A block of flats in the Greenwich Millennium Village is at the centre of a bitter dispute about noise transmission. Although the building originally passed an acoustic test, the residents claim the problem is so bad they cannot sleep.

Kirk Smith, Alan Jones and Bruce Haswell (left to right) took on Greenwich Millennium Village Ltd in a battle to sort out noise problems in their homes
Kirk Smith, Alan Jones and Bruce Haswell (left to right) took on Greenwich Millennium Village Ltd in a battle to sort out noise problems in their homes

Greenwich Millennium Village seems peaceful enough. The streets are car-free and many properties are situated around an ecological park where the only sound is twittering birds.

But inside one of the blocks of flats a row is brewing that threatens to tarnish John Prescott’s blueprint for the redevelopment of the Thames Gateway.

The row centres on noise transmission between apartments in a block of flats in Kilby Court. This has blighted the lives of the residents and turned into a long-running battle between the residents on one side, and the developer, Greenwich Millennium Village Ltd, housebuilder Countryside Properties and warranty provider NHBC on the other.

From the residents’ perspective it is a sorry example of the housebuilding industry fobbing off hapless purchasers with excuses when problems arise – this is particularly pertinent here, as GMVL has publicly set itself a target of zero defects with the homes it builds. From an industry position it has implications for the acoustic regulations, which were only recently revised. It is also a cautionary tale for anyone trying out new construction methods, as the block has been constructed using lightweight steel framing. Finally, it is a story about the determination of the people of Kilby Court to have their voices heard.

“I took up the gauntlet as we experienced difficulties the day we moved in,” says Alan Jones, a psychotherapist who took possession in April 2004. “I immediately knew there was something massively wrong with the building. The problem is impact sound – I can hear any movement upstairs amplified as in a drum, and the frame creaks as well. I can even hear the neighbours above the people above me if they are out. If our neighbours have people around for a dinner party we go out – I’d rather sleep on a friend’s floor for the night.”

Jones’ neighbour, Bruce Haswell, who happens to be a quantity surveyor, suffers from the same problems. “The neighbour upstairs comes home pretty late. I can’t rest until she’s come home, which could be two in the morning.” Another occupant, Kirk Smith, also says he does not go to bed until his neighbour gets in. “I wake up if my neighbour gets up or if they turn over in bed, because of the creaking.”

Nick Raynsford, the MP for Greenwich, became involved in the case later on, and visited Kilby Court to experience the problems for himself. “I was concerned the noise penetration was far greater than I’d have expected from a building constructed to current standards,” he said.

The residents began complaining to developer GMVL. “They said it would improve when I moved my furniture and soft furnishings in,” says Jones. “When I went to customer care they treated it as an isolated case,” says Haswell. Gradually the neighbours began talking to each other about the problem, culminating in the residents of the 16 flats writing a letter of complaint to GMVL. The developer then agreed to carry out an acoustic test.

Unfortunately for the residents, the test indicated that the flat used passed the requirements of the 1992 edition of Part E of the Building Regulations. “It passed by 1 dB mean average above the minimum requirement,” says Jones. “Countryside always kept this confidential and said it was a good pass.” A letter from Alan Cherry, Countryside Properties chairman, written on behalf of GMVL to Jones and his partner Tia Comfort bears the “good pass” statement out.

We have made it very clear we are facing up to this problem. Whatever the problems are they will be dealt with to everybody’s satisfaction

Alan Cherry, Countryside Properties

It says, “Countryside has complied with its obligations to you and the other residents who have acquired these apartments. I have to inform you the company is not prepared to take the matter further.”

Undeterred, Jones, Haswell and Smith formed a group called Core Four Collective to represent the residents of the block and keep up the pressure. They approached warranty provider NHBC to take up their case. “NHBC sent their claims inspector round who said the problem was quite bad,” says Haswell. Because of this, NHBC agreed to an intrusive investigation which took place on 8 September 2004. “He looked at the floors and the ceiling in one flat by sawing a section out and pronounced it was OK,” says Haswell. “We had to keep drawing the inspector’s attention to things. We pointed out the insulation was only 20 mm thick but it was 100 mm thick in the drawing,” says Smith. “He said the building could be made of straw and mud – if it passed the regulations it was okay.” NHBC later admitted that the insulation was too thin in its own report.

Things came to a head later that day. NHBC was there with GMVL on one side and the residents on the other. “There was a big showdown,” says Jones. “We had to stand shoulder to shoulder. It was like that scene in the film Zulu where they stand shoulder to shoulder. It was quite traumatic.” Under pressure Paul Goring, the NHBC inspector, admitted that the acoustic test only passed 1992 Part E by 1 dB, and had been carried out on a floor on which a layer of laminate had been added – Part E tests should be carried out in the “as built” condition. “At that point Goring said we could have another sound test,” says Jones.

Sound specialist RBA Acoustics was engaged by the residents with the agreement from GMVL to carry out the new test on 12 October 2004. This was carried out under strictly controlled conditions – all furnishings and floor coverings were removed from three test areas. The average of the three tests was 62 dB sound transmission through the floor whereas 1992 Part E requires 61 dB – which means the three flats tested fail by 1 dB. Significantly the current, 2003 edition of Part E does not average the results, which means two flats would have passed 2003 Part E (see “How Part E has failed to keep noisy neighbours at bay”, below).

Sean Smith, building acoustics research fellow at Napier University’s Building Performance Centre, says revisions to Part E have not addressed the issue of impact sound. “If you compare our regulations to other parts of Europe its very poor,” he says. “The problem is most acousticians know that values higher than 60 dB are likely to lead to complaints. It needs to be 57 dB.”

Six months later, the residents in Kilby Court are still suffering. One flat has been taken over by GMVL to try and discover what is causing the problem. “It has taken some time because of the complexity of the problem and the nature of the investigations,” says Alan Cherry, chairman of GMVL. “It has led to a series of findings and recommendations. We have made it very clear that we are facing up to this problem. Whatever the problems are, they will be dealt with to everybody’s satisfaction.”

Cherry adds this problem is part of trying out innovative methods of construction. “This is one area where the industry has to be very mindful there is a degree of trial and error, and it is very easy to be discouraged from getting into this process of innovation,” he says. “We were up to 70% with timber frame in the 1980s when

we had to switch it off virtually overnight because of the World in Action TV programme. There were problems with timber frame then but today there aren’t because we have got much better at it.”

The residents are hoping the solution will go beyond the minimum standards set out in Part E. “One of the things that we have had trouble with is the emphasis on the numeric performance of the building,” says Jones. “The 1992 Part E says a building will be considered to have met the requirement if the level of normal noise from adjacent dwellings is such that it allows people to sleep, rest and engage in normal domestic activity. This doesn’t do that.” Jones continues, “The sales literature led us to believe this is a building for the 21st century. We expect to receive that when this ends.”

What might be wrong with the floor

The flats are built using Powerwall System’s lightweight steel stick framing system. Lightweight steel joists were used for the floor structure and the floor design is based on a generic type set out in the 1992 edition of Part E, a type 3A platform floor made using timber joists. Because the floor structure is made from metal rather than timber it has been subject to laboratory testing to prove its performance and to satisfy building control that it would meet the requirements of Part E. At GMV, approved inspector NHBC was used rather than local authority building control. They say it was found to comply with the requirements of Part E based on tests and previous experience of similar types of construction.

Acoustic laboratory test vs real life
The proposed floor design was tested by acoustic consultant AIRO, and achieved a respectable figure of 50 dB, well in excess of the requirements of Part E for separating floors. But there are several important factors that could explain the gap between the test result and the floor’s performance at GMV.

Sean Smith, a building acoustics research fellow at Napier University’s Building Performance Centre, cautions against using lab test results to define real life performance. “A lab test report does not represent how it is in the field. If we use a report we typically add 5 dB because the tests are not carried out with the full wall flanking and the support structures in place,” he says. However this is not enough to explain the difference between the 50 dB lab result and the 62 dB obtained at GMV.

The second reason for the floor’s poor performance could be because it differs from the lab test detail in one important respect. The test detail contains an element called resilient bars, which are missing from the Proctor and Matthews drawing used for the floor construction at Kilby Court. The reason they are missing is unclear. Resilient bars are steel bars fixed to the underside of the joists, running perpendicular to the top side of the ceiling boards. They are intended to reduce the area of ceiling plasterboard in contact with the joists. Mike Langley, technical director of acoustic consultancy Sound Research Laboratories, says resilient bars would reduce sound transmission by 2-3 dB. “Subjectively you would barely notice the difference,” he says.

Moving floors
For unknown reasons, the floors at Kilby Court suffer from deflection as people walk over them. The NHBC acknowledge this is a problem with Alan Jones’ flat – Jones says it is even worse higher up the block. “We were having dinner upstairs and as our friend walked across the floor to the kitchen the flower heads were bobbing up and down in the vase on the table,” he says. Smith says people are affected by deflection. “If a floor deflects by a few millimetres and the room downstairs is enclosed it compresses the air,” he explains. “That pressure change is picked up by the ear drum which is very sensitive and you can feel where the person is upstairs.” Significantly, this phenomenon is not simulated in acoustic tests.

Langley thinks there may be other reasons why the floors at Kilby Court perform so poorly. He says the fact the 100 mm insulation quilt was only 20-30 mm thick in places would make a difference of only 1 dB at the most. He acknowledges deflection could be making things worse and adds that the acoustic insulating layer could have been bridged in some way. “Someone could have put a nail through it,” he says.

No quick fix
Langley does not believe there are fundamental problems with lightweight steel frame. “We recently tested some steel floors that gave very good results which were more akin to the lab tests done by AIRO,” he says. It is impossible to extrapolate this to GMV because the specific differences such as the length of the floor span, which could have a significant effect, may be different, and the definitive cause of the problems at Kilby Court has not yet been found.

However Napier University and Sound Research Laboratories think fixing the problem will by tricky. “GMV should be aware that the complaints won’t stop unless they do something radical,” says Langley. He says a resilient layer on top of the floor won’t make much difference. “This damps out the high frequencies but not the low frequencies,” he says. “The results would suggest the high frequencies have been dealt with.” He says a floating floor on top of the existing floor or a secondary suspended ceiling or both would be needed.

Smith’s colleague, professor Robin Mackenzie, sounds a cautionary note about the specific floor design at Kilby Court. “Platform floors have been largely discontinued in Scotland because they have not performed well,” he says. “I am not surprised there is a problem given the type of construction and the test results. We were continually getting complaints typically at the 61 dB mark. We recommended to our clients that a different form of construction be used.”

How Part E has failed to keep noisy neighbours at bay

The ODPM could learn some lessons from the noise problems suffered by the residents at Kilby Court. It was built to the 1992 version of Part E and failed its requirements in the second set of tests as the mean of the three flats tested was 62 dB. 1992 Part E stipulates a mean of 61 dB. Significantly, two of the three flats tested would have passed the 2003 edition of Part E, which was meant to sort out the problem of noisy neighbours. With the 2003 edition, there is a straightforward 62 dB pass or fail for each flat – the results cannot be averaged between several apartments. Given these facts there are much wider implications not just for the ODPM but housebuilders and homebuyers too as 2003 Part E wouldn't have helped much in this case.

Professor Robin Mackenzie, the director of Napier University’s Building Performance Centre, says acoustic standards have actually got worse. “Airborne noise standards have improved but impact standards have worsened,” he says, adding the 1985 standard was 57 dB rather than today’s 62 dB. “We now have a situation in England and Wales where the standards are out of kilter. A marginal case of airborne noise will generally be accepted by neighbours whereas impact noise will generate complaints.”

Mackenzie says the weakening of impact noise requirements was the result of a BRE study that discovered construction details approved in Part E were not performing as they should. “In the mid-1980s research done by BRE showed 63% of deemed-to-satisfy construction types failed so the standard was lowered from 57 dB to 61 dB.” The 2003 standard for Part E is 62 dB but is potentially more onerous for housebuilders as it requires them to carry out post-completion testing. However housebuilders can escape post-completion testing if they use Robust Standard Details – approved details that have been proven to perform well.

Napier University worked with the House Builders Federation on the RSD project. Interestingly, the details are designed to perform to 57 dB as this builds in a margin of error. It also could mean that homes built using RSDs enjoy a decent level of impact noise resistance. Unfortunately this does not help anyone using innovative forms of construction as RSDs only apply to tried-and-tested methods.