Jeremy Harris

Gaps around boxes are a no-no. The boxes should be tight to the plasterboard, or ideally project into it slightly, so there is no gap. We packed all ours out so they projected a few mm into the plasterboard layer, to both make sure there were no gaps and to make it easier to mark the box cut outs on the boards (just put the board in place and bang it over the box, and it leaves an imprint for the cut out) The workmanship looks bloody awful, and I think there is a strong probability that you will get cracks in all those areas with big voids.

Interesting article on the BBC: http://www.bbc.com/news/uk-england-berkshire-41466281 that suggests that water from either vapour/steam released by the charring PIR, or from the firefighters, could have exacerbated this fire. This fits reasonably well with some of the video evidence, that shows the fire burning fiercely just above the range of the firefighters water jets, and seeming to spread vertically above that area pretty vigorously. Perhaps steam given off from below rose up in the updraft through the ventilation cavity increasing the intensity of the fire by reacting with the cladding and foil on the PIR? Makes sense, but is far from proof that this may have happened.

Another vote for the stuff from C&A Building Plastics. I too bought a roll to seal the ends of some twinwall polycarbonate sheet around 15 years ago, and it's still doing the job OK.

I think that's probably down to the reputation Sika has, mainly from its widespread use in the boat industry. Sikaflex has been the "go to" sealant for everything on boats for decades, and practically everything associated with boats seem to attract a fair mark up. I always used to buy my boat stuff from the grotty fisherman's chandlers in Falmouth and Penryn, rather than in the upmarket "yachtie" places, where the same product would be half as much again.

In our test report there are 10 measurements during the ramp up to 50 Pa for each of the pressurise and depressurisation tests, with the results for both plotted on a graph. The leakage with pressure differential seems pretty non-linear, with zero flow rate being measured until the pressure differential reaches 8 Pa during the depressurisation test cycle and 9 Pa during the pressurisation test cycle. The results show curves that are of opposite characteristics, and indicate that the the rate of change of air leakage increases with increasing pressure during the pressurisation test, and decreases with increasing differential pressure during the depressurisation test. I'm guessing that's because we have outward opening windows that may well seal slightly tighter as the internal pressure decreases.

As @Nickfromwales says, there are several different Sikaflex compounds. Originally, Sikaflex was only a polyurethane, moisture curing, sealant/adhesive. This was the stuff that earned Sikaflex a good reputation as a sealant that would tick like sh*t to a blanket, on damned near anything. The two downsides with it were the slight smell and the fact that it was, if anything, too good an adhesive (damned near impossible to get off when cured). Since then, several manufacturers have come up with other formulations, from polysulphide sealants (excellent stuff for sealing up the joints inside aircraft wing fuel tanks, but smells awful and I'd avoid it like the plague) to a wide range of modified silicone polymers and polyurethanes. CT1, Evo Stik "Sticks like Sh*t" sealant and Unibond F|T101, are examples of MSP adhesive/sealants that are near-identical. Sikaflex EBT is a polyurethane that seems very similar to their old 223 industrial grade stuff, marketed as a building adhesive/sealant. In general, there's not much to choose between any of the MSP or polyurethane products. All are significantly better in just about every respect to standard moisture curing acetoxy silicone sealants, although some of the better neutral cure silicone sealants perform almost as well as the MSP stuff, but there's not a big enough price differential to make it worth using them, IMHO.

The basin in our old house bathroom has had a Sikaflexed in place waste for around 7 or 8 years now. I ended up resorting to using Sikaflex because the rubber seal just wouldn't bed down properly and seal. The basin "counter bore" where the waste fitting goes was too shallow (Victoria Plumb crap again......). In my case, I applied a bead of Sikaflex around the thing, did as you did and stopped the waste rotating whilst I lightly tightened the nut, wiped it all clean with some IPA and left the Sikaflex to go off over night. The nest day I tightened the nut up properly and fitted the trap. It's never given any trouble, and stays cleaner than the other wastes, as there isn't a small dip around the edge of the waste to get filled with discoloured lime scale. I've never seen any mould in there at all, and suspect the regular washing of the surface with water going down the waste keeps it mould free.

As I wrote above, with the link to a recommended compliance document for EN13829 in the ATTMA guidance (which is not the only way of showing compliance with the standard), but really only where there is justification for this, i.e. a case can be made by the tester that all the seals are unidirectional. Given that it takes 5 minutes to switch from depressurisation to pressurisation and do another test, there seems very little reason not to do both. It takes several times longer than that to fit the blower and tape it in place, then remove it again afterwards.

Just to add, doing both a pressurisation and depressurisation test is at the discretion of the tester, but they must take into account the specific construction details of the house. The key points here are the direction in which the doors and windows operate. If a house has openings where every door and window operates inwards, then the tester could decide that a depressurisation test would give the worst case, and so use their discretion to just do this. My personal view is that this is bad practice, looking at our test report, the depressurisation test was undertaken at 08:45 and the pressurisation test was undertaken at 08:50, so there was only 5 minutes to switch from one test to the other, and no real excuse for an air test company not to do this at all. Most houses will have a combination of inward and outward opening doors and windows, so would need both a depressurisation and pressurisation test done, anyway, as I would question whether there would be adequate justification for just doing a single test under such circumstances. The UK accepted compliance test method developed by the ATTMA does vary slightly in this respect from the way other countries interpret the requirements of EN 13829, where bi-directional testing is the normal procedure. For those that want the detail on the current recommended test methodology used in the UK, then here is the current (2016) document: https://www.attma.org/wp-content/uploads/2016/09/ATTMA-TSL1-Issue-3-Rev-0-2016.09.09.pdf

Every height on our main drawings is referenced to ordnance datum. The standard abbreviation when this is the case is to put AOD after the height, for Above Ordnance Datum. If drawings don't have this, then there should be a note somewhere saying where the datum is located. For example, the working copies of our ground works drawings were referenced to a datum nail in the lane that was at 81.48m AOD. It was easy to then just use relative heights to this point as far as the guys on site were concerned, because they could just do a very quick check from that datum nail, without needing to subtract the height above ordnance datum.

In this case is seems as if the air test company wasn't exactly on the ball, and that may have been the real reason for the frame coming loose. 50 Pa is a very low pressure, and I strongly suspect they made an error, didn't fit the restrictor rings to the fan before turning it on and accidentally depressurised the house to WAY below 50 Pa, hence the problems with the frame coming out, their software being out of range etc. This also tallies with them not understanding the proper air test procedure and the requirement to take both readings and use the average as the air permeability figure.

My understanding is that local authorities only have the power to create a byelaw banning certain forms of transport on a very limited set of public highways, principally footpaths, and some bridleways. Cycles cannot be banned from a normal vehicular right of way by a local authority, only by national law, such as the ban on cycles on motorways. Local authorities can place advisory safety notices up if they wish, but these cannot be prohibitions. If the road is a vehicular right of way, then cycles can use it. Cycles have the same rights of way as any other road user, in fact they have greater rights than some, such as they are not required to comply with any speed limits.

The test procedure requires them to do both a positive and negative test, then average the two results to get the permeability, so both tests need to be done. No need to go outside either, they just switch the fan. There's often quite a difference between the positive and negative values, because some door and window seals will tighten on a positive test, some will loosen, and vice versa, which is why the procedure requires both tests to be done. Looking at our blower test report, there was a 15% difference between the pressurisation test and the depressurisation test, with the latter giving the lower permeability figure (we have outward opening windows, inward opening doors).

Which reminds me, one of the most useful sources of technical papers, going right back to the early 1900's, is the NASA/NACA technical reports library: https://ntrs.nasa.gov/search.jsp It's publicly accessible to anyone, and it makes a fascinating research source. Some of the very early NACA papers, mainly from Langley, but a fair few copied and translated German papers captured after the war, and some of the old RAE papers from here, are in there, with some of the really fundamental principles that we still use today.

Our very worst case, when it's -10 deg C outside, is a total heat loss of just under 1700 W. The heating requirement is lower than this, because of incidental and occupant heat input, so the practical worst case heating requirement is probably around 1300 to 1400 W. That equates to a much higher 18 W/m² from the UFH, but it's a situation that so rarely occurs, and lasts for such a short period of time, that it has little impact on the mean heating requirement for the coldest month of the year. In reality the temperature rarely drops below zero here (the January mean heating requirement is only around 260 W, after accounting for winter incidental heat gains) and if the outside temperature does drop below zero overnight that's usually during clear weather, so solar gain tends to be reasonable during the day, and quickly warms up our very sheltered spot, cut back into the hillside. I've found that our outside air temperature tends to be between 2 and 3 deg C warmer than the Met Office historical mean monthly temperatures for our area, and, coupled with the much lower than typical wind speeds we get (so much lower accelerated convective heat loss) means our heating requirement seems to be much lower than I expected. This very low heating requirement, coupled with the 7 kW max output ASHP (only chosen because it was readily available at the right price - 4 kW models seem rare and expensive), is a part of the reason I had control issues with the heating system initially. The ASHP has a pretty good programmable weather compensation curve, that I did play around with a fair bit, but after a lot of testing and measurements I chose to set it to deliver a flat 40 deg C, no matter what the outside temperature, as it turned out that this setting gave the lowest overall electricity consumption. In practice, the ASHP never runs at full power unless the buffer tank has been cooled down well below the 35 deg C that the tank stat is set to. Most of the time the ASHP is barely ticking over, and it only seems to come on once a day, for an hour or two, then stays off. The house seems to stay between 20.5 deg C and 21.5 deg C all the time, with around 0.5 deg C of overshoot from the set point, despite the 0.1 deg hysteresis room thermostat. In practice this range of temperature variation feels just about OK. I'd like to have slightly better control, but the sensitivity of the system, plus the relatively large effects from cooking, showering, having visitors etc is such that I doubt this is practical.

That's a higher heating system demand than our house, by a fair bit. The January UFH mean heating requirement for us is around 3.5 W/m². That makes a big difference, as it means that most of the heat loss in our house is being made up for from incidental heat gains from the occupants, appliances, lighting, etc, and the heating system only needs to contribute a relatively small proportion of the total heating requirement. That changes the control system dynamic quite significantly, and really just amplifies the point that @Nickfromwales made earlier that having no TMV isn't a "one size fits all" solution.

There's no such thing as "thermal mass", because mass isn't related to thermal stability at all in this context; what does impact on the house thermal time constant is the heat capacity and thermal conductivity of the internal surfaces and the house contents, as mentioned above. The practical problem I've found, after three years of experiments, is that, because the heat capacity of the concrete slab is relatively low (the mass heat capacity of water is nearly 5 times that of concrete, for example) and the thermal conductivity of both water and concrete are broadly similar, heat transfer to the slab is relatively rapid when the UFH is running. However, if the UFH is turned off, because the room thermostat has stopped calling for heat, the slab continues to provide heating for a long time, depending on the temperature differential between the air immediately above the slab and the temperature of the slab. What's more, the heat transfer rate during this overshoot period is neither linear nor even, some parts of the slab will transfer heat at a greater rate than others, because of the variation in heat transfer rate caused by local effects (air flow, objects on top of the slab, floor coverings with both a differing surface emissivity and thermal conductivity, etc). I got around the latter problem by running the UFH circulating pump even when the TMV or the valve controlling flow to the UFH are turned off. Circulating water around the loops all the time acts to even out the slab temperature, lowering the temperature of the warmer spots and raising the temperature of the cooler spots. This is extremely effective at controlling solar gain from sunshine that warms the floor, as it lowers the local floor temperature and significantly reduces the rate of heat transfer to the air. This is particularly effective because the rate of heat transfer with temperature differential is non-linear, for a floor surface with a typical emissivity and surface convection rate it approximates to about 8.92 * Δt 1.1 W/m².

See here:

It doesn't work like that. All that happens is that the ASHP stays on for longer on a cold day - I don't fiddle around changing the UFH flow temp. I did spend a lot of time trying to make weather compensation work with a low heat demand house, but failed miserably. I have around 30 variations of software I wrote to try and do closed-loop control of the slab and house temperature, before I gave up. The problem is that the uncontrollable variables are of a similar magnitude to the controllable variable, and can only be indirectly sensed from the room temperature. For example, say the system is stable, with the slab delivering around 300 W of heat to the house (a pretty typical cool, but not cold, winter heating level). We get two visitors arrive, and the house suddenly gets a heat input jump of around 160 to 200 W. The room thermostat detects this when the room temperature rises 0.1 deg C above the set point, but the slab, and the water in the UFH loops, is still sitting at 24 deg C. It therefore carries on putting out heat for ages, even with the ASHP off, until there is a new equilibrium between the heat output from the very slightly cooler slab, the heat input from the occupants and the heat loss (which won't have changed if the outside air temperature is the same). The result is that there is a temperature overshoot, and it's quite easy for this to get to 23 deg C or so, which is uncomfortably warm. This control challenge is, in my view, one of the biggest things to overcome with low energy house design. The easiest way to limit it's impact is to make sure that the unheated inner surfaces of the house are both relatively thermally conductive and have a high heat capacity, as then they will tend to absorb some heat from the overshoot, so reducing it's magnitude. This works reasonably well, but isn't great, as air is not a particularly good conductor of heat, and the rate of heat transfer to the walls and ceilings is not fast.

My experience is that our Carrier ASHP just cannot hold the flow temperature anywhere near a constant temperature, even with a 70 litre buffer. 30 deg C is way too hot for the UFH, even in really cold weather I can't let the UFH flow exceed about 25 deg C, or else the house temperature will over-shoot for the next couple of hours or more, often by up to a couple of deg C, which is pretty uncomfortable. I had to play around to get the TMV to regulate tightly at around 23 to 24 deg C, and even at 24 deg C UFH flow we get a bit of overshoot on the house set temperature, maybe as much as a degree. If I set the ASHP flow temp down to, say, 25 deg C, then what happens is that it drops to around 22 deg C, then rises to around 28 deg C or so, then drops back down to the lower temperature again, as it cycles on and off. A 70 litre buffer, plus the concrete slab, just cannot absorb even the lowest modulated heat output from the ASHP at this very low flow temperature, so it cycles. Running the buffer at 40 deg C and using an accurate TMV is a 100% fix for this, as the very much greater heat output power needed to get the buffer up to 40 deg C means that the ASHP can run at a low modulation level for a modest period, then just shut down for a long time, typically 24 to 48 hours, before firing up again. I spent a long time playing around with different control strategies, including a lot of time trying to use slab temperature feedback as a control method, before finally settling on the system we have now.

What's the minimum flow temp of the ASHP? Mine just won't regulate at all below about 30 deg, and even then it jumps up and down by 3 or 4 degrees. If the flow temperature to the UFH exceeds about 24 deg C then we get big temperature over-shoots in the house, even with room thermostats with 0.1 deg switching hysteresis, so I have to make sure that the TMV can accurately hold the flow manifold temperature at the relatively low set point. This was one of the hardest parts to get right, getting really tight control of the UFH flow temperature, and realising it's significance in house temperature stability control. I also use the buffer to preheat the DHW, so wanted that as hot as I could reasonably get it, within the constraints imposed by defrost cycling, so run the ASHP flow at 40 deg C. I've found that the old-style remote sensor type TMV that was fitted to our Wunda manifold is extremely good at regulating the flow down to about 23 to 24 deg C, below that and it starts to have problems regulating, though.

The cheapest way to make brick or blockwork airtight is to just use a standard cement wash parge coat. Mix it up so it's thin enough to apply with a soft broom and you can do a whole wall in minutes, for very little cost. It used to be the standard way to seal up a wall and prepare it for wet plastering years ago, but it is every bit as good for just making leaky brick or blockwork pretty airtight, and it's dirt cheap, too!

I solved this problem, and the mislaid tape measure problem. I acquired loads of free 3m tapes from the company that supplied a lot of the pipe and fittings I used for the borehole; every time I ordered something they sent me a free tape measure. Before Staples closed down here I went in and bought loads of pencils and sharpies in their sale, and have them liberally scattered around all over the place, along with the free tape measures. The golden rule was to never put a tape measure, pencil or sharpie in your pocket, as before making that rule up they all used to migrate back to the old house, when I emptied out my pockets upon arriving home in the evening. Before long I'd have accumulated a large collection if pencils, sharpies and tape measures on the shelf in the kitchen where I tend to leave the loose change, keys, etc from my pockets....................

The cat's been out of the bag for years, as I remember finding out how easy it was to remove back in 2013, and it being discussed on the old Ebuild forum. I may even have mentioned it in my blog, and suspect I've posted a copy of my de-watermarked design EPC either here or on Ebuild in the past.

Yes, I can confirm that the "draft" is just a graphic on a separate layer. Any decent pdf editor will be able to remove it. If you have a problem finding one, you can use PDF Pro online to edit it: https://www.pdfpro.co/