissible for a judge to approach the matter by analysing each of those causes. If he ranked those causes in terms of probability and concluded that one was more probable than the others, then, provided those were the only three possible causes, he was entitled to conclude that the one he considered most probable, was the probable cause of the event provided it was not improbable.”

On the basis of those decisions, the approach to the question must always be that the claimant must prove the cause of the damage on the balance of probabilities. If there are a number of possible causes, the mere fact that any cause is more likely than another cause does not mean that such a cause makes it the probable cause because a number of causes can each have a low probability. Where a defendant puts forward a number of exculpatory causes then that does not shift the burden so that the defendant has to prove a cause on the balance of probabilities. Rather, the claimant has to prove that, despite the evidence put forward by the defendant, the cause of the damage on the balance of probabilities is still that put forward by the claimant.

Secondly, it is necessary to consider what impact, if any, the conduct of Wolseley has in the context of the decision in the Birmingham Proceedings. In the Birmingham Proceedings the trial judge found, as is now common ground, that the cause of the fire was the float switch cable. However, the judge also found that the cause of the fire was the shortening of the float switch cable to below the necessary 10cm by Hi-Lite so that a fatigue fracture of the cable occurred where the cable was clipped into the body of the Pump. That is not now the cause put forward by any party, including Hi-Lite, in these proceedings.

Hi-Lite has sought to assert that there was something in the conduct of Wolseley that, effectively, made it unfair for Wolseley not to be bound by the outcome of the Birmingham Proceedings. It is therefore necessary to consider the position of Wolseley in relation to those proceedings. The fire broke out on 24 January 2006 and on 12 April 2006 Wragge & Co, solicitors for Obsession, wrote to say that their investigations indicated that the fire was caused by faulty electrical wiring in the Pump and asking Wolseley to notify their insurers. There was then correspondence with insurers and on 22 June 2006 Hi-Lite’s insurers wrote to Wolseley’s insurers inviting them to take over the handling of the claim. That was rejected by Wolseley’s insurers.

Obsession commenced the Birmingham Proceedings against Hi-Lite in 2006 and, in due course, a trial of liability was fixed for 16 to 18 June 2008. On 30 May 2008, shortly before the trial, Hi-Lite’s solicitors wrote to Wolseley’s insurers to say that on the basis of their expert evidence they had decided not to commence Part 20 proceedings against Wolseley. They said that, if notwithstanding the defence, Hi-Lite was found liable, then Hi-Lite would pursue an indemnity against Wolseley for its liability to Obsession. The letter concluded by saying that it enclosed the pleadings and expert evidence and said:

“For the avoidance of doubt all of the material is disclosed to you so that you may consider Obsession’s claim against Hi-Lite and then indicate whether there are any additional matters which Hi-Lite should raise in its defence...”

No response was received from Wolseley’s insurers to that letter. Hi-Lite’s solicitors wrote again on 20 June 2008 to say that the trial had been adjourned to September 2008. On 22 December 2008, having received the judgment on liability Hi-Lite’s solicitors wrote to Wolseley’s insurers indicating that, subject to an appeal in the Birmingham Proceedings, they intended to commence proceedings against Wolseley.

In these proceedings Hi-Lite sought to contend that it would be an abuse of process for Wolseley to dispute the findings of fact made by the trial judge in the Birmingham Proceedings. That contention was rejected by Burnett J in July 2009.

This is a case where Hi-Lite decided not to bring Part 20 proceedings against Wolseley in the Birmingham Proceedings and, in such circumstances, there is no basis for criticising the conduct of Wolseley in relation to those proceedings. Wolseley’s insurers were only notified of the proceedings shortly before trial and, without being a party to those proceedings or having been the subject of any pre-action procedure, cannot be criticised for not having dealt with its defence in those proceedings. The risk of not bringing Wolseley into the Birmingham Proceedings is that there will be inconsistent findings between the two sets of proceedings. That risk arises particularly where, as here, the experts in the Birmingham Proceedings have not been instructed and three new experts, including Hi-Lite’s expert, have now provided opinions which are inconsistent with the findings in those proceedings.

Thirdly, there is the question of the way in which the case has been pleaded and the effect of the expert evidence in this case. Originally in the Defence served by Wolseley they pleaded at paragraph 43 of the Defence that:

In November 2009 Wolseley were given permission to issue and serve a third party claim against Leader. There was an adjournment of the trial originally listed for July 2010 and directions were given for an expert statement pursuant to CPR 35.12(3), followed by expert reports to be filed and served in July 2010. In the event, the experts met between May and December 2010 and produced the helpful and detailed Experts’ Joint Statement dated 2 December 2010.

In the light of the Experts’ Joint Statement Wolseley applied for and obtained permission to amend the Defence to delete the original paragraph 43 and to plead the case, which included an allegation that the cable was damaged in the course of installation and/or service.

That pleading clearly put forward the case which has now formed the basis for the contentions of Wolseley and Leader in the closing submissions. Whilst I accept that the amendment was made comparatively late, it was based on the expert evidence which had only developed in the course of the meetings leading up to the Experts’ Joint Statement. Given that the case now pursued by Hi-Lite that the damage to the float switch cable was caused in the manufacturing process of moulding the cable and the float switch, which is not clearly pleaded but which is referred to in the expert evidence, I do not consider that criticism can fairly be made by Hi-Lite of the way in which Wolseley has pleaded its case.

Having dealt with those issues, I can now turn to the central question of the cause of the fire, which depends on the cause of the damage to the cable. Whilst, as I shall consider below in relation to the impact that an RCD may have had, the experts differ on the precise way in which the fire started, they are agreed that the mechanism involved the fracturing and arcing of the conductors in the float switch cable which then caught fire. That fracturing and arcing took place at the place where there was damage caused, on Hi-Lite’s case, by the manufacturing defect and, on Wolseley and Leader’s case, by damage to the cable during installation or in service.

Given that those are the issues based on the expert evidence, I do not consider that the suggested approach by Mr Shepherd QC is appropriate in this case. He submits that, as the float switch cable to the Pump failed in nine weeks that should raise an inference the court should find that there was a failure of satisfactory quality in the cable. Whilst in simple cases where there is no developed expert case on causation that may be an appropriate starting or even ending point, in this case any such inference is weak even as a starting point and has been displaced by the detailed expert evidence on the cause of the defect in the cable.

I consider it appropriate to start by considering Hi-Lite’s case that there was a manufacturing defect in the cable caused by the moulding process when the cable and float switch were pressed, which is the only basis on which Hi-Lite seek to establish liability.

The possibility of this as a cause was first raised in the Experts’ Joint Statement when the three experts were looking at potential causes. At paragraph 8.2 of that statement the experts put forward five “theoretical possibilities”. The first was “a defect in manufacture of the float and cable assembly”. At paragraph 8.4 of the Experts’ Joint Statement the experts agreed that a manufacturing defect was a theoretical possibility. They agreed that a pre-existing cable manufacturer’s manufacturing defect was unlikely. Mr Coates view was that “a manufacturing defect cannot be entirely discounted. The point of failure was close to the moulded strain relief at the float end. This was where most damage occurred in the assembly tested by Dr Lipczynski. An error during moulding of the strain relief onto the cable could have damaged the cable at this point. Possibly by an over temperature during moulding. Such damage may not be revealed by a visual inspection or electrical testing.”

In his report, Mr Coates stated that for the float switch cable to fail after only two months would require some other factor acting at the end of the strain relief on the float switch cable. He says that “Such a factor could be mechanical damage to the cable during installation or use or a manufacturing error”. He adds “The float and its strain relief are moulded onto the cable with the application of heat and pressure. If excessive heat was applied during the moulding it could damage the cable sheath at the stress relief. The excessive heat could be an over temperature of the mould or an excessive time in the mould.”

Mr Champion’s view was that a manufacturing defect could not be entirely ruled out but, assuming that the pumps were tested electrically and mechanically prior to shipment to comply with the appropriate standard, he considered a manufacturing defect to be unlikely. Mr Bullen also considered that a latent manufacturing defect was improbable and pointed out that no defect was noted at purchase or on installation. He noted that, so far as he was aware, there was no history of float cable failures associated with moulding errors. His view was that a gross moulding error would be demonstrated by inspection and/or the 100% high voltage test on completion of both the float sub-assembly and full assembly. In his opinion a gross defect would also be visually apparent at the time of manufacture, purchase and installation.

In his report Mr Champion said that a manufacturing defect was an unlikely explanation and that such a moulding error would not affect a single float switch but a number of float switch mouldings would be affected and suffer the same defect before the temperature control or heating duration failure or mal-adjustment was corrected. He refers to Mr Panesi’s evidence that the moulding process temperature was 210°C and that there was no evidence of pump returns due to float switch failure.

As set out in his report, Mr Bullen said that, given the robustness of the cable and the relatively low melting temperature of the poly propylene float casing, it was difficult to envisage how some form of latent damage could be caused remote from the moulding. He said that when taken with the absence of discovery of the supposed defect as a result of observation or electrical testing in China or during assembly in Italy or by observation during procurement or installation or immediately thereafter, such a manufacturing defect was highly unlikely. Mr Bullen also referred to a test he had carried out on the cable by exposing it to a temperature of over 230°C for 30 minutes. He said that this produced no obvious change in condition of the cable.

In his evidence Mr Coates said that the moulding operation was carried out at 210°C which was higher than the manufacturer’s recommendation of 200°C for 5 seconds. He referred to standard EN60335-1 which warned about the possibility of damage to the cable sheath during a moulding operation. It stated “the insulation of the supply cord shall not be damaged when moulding the cord to part of the enclosure. Compliance is checked by inspection.” He said that visual inspection is probably the only thing that can be done but that it is not a “guaranteed, 100% method” for determining that there has been no change in the physical properties of the material of the cable.

In cross-examination he accepted that, based on Mr Panesi’s evidence, he could not see how such an error could occur in practice. He said that a manufacturing defect was “just a possibility” and he had no evidence to say it actually occurred. He was referred to the passage in the Experts’ Joint Statement where he had said that it was something that “cannot be entirely discounted”. He accepted that, whilst he was not willing to rule it out completely, he was not suggesting that it should be at the forefront of any explanation for what occurred. His evidence was to the effect that a manufacturing defect was no higher than a possibility that could not be entirely discounted.

Mr Bullen and Mr Champion gave evidence that essentially supported the views expressed in their reports. Mr Bullen considered that a manufacturing defect was improbable, given that the moulding process had to be very carefully controlled to make sure that the molten plastic did not damage the upper or lower shell of the switch. He confirmed that the cable tested for 30 minutes at 230°C seemed to be unaffected by that process. There was no shrinkage or significant change that he observed. Mr Champion said that, in his view, if anything had gone wrong with the heating time or level of heating there would not just be one failure but there would have been reports of many failures because there would have been a batch of failures rather than a unique single failure arising from the incident.

The evidence of how Leader manufactured and tested the float switch cable was given by Mr Panesi, the production director of Leader. His evidence which was not seriously challenged on this aspect was that he was not aware of the moulding process ever having damaged the cable nor did he see how in practice it could do so. He referred to the testing carried out both on the float switch cable assemblies in China and on the final pump assembly in Italy. He says that he does not remember Leader having to reject any of the float switches or cable after testing and has no experience of pumps being rejected or returned because of moulding defects. I accept his evidence on this.

I have come to the conclusion that it is unlikely that there was a manufacturing defect in the float switch cable caused by a manufacturing defect in the course of moulding the strain relief and float switch onto the cable. First, whilst Mr Coates appeared to provide some support for this being a cause, it was clear in his evidence before me that he was doing no more as putting it forward as a possibility and a remote possibility at that. I found the evidence and reasoning of both Mr Bullen and Mr Champion more cogent and better supported on this aspect.

Secondly there would have to be some evidence to support a case that moulding defects would occur. There is no evidence that there have been defects in the moulding process which as Mr Bullen said would be carefully controlled. Mr Panesi’s evidence and the limited “returns” evidence do not support the existence of manufacturing defects in the moulding process.

Thirdly, there would have to be some evidence that heating at higher temperatures or for longer would cause damage to the cable. The overheating of the cable, even at 230°C for 30 mins, as tested by Mr Bullen, did not have a noticeable effect on the cable. This strongly suggests that a defect would not be caused by overheating during the moulding process.

Fourthly, there would have to be a defect which in turn could cause a fatigue failure of the cable. The most that might happen would be that the cable might stiffen but this would be likely to be within the mould and not some 1cm away from the end of the strain relief. Hardening would tend to make the cable less flexible but would be unlikely to lead to local damage to the sheath or fracturing of the conductors in the in a nine week period.

Fifthly, as set out below, I find that damage having occurred either during installation or in service is a much more likely explanation for the failure.

Damage in service and damage in installation were another two of the five “theoretical possibilities” put forward by the experts at paragraph 8.2 of the Experts’ Joint Statement.

At paragraph 8.5 of the Experts’ Joint Statement the experts agreed that “damage in service...might have occurred in view of Hi-Lite’s previous evidence that a knife was found in the sump, presumably to assist with cleaning the pump or sump”.

So far as damage during installation was concerned, the experts were of the view at the time of the Experts’ Joint Statement that this depended on whether Mr Fletcher had shortened the cable to below 10cm, an issue which, as explained above, was not subsequently relied on as causative of the damage.

In his expert report Mr Coates merely noted that an additional factor such as “mechanical damage to the cable during installation or use or a manufacturing error” could have caused the failure. Mr Champion in his report continued to state his view that shortening of the cable was a cause. Mr Bullen, however, referred to the evidence of Mr Fletcher that a knife, towels and gloves were found in the sump. He said that a “nick in the outer sheath of the float cable would not need itself to penetrate the sheath and conductor insulation: it would act as a focus for enlargement in response to cyclical stress”. He expressed the view that damage during installation or cleaning was a more probable cause than moulding damage.

The experts had developed their views further by the hearing. Mr Coates identified “stress raisers” which would increase the stress and said that damage to the outer sheath or to both the outer sheath and the insulation of a conductor in the float cable could be a stress raiser. He said that the damage would have to be coincident with the point of higher stress just outside the strain relief. He accepted that there was a short length of float switch cable between the clip on the Pump body and the end of the strain relief and that it would not be surprising for damage to occur in that area if it was assumed that somebody was cleaning that piece of cable with an instrument capable of cutting the sheath. He also accepted that if there was damage to the cable the most significant damage would be any damage close to the change in section of the cable at the end of the strain relief. He accepted that it was a rational explanation for the damage but stated that it was still a coincidence.

In relation to the possibility that damage was caused by a screwdriver used by Mr Fletcher to prise the cable out of the handle, Mr Coates thought this was unlikely based on the fact that the nominal length of cable when the Pump was produced by Leader was 10cm but accepted that damage in adjusting the cable and unclipping and clipping it was something he could not entirely discount. He accepted that if there was damage at the point around the area of maximum stress a “nick” in the outer sheath could open up and develop an outside to inside fault and he accepted that over nine weeks what might have been a very small nick originally might have developed into the failure which caused the fire.

Mr Bullen said that the cable was flexible and was capable of being damaged. In relation to the possibility of damage during installation or in service he said that if insufficient care was taken, either during the installation or during cleaning, there was the possibility of damage to the cable. Mr Champion considered that the frequency of access played a part in trying to assess the probability of when the damage occurred. He said whilst it was possible for Mr Fletcher to have damaged the cable he was there for two hours making adjustments whereas if the Obsession staff carried out cleaning in the vicinity twice a day there was more opportunity to create damage to the float switch cable. As a result he thought the most probable source of the damage came from the Obsession staff rather than Hi-Lite.

One of the pieces of evidence relating to damage to the cable was the evidence of Ms Claire Robinson that she had received an electric shock.

In the Experts’ Joint Statement, the experts agree that the electric shock incident could be an indication of water penetration into the Pump, the float switch or the pump cables. Mr Bullen refers to the evidence of Ms Robinson and says at paragraph 9.3.1 that, in his opinion, a more probable interpretation of her evidence is that an incipient float switch cable failure, involving breach of the insulation, had occurred. Mr Coates at paragraph 9.3.3 said that, in his opinion, it is more likely that the shock incident was related to the cable failure. He said “Damage to the cable sheath could have led to water ingress down inside the cable sheath to the inside of the float. This would leave a conductive path from the water in the tank to the live terminals in the float.”

In his expert report at paragraphs 8.18 and 8.19 Mr Coates said that the shock felt by Ms Robinson was probably indicative of some damage to the float cable sheath in the region of the strain relief and that the shock felt by her was likely to have been less than 0.5mA. Mr Bullen reports on the joint view expressed by him and Mr Coates that the electric shock was likely to be a consequence damage to the float switch cable.

When Mr Coates came to give evidence he said that his thoughts on Ms Robinson’s electric shock incident had moved on. He said that if the shock related to the fire then the shock would be indicative of there having been damage to the cable sheath. He said that if it were damage to the cable sheath only and not a cut through the conductor insulation as well so that a live conductor was directly in contact with water, then for Ms Robinson to receive a shock there would have to be a sufficiently conductive path from the point of damage of the cable sheath through to a live part, being the termination of the go conductor in the float switch. He then explained some calculations which he had just done. He said that the conductor path was approximately 13cm and if it were assumed that Ms Robinson received a shock of 1mA, the limit of perception being 0.5mA, and if the voltage was 230 Volts, then with water having the conductivity of laundry water he calculated that a sectional area equivalent to 3mm² of water would be needed between the damage to the cable sheath and the switch. He said that he thought that there was hardly room for 3mm² of water within the cable.

Mr Coates said that he thought it unlikely that the shock was caused by a cut to the “go” conductor because that conductor had only partially failed whereas the other two conductors had fully failed. His conclusion was that this did not mean that the shock could not have been caused by damage to the sheath but it was an factor that was leading him towards the view that it was more likely that the electric shock to Ms Robinson was a coincidence and was not related to any damage that have might have led to the fire.

In his evidence in answer to questions from Mr Dougherty he said he was not saying that his original view was necessarily wrong. He thought that his original view was less likely because he had not considered what was necessary to achieve the electric shock. He said he had not previously done a calculation of the area available within the cable but had inspected it. He accepted that it was rather difficult to determine the cable area by looking at it as the area was not straightforward because the internal diameter of the sheath was not circular as the sheath material protrudes into the gaps in the trefoil between the three cords.

In relation to the evidence of the electric shock to Ms Robinson, Mr Bullen said that there had been a failure of the Pump and a few days before the incident Ms Robinson had felt an electric shock. In his view there was a likelihood of a relationship between the electric shock and the fire. He said that her evidence pointed to electricity leaking somehow from the Pump. The most obvious location was damage to the float cable which was electrifying the water. His view was that the most likely damage was mechanical damage which affected the outer sheath which had also acted as a focus for the development of fatigue. Mr Bullen considered that the shock could have come either by an electric path to the float switch, if only the sheath was damaged or it could come from an exposed conductor, if both the sheath and the insulation to that conductor were damaged. He said that, in principle, both were possible and it could be an exposed go conductor. In relation to Mr Coates’ calculations it was pointed out that, in his report at paragraph 8.19, Mr Coates had said that it was likely that the shock felt by Ms Robinson was less than 0.5mA but yet based his calculation on 1mA. Mr Bullen said that if 0.5mA were used, the area required in Mr Coates’ calculations would be reduced to 1.5mm².

Mr Champion agreed that there were the two possibilities referred to by Mr Bullen but his view was that the carbonisation process had occurred and the charred insulation was giving a conductive path to the water. He thought that if there had been mechanical damage to the “go” conductor anyone putting their hands into the tank would have received a shock.

I consider it very likely that there was some damage to the sheath of the float switch cable caused most likely by the staff of Obsession carrying out cleaning operations to the Pump. First, the cable between the handle of the Pump and the float switch is an area where hair and “gunk” which would evidently become caught. Secondly, there was clear evidence of there being hair and “gunk” in the sump which had not been caught by the filter both when the original pump was replaced by Mr Leadbrook in 2005 and before. There was then the evidence of Mr Fletcher and Mr Brown on what they saw when they installed the Pump in November 2005. General support is also given by Ms Wharton, Mr Raistrick and Ms Ridding. Thirdly, that evidence also shows that members of staff at Obsession had access to and did access the sump. The precise relationship between the filter and the tank is not wholly clear but what the evidence of the Obsession personnel shows overall is that they did gain access at regular intervals to the sump for the purpose of cleaning hair and other debris. Fourthly, the fact that, as Mr Brown said, a table knife was found in the sump together with towels and gloves show the use of a knife in the past to clean the debris and this is also cogent evidence that an instrument was needed to cut hair away. A knife or scissors in a hair salon would be the obvious instruments to do this. Fifthly, as Mr Champion observes, there would be more opportunity to damage the cable by those people in the salon having access over weeks before the fire than for Mr Fletcher who carried out the installation of the Pump over the period of a few hours.

On that basis I consider that, on the balance of probabilities, it is likely that during one of the cleaning operations, a member of staff accessed the sump to clear hair and “gunk” using a sharp instrument and in doing so cut the sheath at the vulnerable length of cable near the strain relief where that debris had become caught up.

I consider that it is unlikely that damage was caused to the cable during installation by Mr Fletcher. First, I am far from convinced that Mr Fletcher would have needed a screw driver or other instrument to lever the cable out of the clip on the Pump. As observed in court, the operation seemed possible without the need for a tool to lever it out and there appears to be no need otherwise for a tool to be used in relation to the float switch cable. Secondly, whilst Mr Fletcher’s evidence was unsatisfactory in a number of respects, he is a qualified electrician and would know the importance of not cutting the sheath of an electrical cable which was operating in water.

The existence of damage to the sheath of the cable is, in my judgment, confirmed by the evidence of the electric shock felt by Ms Robinson shortly before the fire. I found Mr Bullen’s evidence on this more convincing than the doubts which were more recently expressed by Mr Coates, who had formerly relied on that evidence. This evidence is also more consistent in my view with there being a “nick” or cut in the sheath of the float switch cable rather than some manufacturing defect caused by excessive heat being applied to the cable during the moulding process.

I can now summarise my conclusions as to the cause of the fire. On the expert evidence, the fire was caused by a fatigue failure of the float switch cable at about 1 cm from the strain relief. That fatigue failure was not caused by an excessive number of cycles of the Pump nor because the cable may have been shortened. It was caused by damage to the float switch cable which on the balance of probabilities was caused by staff at Obsession having access to the sump and cutting the cable in the vicinity of the end of the strain relief with a sharp object whilst cleaning the hair and “gunk” which accumulated in the sump and over the cable. I consider that it is unlikely that any damage was caused by Mr Fletcher when he and Mr Brown installed the Pump. It follows that Hi-Lite have not proved on the balance of probabilities that the cause of the fire was a manufacturing defect because of excessive heat being applied when the float switch was moulded onto the cable and there is no other likely cause which would give rise to Wolseley being liable under s.14 of the Sale of Goods Act 1979.

On the basis of the above finding it is not strictly necessary for me to decide whether Hi-Lite should have fitted an RCD and, if so, whether that would have prevented the fire. However, having heard full argument on this issue it is convenient to deal with it.

The starting point for an RCD being needed is the reference in the instruction manual that “Before inserting the plug, check that: (1) The electrical installation is equipped with a life-protector switch (breaker box-I_Δn=30mA)…” which is a reference to an RCD. The experts agree that a competent electrician reading those instructions would have concluded that the manufacturer recommended the installation of an RCD to protect the Pump circuit.

In the Experts’ Joint Statement Mr Coates said that there was no good practice requirement, other than the need to follow the manufacturer’s installation instructions, which required the installation of an RCD in the Pump circuit. Mr Coates disagreed with the view of Mr Champion and Mr Bullen that, in the circumstances of this case, BS7671 required an RCD to be fitted in the circuit supplying the Pump.

Clause 471-8-01 of BS7671 provides that: “For installations of increased shock risk such as those in Part 6, additional measures may be required such as [an RCD]”.

Part 6 of BS7671 describes the special situations, which include wet situations such as a bath, shower cubicle, swimming pool, garden pond, septic tank and construction sites. It does not refer to immersed sump pumps and Mr Coates did not consider that it came within the definition of “installations of increased shock risk such as those in Part 6”. Mr Champion and Mr Bullen considered that the situation of a submersible pump with a flexible cable gives rise to similar dangerous wet conditions to those in Part 6 of BS7671. As set out above Mr Champion and Mr Coates disagreed on the impact of various other provisions of BS7671 and BS EN 60335-2-41 on the need to provide an RCD.

Mr Coates’ evidence was that installing an isolating switch would have been sufficient on its own. That however was not done. He agreed that in those circumstances the Pump could not be isolated and that there was a risk of an electric shock arising. In these circumstances he had to concede that an RCD was the preferred option and should be provided if somebody was going to have access to the Pump while it was still connected to the mains.

I have come to the conclusion that in this case an RCD should properly have been installed by Mr Fletcher. The Pump was being installed in circumstances where the mains connection was being made by wiring the cable into a fused spur. There was no switch on the spur. The Pump came with a normal 13A fuse attached to the end and this had to be removed by Mr Fletcher so that it could be wired to the spur. The way in which the Pump was connected to the mains therefore meant that it could not be isolated and, in any case in order to check that the Pump was working, any person having access to the sump would have to deal with the Pump whilst it was connected to a live power supply. Mr Fletcher was evidently aware that the staff at Obsession had access to the sump as he had found the knife, towel and gloves in the tank and that the Pump would be live whilst untrained staff were dealing with it.

Whilst there may be arguments as to whether the use of the Pump as a fixed installation in a sump would generally require the installation of an RCD as being an installation of “increased shock risk” such as those in Part 6 of BS7671, the situation where untrained operatives in the hair salon were having access to clean the Pump with knives and when the Pump was connected to the mains, makes it unarguable, in my judgment, that there was an increased shock risk. Indeed, whilst Mr Coates was reluctant, at first, to accept this, he finally did accept that an RCD was the preferred option in the present circumstances where people were working with a live pump.

In addition, as Mr Coates stated, another reason for installing the RCD was that it was stated to be required in the instruction manual which stated that the installer should check that the Pump was equipped with an RCD before inserting a plug or, in this case, connecting the Pump by wiring the mains cable to the fused spur. In this case there was no reason not to follow the instruction manual and every reason to follow it. The fact that no RCD was installed with the previous pumps does not mean that an RCD should not have been fitted, as I hold it should.

As a result, an RCD should have been fitted when Mr Fletcher installed the Pump in the sump at Obsession.

This raises the difficult technical question of the effect of installing an RCD: would it have tripped and therefore prevented any fire? That depends on a view of the mechanism by which the fire started and involves the experts in having to consider what sort of electrical fault would have developed as a result of the fire and whether it was the type of electrical fault which would have caused the RCD to trip.

At paragraph 11.9.1 of the Experts’ Joint Statement, the experts agree that if an RCD had been fitted in the pump circuit the RCD would be expected to operate in the event of an earth leakage fault. In this case the evidence is that this would be either a fault between the “go” and “earth” conductors or, when the float switch was on, a fault between the “return” and “earth” conductors.

I have set out above the general sequence of the mechanism for the start of the fire which the experts agreed at paragraph 7.1 of the Experts’ Joint Statement. The question was what would happen after multiple cycles of arcing had led to carbonisation of the “return” conductor insulation from within and further heat damage and damage to the other conductors and the sheath.

Mr Coates considered that the degradation of the insulation and sheath materials due to the heat from the arcing would generate gases within the cable sheath. These gases would then pressurise the insulation and cable sheath and the arcing would also heat and soften the insulation and sheath materials. The internal pressure due to the gases combined with the softening of the insulation would then cause the insulation on the conductors to rupture and the degradation process would continue to generate gases and internal pressure which would then cause the sheath to rupture explosively. Mr Coates relied on evidence from cables subjected to rapid heating during high current, short time tests. His view was that the sheath would rupture before an electrical connection was made between two conductors within the cable and therefore before there was an earth fault to trip the RCD.

Mr Champion did not agree with that mechanism of causing the fire. He was of the opinion, taking into account the close proximity of the internal conductors in a helical formation and the thickness of the conductor insulation, that the carbonisation of the insulation to the “go”, “return” and “earth” conductors would cause that insulation to reach a high resistance value leading to earth leakage currents between the “go” and “earth” and between the “return” and “earth” conductors before the sheath ruptured and a fire could be initiated. Mr Champion stated that Mr Coates’s observations regarding an explosion arising from internal gas pressure was based on tests on cable subjected to rapid heating during high current short time tests. Mr Champion’s view was that in the “return” conductor the current was limited to a maximum running current of about 1.8A and a starting current of 7.2A for one to two seconds which did not represent the type of high current conditions which were required for Mr Coates’s mechanism to operate. As a result, Mr Champion’s opinion was that a high resistance, low energy, electrical connection between two of the internal conductors was likely to occur before penetration of the sheath.

Mr Bullen endorsed Mr Champion’s views and was of the opinion that the explosive failure process was not supported by tests done by the previous experts and did not seem to be consistent with the sequence of arcing events envisaged. Mr Bullen said that although he appreciated that an explosive failure could occur in high current faults this was a case with a relatively small current. He was of the view that any process of explosive eruption would tend to blow away the initial fuel from the fire, based on his experience of cable faults in domestic irons, with the net result that there would be a bang but no sustained fire to follow. Although a fire might have developed, in due course following such an event, the initial explosion may well not produce a flame. Mr Champion did not think that that process would liberate sufficient gases for the explosion relied on by Mr Coates and that any gases which were produced would be vented along the cable.

Whilst I accept that Mr Coates has experience of explosions arising from internal gas pressure where cables are subjected to rapid heating during high current short time tests, I consider that those tests cast little light on the way in which the cable in this case would act at the comparatively low current values. The process in this case is, in my judgment, one of gradual heating over a long period of time at those low currents. The process of arcing would be intermittent and, as I have found that there was, on the balance of probabilities, a cut or a “nick” in the cable, Mr Coates’ theory would have to depend on an explosive rupture of the insulation on the individual conductors because the outer sheath would already be ruptured. The explosion would then have to cause the fire before the power was interrupted by the RCD because of a leakage of current from the “go” or “return” to “earth” conductors.

Mr Champion’s experience of leakage paths on control circuits provides in my view a more realistic basis for assessing what would happen with low current intermittent arcing. I am persuaded that the mechanism would be the gradual breakdown and carbonisation of the insulation rather than an explosive rupture. This was supported by the evidence in the photographs of Mr Braund’s tests. In these circumstances, the carbonisation itself would provide sufficient conductive leakage paths to permit the current leakage to the earth conductors even without the insulation completely breaking down. In this way, the fire would not have developed to the stage where it was set before the leakage path would have tripped the RCD.

In summary, I therefore prefer the evidence of Mr Champion and Mr Bullen on this aspect and it follows that, with the mechanism they put forward for causing the fire, the RCD would cut off the power supply and avoid the fire, whatever the underlying cause of the fatigue failure of the float switch cable.

I now turn to consider what effect the failure to install an RCD would have had on the liability of Wolseley, on the basis that there had been any such liability. The issue raised by Wolseley is whether the failure to fit the RCD, which on the above findings would have prevented the fire, was a novus actus interveniens.

As Aikens LJ said in Chubb Fire Limited v The Vicar of Spalding [2010] EWCA Civ 981 at [63]:

“Further, in my view it also does not matter whether you regard the doctrine of “new intervening act” as part of the law of “causation” or “remoteness of damage”. The doctrine of “new intervening cause” is used by the courts as one of a number of means by which to decide whether a defendant, whose breach of a duty of care to a claimant has been established, will be responsible for certain consequences of that negligence and the damages that are claimed to flow from those consequences.”

In the context of causation in tort the law is summarised in Clerk and Lindsell (20th edition) at 2-105:

“The question of the effect of a novus actus ‘can only be answered on a consideration of all the circumstances and, in particular, the quality of that later act or event’. Four issues need to be addressed. Was the intervening conduct of the third party such as to render the original wrongdoing merely a part of the history of events? Was the third party’s conduct either deliberate or wholly unreasonable? Was the intervention foreseeable? Is the conduct of the third party wholly independent of the defendant…?”

As Gross LJ stated in Borealis AB v Geogas Trading SA [2010] EWHC 2789 at [42] to [47] the following principles can be derived from previous decisions in relation to causation:

In this case whilst there was, in a sense, an intervening act, the true position is that the failure by Hi-Lite to fit the RCD occurred before the fire but the absence of the RCD later failed to prevent the fire. There was therefore no subsequent conduct by Hi-Lite and questions of Hi-Lite’s knowledge at the time of and following any breach would not be material considerations. I would characterise the failure to fit the RCD as being a negligent failure by Hi-Lite which essentially is unreasonable conduct. In this case because the claim against Wolseley would have been for breach of contract and Hi-Lite would have been negligent as set out below, there is no ability to rely on contributory negligence.

In such circumstances, whilst the existence of the RCD would have prevented the fire, I do not consider that the failure to fit the RCD could be said to deprive any breach by Wolseley which caused the fire of its causative potency so that it would still remain an effective cause and Wolseley would not have escaped liability. At the same time it could be said that the fire was caused by Hi-Lite’s failure to install an RCD but the fact there are two causes does not, in law, deprive one cause of its causative effect.

In my judgment this case has some similarity to the situation dealt with in the case of Supershield Limited v Siemens Building Technologies FE Limited [2010] EWCA Civ 7 where a flood occurred which could have been prevented had there been clear drains and an operative building management system. As a matter of causation the flood was still the cause of the damage.

On this question, Mr Moody QC adopts the submissions made by Mr Dougherty that the loss claimed in this case is too remote because Hi-Lite is only entitled to recover loss of a type which was at the time of the conclusion of the contract within the contemplation of the parties. He submits that it is not of itself sufficient that the type or kind of loss is reasonably foreseeable but that it is necessary that the type or kind of loss was ‘not unlikely’ to result from the breach in question; or that there is a serious possibility of the loss arising or that the loss is ‘easily foreseeable’.

He refers to Victoria Laundry v Newman [1949] 2 KB 528 at 539 to 540; The Heron II [1969] 1 AC 350 at 382 and 385 to 356 (per Lord Reid); Transfield Shipping v Mercator [2009] 1 AC 61 at 68 to 71 (per Lord Hoffman), 72 to 73 (per Lord Hope), 76 to 78 (per Lord Rodger). 82 to 86 (per Lord Walker) and 90 (per Baroness Hale) and Chitty on Contracts (30th edition) 26-051 to 26-063. He submits that “a result which though foreseeable as a substantial possibility would only happen in a small minority of cases should not have been regarded as having been in their contemplation” The Heron II, per Lord Reid at 384.

Mr Dougherty submits that the breach of contract would have been that the float cable was not of satisfactory quality in that it was liable to fracture and would not have been the arcing or the fire which are the consequences of the alleged breach, not the breach itself.

In relation to the type of loss, he submits that it was not in the contemplation of any of the parties, at the time of contracts between Obsession and Hi-Lite, Hi-Lite and Wolseley or Wolseley and Leader, that a breach of contract with regard to the design of the float cable might give rise to fire damage or loss arising from fire. He submits that it is not sufficient simply to foresee some property damage, as Hi-Lite appears to contend, but it is necessary to foresee the type of property damage in the form of fire rather than flood. He refers to Victoria Laundry v Newman where a line was drawn between different types of profit and to the Wagon Mound [1961] 1 AC 388 where, in the context of remoteness in negligence, fire damage was treated as being different from general pollution damage caused by the furnace oil.

He submits that whilst it might have been within the contemplation of the parties that a failure of the float switch cable might give rise to a flood, a fire could not be said to be a “serious possibility” or “not unlikely”. He says that the float cable is submerged at the time when the greatest current passes through it when it switches on and, as Mr Coates noted, you would not expect an electrical fault in a submersible pump to lead to a fire He submits that the fact that fire damage is too remote is further evidenced by the fact that at the time of contract many millions of these types of pump had been sold without any case of fire.

Mr Shepherd QC submits that, as Mr Bullen broadly agreed, the sequence from a break in a conductor to fire was “logical”. Whilst it might be said that no-one expected the Pump to burst into flames, he submits that if there was a manufacturing defect in the cable of the Pump, an electrical fire would be likely to be caused. In any event, at the time of contracting, the parties must have contemplated that a breach of this sort could lead to Hi-Lite being liable to a customer for property damage; the type of loss suffered was within the reasonable contemplation of the parties and the precise mechanism by which that loss might arrive need not be contemplated.

The correct question, in my judgment, is whether at the time of the contracts, a fire was not unlikely to result from a manufacturing defect in the cable or that there was a serious possibility of loss by fire arising from that manufacturing defect. The important aspect is what would be expected, at the time of the contract, on the basis of a breach of contract in the form of a manufacturing defect in an electrical cable. Posed in that way, I consider that an electrical fire was not unlikely to result from a defect in an electrical cable. There is a clear logical sequence of not unlikely events from the breach to the fire: a defect in an electrical cable causing wires to break, causing arcing, sparks and heating, causing the cable to ignite. Once that type of damage is not too remote, I do not consider that liability to Hi-Lite for sums that Hi-Lite had to pay to another party for damage caused by the fire would be excluded on remoteness grounds. The type of damage is the same and payment to the person who suffered the fire at their premises to cover losses would not be too remote.

Finally, I deal on the same contingent basis with the question of apportionment raised by Wolseley and Leader. They submit that if they are liable for the fire then the fire was also caused by Hi-Lite because of the failure to install the RCD and, on that basis, there should be an apportionment of liability.

They refer to the Court of Appeal decision in Tennant Radiant Heat Ltd v Warrington Development Corp [1988] 1 EGLR 41 and say that whilst the scope of the decision was questioned obiter in Bank of Nova Scotia v Hellenic Mutual War Risks Association (Bermuda) Ltd [1990] 1 QB 818 at 904, it was applied by His Honour Judge Hicks QC in W Lamb v J Jarvis & Sons Ltd (1990) 60 Con LR 1 at [88] to [99]. They submit that, as in this case, the conduct of both parties in Lamb was held to have contributed to the loss although the conduct of both parties was not independently actionable at the suit of the other and that Judge Hicks QC held that, notwithstanding the obiter comments in Bank of Nova Scotia, he was bound by Tennant and the principle in Tennant accorded with justice.

They also refer to the judgment of Judge Hicks QC at [94] to [96] where he held that such apportionment did not depend on there being reciprocal liability between the parties. They say that whilst Hi-Lite owe no duty to Wolseley, insofar as a concurrent cause of the fire was the negligence in installing the Pump, Hi-Lite’s liability to Obsession which they seek to pass on to Wolseley is, in part, due to their own independent breach of duty to Obsession for a negligent installation, for which Wolseley are not responsible. Accordingly, although there is no reciprocal liability in this case, they submit that the causal allocation would still be based or premised on a wrong, namely Hi-Lite’s independent breach of duty to Obsession.

They submit that the effect of apportionment would be similar to the effect of the Civil Liability (Contribution) Act 1978 if Obsession had bought the Pump from Wolseley and contracted with Hi-Lite for its installation. If Hi-Lite and Wolseley were both held liable to Obsession, they would be entitled to claim contribution from each other. They draw some support for the concept of apportionment from the Scottish decision in City Inn Ltd v Shepherd Construction both in the Outer House ([2007] CSOH 190) and in the Inner House ([2010] CSIH 68) where the court apportioned delay and sums due for delay on the basis of causal potency.

They submit that Tennant is cited by the editors of Chitty on Contracts (30th edition) at 26-050, Benjamin on Sale of Goods (8th edition) at 16-051, and Keating on Construction Contracts (8th edition) at 8-022 and that the principle is sound and enables a Court to do justice between the parties in cases of concurrent causes where contributory negligence or contribution is not otherwise available.

Mr Shepherd QC submits that there is no general ability to apportion based on causation and that, in law, apportionment must be based on liability. He refers to the decision of the Court of Appeal in Loftus-Brigham v London Borough of Ealing [2003] EWCA Civ 1490 at [29]. He refers to the comment on Tennant in Chitty on Contracts (30th edition) at 26-039 where the editors say: “In one case, the Court of Appeal unusually allowed apportionment of causation between claimant and the defendant.”

He points out that Tennant was doubted in Bank of Nova Scotia v Hellenic Mutual War Risks Association (Bermuda) Ltd at p.904 and would appear to run directly contrary to the principle set out in Barclays Bank v Fairclough Building Ltd [1995] 1 QB 214. He submits that if it is ever right to apply the Tennant case, it is not right to do so here, where numerous different causes have conspired to produce a single ignition which has resulted in all of the damage.

In addition Mr Shepherd QC submits that in sale of goods cases the seller has a strict obligation to deliver satisfactory goods and Hi-Lite owed Wolseley no duty at all save to pay the price. He submits that now that shortening of the cable has been abandoned there is no contributing act that caused the fire and no blame to be apportioned and a failure to fit the RCD is not a breach of any legal obligation owed by Hi-Lite to Wolseley. There is, therefore, he submits no room for apportionment, in any event.

The starting point for an analysis of this issue is the decision of the Court of Appeal in Tennant where a landlord brought proceedings against a tenant and the tenant brought proceedings against the landlord arising from the collapse of the roof at an industrial unit. The tenant alleged that the landlord was in breach of the terms of the lease, alternatively liable in nuisance or negligence. The landlord counterclaimed for the cost of repairing the roof under the tenant’s repairing covenant under the lease. The collapse of the roof was caused by water accumulating on the roof because the 24 rainwater outlets had not been kept clear of debris. The landlord was in breach of the lease for failing to keep the outlets free from debris but the tenant was also in breach of the covenant in the lease by failing to keep the one outlet above the unit clear of debris.

The claim by the tenant against the landlord was a claim in tort but the claim by the landlord was a claim in contract and the Court of Appeal held that the Law Reform (Contributory Negligence) Act 1945 had no application. Dillon LJ said this about the position as between the landlord and the tenant:

“The problem which this court faces, on claim and counterclaim alike, is in my judgment a problem of causation of damage. On the claim, the question is how far the damage to its goods which the lessee has suffered was caused by the corporation's negligence notwithstanding the lessee's own breach of covenant. On the counterclaim, the question is how far the damage to the corporation's building which the corporation has suffered was caused by the lessee's breach of covenant, notwithstanding the corporation's own negligence. The effect is that on each question, apportionment is permissible. This is the same result as the 1945 Act would produce, but it is not reached through the Act, because the obstacle which the 1945 Act was passed to override is not there on either claim or counterclaim in the present case.”

As to the method of making the apportionment, Dillon LJ said:

“... I regard as particularly significant the number of outlets in the relevant half of the roof for which the corporation and the lessee were respectively responsible, and on a broad assessment I would apportion liability as to 90% to the corporation and as to 10% to the lessee.”

The landlord therefore had to pay 90% of the tenant's damages and the tenant had to pay 10% of the landlord's damages.

Croom-Johnson LJ added this:

“If the 1945 Act has no application, what is the position on the two claims? The evidence clearly indicates that the damage to the plaintiff's premises was attributable to two concurrent causes, both operating contemporaneously. One was the defendant's negligence and the other was the plaintiff's breach of covenant. I agree with the finding of Dillon LJ that simply as a matter of causation, based upon the amount of water which was wrongfully on the roof, the plaintiff's breach of covenant was a factor of one-tenth of that united cause, and the defendant's negligence a factor of nine-tenths. If the collapse had been caused by the combined negligence of the defendant and some third party, the plaintiff would have been entitled to recover all its damages from the defendant, leaving it to get contribution from its fellow tortfeasor: see Grant v Sun Shipping Co Ltd [1948] AC 549 at p 563 per Lord du Parcq. But that does not apply when the other cause comes from the plaintiff himself.

...

Where one is dealing with two contemporaneous causes, each springing from the breach of a legal duty but operating in unequal proportions, the solution should be to assess the recoverable damages for both on the basis of causation.”

In Bank of Nova Scotia v Hellenic Mutual War Risks Association (Bermuda) Ltd [1990] 1 QB 818 the Court of Appeal dealt briefly and obiter with questions of apportionment at 904. In doing so they accepted that no apportionment was possible in that case under the Law Reform (Contributory Negligence) Act 1945 as this was a case under the first category in Forsikringsaktieselskapet Vesta v Butcher [1986] 2 All ER 488 upheld on appeal at [1989] AC 852. May LJ giving the judgment of the court then dealt with Tennant as follows:

“Similarly, we think that the facts and circumstances of the present case are such that it can and should be easily distinguished from those in [Tennant]. We merely add respectfully our view that the scope and extent of this last mentioned case would have to be a matter of substantial argument if the principle there applied were to arise for consideration in another case.”

In Lamb Judge Hicks QC dealt with a case where water had leaked into a petrol tank at a superstore and remedial work was carried out by a sub-contractor who claimed payment. The judge found that the cause of the water leaking into the petrol tank was both defective jointing by the sub-contractor and defective groundworks by the contractor. He held that he was entitled and bound to apportion by the Court of Appeal decision in Tennant.

He referred to the decision of the Court of Appeal in Bank of Nova Scotia and said this at [96]

“That is clearly intended to be a warning against following the Tennant case uncritically but is itself plainly obiter, and it is therefore not entirely clear to me where it leaves a first-instance judge who is bound to follow Tennant unless it can be distinguished, which in my view can be done only rather artificially and unsatisfactorily, since the crucial justification there for apportionment seems to have been not reciprocal liability (which might rather have been a reason for set-off) but the fact that on both claim and counterclaim the concurrent cause arose not from the act of a third party but from that of the claimant itself.”

The right to apportion liability has now been further considered by the Court of Appeal in Barclays Bank Plc v Fairclough Building Ltd [1995] QB 214. In that case the plaintiff succeeded in proceedings against a contractor for breach of strict liability provisions of a contact in failing to carry out roof cleaning operations properly. The judge also held that the plaintiff was negligent in failing to supervise the contractor’s work and apportioned liability under the Law Reform (Contributory Negligence) Act 1945. The Court of Appeal held that he was not entitled to do so. Beldam LJ in giving a judgment with which the other members of the court agreed referred to December 1993 Law Commission Report: Contributory Negligence as a Defence in Contract (Law Com No 219). He then referred to the classification of contractual duties, of which category (i) is where a party’s liability arises from breach of a contractual provision which does not depend on a failure to take reasonable care. He said this at 229 C to D:

“On the other hand, in category (i) cases there is no decision in which contributory negligence has been held to be a partial defence. There are powerful dicta to the effect that it cannot be: see the judgment of the court in Tennant Radiant Heat Ltd. v. Warrington Development Corporation [1988] 1 E.G.L.R. 41 , in Bank of Nova Scotia v. Hellenic Mutual War Risks Association (Bermuda) Ltd. [1990] 1 Q.B. 818 , 904, and the observations of Nolan L.J. in Schering Agrochemicals Ltd. v. Resibel N.V. S.A. (unreported), 26 November 1992; Court of Appeal (Civil Division) Transcript No. 1298 of 1992, noted in (1993) 109 L.Q.R. 175 , 177.

The defendant's argument that, because the plaintiff owed duties to its employees it was therefore under a duty in its own interest to see that the defendant fulfilled its obligations under the contract, is inconsistent with many cases in which it has been held that employers and others liable to third parties for failure of plant or equipment are entitled to rely on warranties given by their suppliers: see for example Mowbray v. Merryweather [1895] 2 Q.B. 640 ; Sims v. Foster Wheeler Ltd. [1966] 1 W.L.R. 769 , 777 and Lambert v. Lewis [1982] A.C. 225

...

That a contracting party is entitled to rely on the other party to a contract to carry out his undertaking and to act carefully in doing so was emphasised by Devlin J. in Compania Naviera Maropan S.A. v. Bowaters Lloyd Pulp and Paper Mills Ltd. [1955] 2 Q.B. 68 , 77, where he said:

“Indeed, I think business, whether maritime or otherwise, might be gravely impeded if the ordinary principle were not allowed to operate freely - and by the ordinary principle I mean that, generally speaking, a man is entitled to act in the faith that the other party to a contract is carrying out his part of it properly. It does not lie in the mouth of the promisor to say that a promisee has no right to assume that a promise has been faithfully carried out and should make his own inquiries to see whether it is or not. If everything done under contract has to be scrutinised and tested by the other party before he can safely act upon it, many transactions might be seriously held up...””

At 230 Beldam LJ concluded as follows:

“In my judgment therefore in the present state of the law contributory negligence is not a defence to a claim for damages founded on breach of a strict contractual obligation. I do not believe the wording of the Law Reform (Contributory Negligence) Act 1945 can reasonably sustain an argument to the contrary. Even if it did, in the present case the nature of the contract and the obligation undertaken by the skilled contractor did not impose on the plaintiff any duty in its own interest to prevent the defendant from committing the breaches of contract. To hold otherwise would, I consider, be equivalent to implying into the contract an obligation on the part of the plaintiff inconsistent with the express terms agreed by the parties.”

In the light of the decision of the Court of Appeal in Barclays Bank v Fairclough I do not consider that in the present case if Wolseley were to be liable under section 14(2) of the Sale of Goods Act 1979 that Wolseley could then seek an apportionment on the basis that Hi-Lite were negligent in failing to provide an RCD as part of the installation of the Pump. The provisions of section 14(2) are strict liability provisions and, as set out in Barclays Bank v Fairclough a party liable in relation to such a contractual duty cannot reduce that liability by an apportionment to take account of the negligence of the other party.

In Lamb the sub-contractor had carried out remedial work which had been necessary because of the leaks. The sub-contractor sought the cost of that work from the contractor. The sub-contractor was however liable for the defective pipework which had been an effective cause of the leaks but the contractor’s groundworks had also caused the leaks. The sub-contractor was liable on the basis of strict liability in contract to the contractor for the defective pipework but the contractor was not liable to the sub-contractor for the defective groundworks. On the judge’s findings, the remedial works had been necessary because of the leaks caused both by the sub-contractor and the contractor.

In my judgment, if there was to be any reduction in the sums for which the sub-contractor was liable in Lamb, that would have to be on the basis that the groundworks were carried out negligently by the contractor. However, as made clear in Barclays Bank v Fairclough that provides no basis for an apportionment in relation to the strict contractual liability of the sub-contractor. Further, as Judge Hicks QC said at [55] this was a case where there were difficulties in apportioning damages under the Civil Liability (Contribution) Act 1978 on the basis that the two parties were not both liable to a third party in respect for the same damage.

As was observed in Law Commission Report No 219 at paragraph 3.13, the decision of the Court of Appeal in Tennant has been the subject of criticism on the basis that there is no good reason to limit the principle to cases where the contributory fault amounts to breach of duty to the other party and, on this basis and in the light of what was said obiter in the Bank of Nova Scotia case it was said in the Report that it was likely that Tennant would be interpreted restrictively and limited to its own facts.

In my judgment, there are no grounds for apportionment in this case based on the approach in Tennant. I consider that the decision in Tennant is one on the particular facts of that case which are easily distinguishable from the facts of this case which come within the general principle acknowledged in Tennant that contributory negligence is not a defence to strict contractual liability. I do not consider that, with the greatest of respect, the apparent application and extension of the approach in Tennant to the facts in Lamb was justified or can be justified in this case. Absent apportionment under the Law Reform (Contributory Negligence) Act 1945 or the Civil Liability (Contribution) Act 1978, I do not consider that there is any other general ability to apportion damages between two parties. To the extent that the Scottish decision in City Inn v Shepherd decides otherwise or supports an opposite conclusion, I do not consider that it represents the current position in English law.

For the reasons set out above, the fire at the Obsession salon was caused by a fatigue failure of the float switch cable which, on the balance of probabilities, had been damaged by the staff of Obsession who accessed and used sharp tools to clean hair and other debris from the Pump located in the sump. It follows that Wolseley have no liability to Hi-Lite for that damage or for the subsequent fire and, in turn, Leader has no liability to Wolseley based on those facts.

In any event, Hi-Lite failed to carry out the installation properly because they failed to provide an RCD when installing the Pump. If an RCD had been fitted then, on the balance of probabilities, the fire would not have started before the RCD had tripped because of leakage from the “go” or “return” conductors and the “earth” conductor.

However, if Wolseley had been liable as alleged by Hi-Lite for a manufacturing defect in the Pump which caused the fire then Wolseley’s liability to Hi-Lite for damages would not be reduced or avoided by arguments based on the failure by Hi-Lite to install an RCD. The causative effect of Wolseley’s breach of section 14(2) of the Sale of Goods Act 1979 would not have been affected by the failure to fit the RCD. Nor do I consider that the fire or the damages which Hi-Lite had to pay Obsession would have been too remote. Further I consider that any negligence by Hi-Lite would not give rise to any apportionment as between Wolseley and Hi-Lite.

In those circumstances, Hi-Lite’s claim against Wolseley fails and, in turn, Wolseley’s claim against Leader fails.