Jeremy Harris

I'm a bit taken aback by the amount of steel in that slab, as well as the thickness. @jack has a similar slab to ours, with UFH pipes tied to the single thickness steel, and a polished concrete finish that looks really nice. AFAIK he didn't have an cracking problems, and his slab is very similar to ours, I believe, so 100mm thick in the centre.

I'm with @PeterW on this, there's no merit at all in the top mesh, all it does is add complication and there's not really any significant cracking risk anyway. We have a 100mm slab, with mesh in the centre and UFH pipes tied to it, no fibres in the concrete, no addition mesh anywhere, and it's not a problem. There are lots of similar slabs around, and I'm not aware of any that have had cracking problems. If you want a belt-and-braces approach, then add fibres, as Peter says, but I'm not convinced they are needed.

The guys that supplied and fitted our Silestone worktops (a bit similar to Corian) couldn't cut it on site, but did joint it on site. They cut it in their workshop, using a water cooled diamond saw, to get the dead straight and square edge needed to make the joints, and each joint has a chamfer at the top that is filled with the epoxy resin jointing compound. I watched them fit the worktops, and they took a lot of time making sure that all the surfaces were level, then fitted them with silicone adhesive to the units, with the carefully colour matched epoxy jointing adhesive in all the joints (we had two joints in a "C" shaped layout). The jointing adhesive cures fairly quickly, and they then roughly cut the overspill away with a Stanley knife blade, before the adhesive was fully cured, then waited until it was really hard before doing the finishing. Finishing was done with a polishing machine, fitted with what looked like Scotchbrite pads, that was used to make the joints dead flush with the surface. The end result was almost invisible joins between the worktop pieces, and no sign of movement in the two years since it's been fitted. I did spend a lot of time making sure that all the kitchen base units were set dead level, and all secured to each other very rigidly, so that there was little chance of any movement, and this may well have helped make the worktop fitting go a bit more smoothly. The guys fitted all the worktops, including the island, in about half a day, with most of the time spent waiting for the adhesive to cure and polishing up the joints. I'd say that it's not really a DIY job to cut and fit this stuff, as you would probably spend more on tools and materials than a decent fitter would charge to do the job. I'll tackle most things, but I think I'd just pay to have someone fit worktops like this.

Not bamboo, it spits and explodes violently, according to our neighbour who tried to burn some. I think it's something to do with the way it's manufactured, by pressing strips of bamboo together under high pressure, with some sort of resin to bond it all together. It may also have something to do with the resin coating that seals the stuff. Either way, it's probably not a good idea to burn it. All our offcuts of softwood, like the larch cladding, plus the offcuts of oak from the interior joinery, were given away to a neighbour as firewood. I reckon he had around 4 or 5 wheelbarrow loads of the stuff all told.

I've been getting away with sticking a rubble bag or two of waste in our wheelie bins every other week. As long as I don't overdo it things are OK. I thought I'd get caught out yesterday, as there were two rubble bags packed with short bamboo flooring offcuts, covered with ordinary refuse, all within a wheelie bin bag that had the top tied up. It wasn't till I went to wheel the bin down the drive that I realised just how heavy it was..................

I would guess so, as I don't think anyone uses the non-bonded bead system anymore, for the reasons given above. When we had our cavity walls insulated a few years ago (in the old house) they used bonded beads, and there wasn't an option to have just the free flowing beads as far as I can recall.

That makes a lot of sense; I can see how there could be leakage to earth on the DC side, especially given some of the dodgy installations that were pretty commonplace during the big FIT boom, when every man and his dog was out fitting panels. Interesting that the SB is transformerless, I'll have to see if I can dig out how they have done that. It's the transformer in our inverter that makes the damned thing so heavy, I'm sure.

I'm pretty sure I've related this tale before, but I got caught the same way, via ANPR. The snag is that it's a permanent ban, and because I've got a personal plate the ban remained in force when I changed cars. The only way around it for me has been to drive over the border and use a recycling centre in Dorset. That gets around the problem, but isn't really fair, as I pay Council Tax to Wiltshire..............

I wonder if it might be related to internal filtering? Some types of harmonic filter can cause a small current imbalance, I believe, perhaps enough to cause a sensitive RCD to trip. I'd heard the same as you, that PV inverters didn't like RCDs, and that's why I'd fitted an MCB in the PV slot in the CU, and was a bit surprised when the installer insisted on replacing it with an RCBO, which was when I went digging through the MIs.

That's interesting, Dave, any idea why? The MI's for ours specify that an RCD or RCBO must be fitted, something I only picked up when the installer went to connect it up. I thought that it would need an MCB, so had provided one ready for it, but the installer pointed out the bit in the MIs and replaced the MCB with an RCBO.

It's fine to have an RCBO on the PV circuit, in fact it's good practice to fit one, rather than just an MCB. I'm a fan of all-RCBO set ups, double pole ones for all outside stuff if practicable.

I think Declan is spot on, the old way of blowing in EPS beads was just to blow them in loose. This caused several problems, from the stuff running out if a hole was drilled in the wall to it slumping over time. To get around these problems they came up with the idea of blowing the beads in with a thin coating of wet PVA adhesive, the bonded-bead system. This causes the beads to lock together as a cohesive block after they'd been blown in, so eliminating the problems of slump or them flowing out of holes.

The first house we bought had a disaster area of a bathroom, where the rough granite walls (it had been an outbuilding) were dry lined with chipboard and then tiled. When I ripped it all out it was a disgusting mess of soggy, rotten chipboard, really manky.

Looks like EPS to me.

Yes, Windows can't normally see all the partitions on Linux boot device like this. You can see all the partitions if you plug it into a Linux machine, but be careful, as this is a boot device. If you just plug it into a Linux box the chances are that it will automount and all three partitions will come up in the file manager. Linux doesn't always behave nicely when hot-unplugging devices, to it's worth remembering to unmount the device before unplugging it, just to be on the safe side. Best to just image it first to be safe, as once you have an image file saved you can easily make a duplicate card.

I'm reasonably handy with TIG on steel, but even after a lot of practice (using a big commercial TIG set) I'm not great at welding aluminium, particularly architectural type alloys like 6063. I'd just get someone else to weld these up, if welding it the only way to go, bearing in mind that most architectural or structural alloys will be supplied to T6 temper (or an equivalent) and that the heat affected zone around the weld will be reduced to T0, and will take a couple of years or more to age harden back to something like T6 If you can use mechanical fastenings, then that would be a lot easier. I tend to use stainless pop rivets to fasten aluminium structures, as they are strong and reliable. Dip the rivets in chromate paste before inserting them if you want Rolls Royce corrosion resistance, but even if you don't the joint will probably outlast you. The aluminium alloy framed boat I made a few years ago is still fine, and that's just pop riveted together, without chromate paste:

22mm seems a bit of an overkill to me. 10mm microbore pipe can deliver 2.5kW at a Δt of around 30 deg C and a flow rate below the noise threshold of around 1.5m/s, so that comes down to about 1kW at a Δt of 12 deg C, which is more than enough - I doubt that we'd need more than around 500 to 600W of cooling up at the top of the house, even in really warm weather. There would need to be a condensate drain, but that could be small, as I'm guessing these units have the same sort of peristaltic drain pump that air con internal units use, and they use an 8 to 10mm drain pipe. The existing pump is on it's very lowest setting, so there is plenty of spare capacity available.

You can't just copy the files from the card with a file manager in Windows. Windows cannot see the Linux files on the card in the two other partitions, it can only see the boot partition usually, so they won't show up in something like file manager. Also, this is a bootable SD card, with the operating system on it. You can image (not copy) the SD card easily enough, either from the Linux command line (not recommended unless you know what you're doing, as it's possible to screw up) or by using an imaging tool. For Linux and Windows there is an open source tool called Etcher that will so the job, but it's a bit clunky. For Windows there's a free utility called Win32DiskImager that does a pretty good job. These tools make an image of the SD card, complete with all the partitions (there will probably be three on that card). This saved image can then be used to create another bootable SD card, using the same imaging tools, but in write, rather than read, mode.

The earth layer will be around 8 deg C below about 0.6 to 1m, not 4 deg C, and may well be warmer than this. The earth under our passive slab stays at around 9 deg C, and all the caves in South Wales I've been in were around 8 deg C, perhaps a bit warmer in the really deep and dry ones, colder in the ones with active streamways. A rough and quick calc suggests that you need more than 100mm of PIR around those columns to mitigate the thermal bridge to an acceptable level, there seem to be so many of them that the surface area is large, and the thermal conductivity of concrete, plus the relatively thick section suggests a fairly high heat loss without very good mitigation, plus a substantial condensation risk. I think 150mm of PIR might be the minimum, plus a very well sealed VCL, then cosmetic cladding. I think that taking the roof off and removing the columns, then building inside the tank may well be the better solution. This has so many advantages, that it seems almost a no-brainer, especially as the new roof structure could be made lighter, allowing greater spans, and yet still retain an earth covering to maintain the external appearance. The construction of this tank will have been driven very much by cost, so the cast in-situ floor, walls, support columns and roof won't have been engineered with domestic use in mind, and the roof is almost certainly massively over-engineered, to allow for the deep earth covering, and possible machinery use surcharges on top.

I've just been looking around, and the Panasonic units don't seem to look much different from a conventional split air con indoor units, and seem very slim (130mm). They are pricey for what they are, though. I'll do some more hunting around and see if there are other options. I've just done a few quick sums and it looks like I could get away with 10mm pipe runs to the sort of size unit we'd need, and running a couple of lengths of 10mm plastic pipe would be easier than running 15mm, I think.

A lot depends on the SD card, I believe, as some have built in wear levelling, some don't. I made the mistake of buying a cheap, no-name Compact Flash card for one of my video recorders and it failed after just a few months, decent brand name cards last for years. You can reduce the size of the card image a lot by using a partitioning tool to reduce the size of the largely unused partition, then imaging the card. I tend to use 8Gb cards with RPis, and imaging the whole 8Gb for every back up is a pain, so I tend to create a master image that has the big partition reduced right down to as small as I can make it, then image the card and that way the image file is a lot smaller. Best to make a back up image (not a back up of the files) first, and then play around reducing the partition size on that, rather than risk messing up the original card. It's also a lot easier to play around with partition resizing in Linux than it is with Windows. Making an image is easier and faster with Linux, too, I've found. There are some online tutorials on reducing RPi image files sizes that are useful.

That sounds like a neat solution. Our bedrooms tend to be a degree or two warmer than downstairs in very warm weather, and I'd been thinking of fitting a cheap air-to-air heat pump, high up in the vaulted hall. The challenge was going to be getting the refrigerant pipes neatly arranged, especially as this couldn't easily be a DIY job (the pipe runs are too long for one of the pre-gassed systems). However, I think I could run a couple of lengths of 15mm pipe to where I need to get to with a bit of effort, and then connect them to the cooling circuit on the ASHP. I'll have to have a look around at these air rads - presumably they are a bit like an air con unit indoor unit, are they?

The event logs will be on the SD card, and may well be commissioning test data that may not have anything to do with your unit at all, especially as the card didn't come with the unit. It's well worth imaging the SD card now, before doing anything at all, so you have a back up, as there have been reports of SD cards wearing out under some conditions on the RPi, especially is there are a lot of log files being written to it on a regular basis. IIRC, a card will tolerate around 100,000 write cycles or so, which sounds a lot, but some have reported that cards fail after a while all the same. My personal experience is that I get around 3 years life out of a 32Gb card on my dash cam, and that writes a lot of data to the card all the time the car is running. Even so, for the sake of 15 minutes time and the price of an SD card, making an image and storing it safely seems a good plan.

Just to be absolutely, transparently, clear, I have never, ever, not once, suggested that anyone should keep all their windows closed if they have MVHR! Yes, opening windows will unbalance the system, so just turn the damned thing off when you do this, to save wasting energy..............

I agree, I like the pillars too, plus they look very definitely structural and there to support the roof loads, which will be pretty high, so removing them could be challenging. Solving the thermal bridge problem, if going for internal insulation, is going to be pretty difficult, as insulating every one of those pillars would significantly reduce the space available, and would make them look out of proportion. In terms of decrement delay and the heat capacity of the interior structure (thermal mass has no physical parameters and cannot be measured, so is a bit of a myth!) then I doubt it's an issue, as being earth sheltered will massively reduce diurnal temperature variations anyway, even with one open side. It looks quite deep to me, so the chances of getting significant solar gain through the one open side is probably pretty small, even if it faces due South. I think it's well worth costing up both an internal and external insulation option, to be honest. External insulation has the advantage of a larger internal space and greater freedom to utilise it in the design, but has the major problem of how to insulate the floor. Adding internal floor insulation, with external wall and roof insulation, will still leave some massive thermal bridges, both through the pillar bases and around the edges (although deep perimeter insulation might mitigate the edge problem a bit). Internal insulation solves the problem of the wall to floor thermal bridges, and means less work in removing soil, etc, but still has the big problem of the pillars and the thermal bridging they create. All told I think the ideal solution (which may well be too costly) might be to remove the roof and the pillars, and internally insulate, adding a new green roof to get back the original external appearance. This would be far more flexible in terms of internal layout, would allow the fitting of roof lights or light pipes to get daylight into the furthest reaches, and would also allow a great deal more design freedom. You could opt for something like the design of Mole Manor, with a central atrium to allow light into the middle of the space, for example The roof light over this could be made invisible from all the elevations easily enough, so from outside no one would know it was there (unless they flew overhead). As a source of ideas, it might be worth looking back at some of Arthur Warmby's designs, and through the archive of the British Earth Sheltering Association. They have long since disappeared from the web, but I was a member years ago and downloaded quite a lot of their articles, but I must have archived them somewhere, as right now I can't find them. The chances are they are saved on an old portable drive somewhere, though, so if you're interest I could try and find them.