TerryE

I bought a batch of 10 (later bumped 20) canned DS18B20 thermometers, much as @Adrian Walker pic above. I did however calibrate them against each other by immersing them as a bundle in warm water (~50°C) in an insulated container but without a lid and left this to cool overnight logging the temp every 5 mins. I also did a run in water with melting ice. I discarded a couple that didn't track the herd, the remainder had a spread of less than 1°C. I recorded the temp offset of each device by OW address and my app adjusts all the readings by the device offset. Less the discards, the remaining thermometers (and with this offset applied) tracked within ¼°C over a repeated cooling test run. I didn't run a second reference temperature (e.g. 100°C) to calibrate gradient accuracy. However I felt this was good enough for my purposes.

You have to do the sums. The CoP does go down with temperature, but I suspect that the day - night drop is rarely over 10°C and this typically drops the CoP by less 1 one so as long as you do most DHW heating, washing machine and dishwashing overnight, then I suspect that you'll still make a material saving.

Yup, if you want to stay M$ WinX for your PC, then Samba is the best way to go as this talks native Windows Stack. You could also symlink (ln -s) the music folder into your FTP share hierarchy.

I've got 2× SunAmp PVs (the early models): Mechanically well engineered. Compact and very low heat losses. We've never had a problem with the instant heat rates. The control circuit boards have some fundamental design flaws which have never been addressed (mainly no decent 240V isolation and 12A traces totally inadequate, overheating of the main 12A relay). I've had 2 board failures in the first 2 years of operation. Not good. I really like the form factors and the functionality. My long term intent is to keep them but dump the internal controller boards and replace them with some direct control. (I already have an dedicated RPi which does the CH and DHW control, and it would be straightforward to control the SA heaters and pumps direct from this.) The thermal density and energy release characteristics of a PCM store are very compelling IMO, we just need a competent supplier.

All our loos are wall-hung Gerberits. Jan and I did the plumbing. Brilliant. No problems so far after 3½ years.

@DamonHD, Thanks for the link; the site needs a proper review. IMO some form of demand-side adaptive use is a good tool to help get us towards a zero carbon goal, and ToU pricing is one good mechanism to implement this. AFAIK, the Octopus Agile tariff is the first such UK offering available to consumers. I also believe that members of this site -- those with the drive to self-build -- will tend to be early adopters of ToU services. This particular thread is branching into various facets, because of the various usecases that might apply here, and the sweet spot is going to be different depending on individual circumstances and abilities. For example in my case, my annual energy bill is about half of my annual council tax bill and my current energy solution doesn't need third party maintenance or material related consumables , so another 30% savings, say, would be a very small change to our overall outgoings. So to the usecases: Many here will have PV and some Tesla PowerWall-style storage. Maybe more so in the future the "powerwall" will be a flatpack with a wheel at each corner and only plugged in when not taking the occupants out and around ? These are capital intensive and complex systems, and merit a separate thread / discussion, I think. At the other extreme, many will find the challenges of adaptive demand to match ToU in the "just too hard" category. Many will be comfortable with adopting essentially static time-of-day strategies that can be implemented with simple timers etc. That's what we do to make best use of our current E7 tariff: Most of CH, all of our DHW, our washing machine and dishwasher use E7 cheap rate electricity, and this drops our energy bill by ~30% compared to an equivalent single rate tariff. Where things start to get really complicated is when you wish to adapt your household demand truly dynamically to daily ToU pricing, and one option here seems to be the way OVO is going and to use a third-party (e.g. energy provider) AI-based service and allow it to schedule your devices and internal services through technologies such as Alexa. This will be the simplest option for many consumers but this one gives me the willies, personally, because of the security, privacy and 3rd-party continuity risks. Another alternative is to have a largely independent system within the household the executes control locally and only depends on open (therefore easily replaceable) and usually free external services, but with all of the smarts running on a computer within the household. I am also IT literate and a strong programmer, so this is the option that I am most comfortable with, and one that I wish to promote and to support /help any other members that are thinking of going this route, but again perhaps the details merit a separate thread / discussion. What I will say here is that there is now a bunch of cheap and incredibly powerful options (e.g. RPi, Home Assistant, ZigBee / ZWave and loads of devices that employ these, REST-base webservces) that make this very feasibly for anyone with some IT literacy and very basic programming skills to go this route. Perhaps others can expand on cases that I have missed.

I've just had a play with a few approaches. Recall that the way my heating calcs go, I compute how much heat I need the following day each midnight. The best algo seems to be to split this into ½hr slots so let's say I need N slots tomorrow, then pick the N cheapest slots in the next day's ToU tariff and run the heating during these slots whenever they are. Depending on the month this can reduce my heating bill by 25-60%, so cancel my comment about the OVO E7 tariff being cheaper. PS. The price can go negative, so I need to think about what I do on these days!

I had to do a search of you past posts to get the backstory. A 24 kW boiler on for ~2hrs a day: this is a peak single circuit load of 96A. This seems to be case of mega-over-sizing to me. Obviously whoever sized your system took no cognisance of your SAP calculations, and ended up costing you a lot of unnecessary expense and running cost / maintenance uplift. Split milk, I guess, but it must piss you off. ? Out of interest, the maximum we've drawn in any ½hr slot (out of the 40 months that I have been capturing and logging smart meter data) is 4.6 kW and only 2% of the 30slots are over 3 kW (e.g. always-on load + one of the SunAmps on + Willis + washing machine, say). I already run daily queries in Node-RED to get weather forecast data and the previous day's usage. Running one in the evening using the REST API to pull down the following day's prices is straightforward. One thing that you screen grabs do underline is that I should histogram out the various slots to get an idea of micro-placement is worthwhile. I don't have PV, so can't justify any battery storage. I can't justify a RoI based on ToU shifting alone.

Not sure where to put this, but I thought that this might be useful for other members considering switching to a ToU tariff. I've been considering moving my Home system to use a ToU tariff and Octopus seems to be the only supplier offering one ATM. My main concern was that they have a spreadsheet for historic pricing for 2019, but no hard data on current pricing on their website. You can't get access to their forum until you have an account, which is a bit of a problem if you want the data to make an informed switch. Luckily they do offer and document a REST API where you can query the ToU tariff data. I used this to pull the data for the last 2 years, and aggregated this for time-of-day by Month. I've attached the quick and dirty scripts (minimum effort to write) that I used for this as a starting point if anyone is interested in doing this themselves. I used httpie and rs (because Octopus did and I couldn't be bothered to recode these; you can just sudo apt-get these utilities on most linux distros.) I've also added the main summary graphs below. My main conclusions are as follows: Electricity prices are rather high at the moment (winter; an increased demand due to Covid home working; and possibly a Brexit hike), and in my case I have a 24 month fixed E7 tariff from OVO which isolates me from this hike, and so the rates will need to fall somewhat for it to be worthwhile for me to switch. However the ToU tariff is cheaper in comparison to current single rate and E7 "flexible" tariffs, so long as you can stick to some basic rules: Use the Midnight - 6AM window as much as possible for high-load stuff. Avoid the 15:30 - 18:30 peak demand pricing where practical and move as much daytime use out of this window. About the only use we have in this window is for "always on" devices and cooking and this is rarely more than 1 kWh during this window. The price can fall considerably on an infrequent basis (11:00 - 14:30 on sunny days in the summer; anytime the wind is blowing). If you have a Home automation system and use the daily price forecast API, then you can reduce your electricity overall costs further. Analysing our past usage E7 saves us about £15 / week in the winter months and a ToU tarrif would increase this saving to £20-25. Be aware that you need a converted SMETS1 or a SMETS2 meter fitted to allow your smart meter to be switchable between suppliers. In our case, we have SMETS1 meter that is due to be converted in by July. (This OVO form post gives the back story.) Hope this is useful. These are all calendar monthly aggregates by 30-min pricing slot. Min and Max are the usual definition; the remainder are averages over different time windows. Get-tariffs #! /bin/bash BASE_URL="https://api.octopus.energy" PRODUCT_CODE="AGILE-18-02-21" TARIFF_CODE="E-1R-${PRODUCT_CODE}-C" TARIFF_URL="$BASE_URL/v1/products/$PRODUCT_CODE/electricity-tariffs/$TARIFF_CODE/standard-unit-rates/" W=~/Desktop/agile for ((i=0; i<$2; i = i+1)) ; do START=$(date -d "$1-01 $i months" +%Y-%m-%d) END=$(date -d "$START 1 months" +%Y-%m-%d) MON=$(date -d $START +%y-%m) echo Processing $MON http $TARIFF_URL period_from=="${START}T00:00Z" period_to=="${END}T00:00Z" page_size==1490 | \ jq -r '.results[] | [(.valid_from|capture("^........(?<m>..)T(?<t>..:..)")|"\(.m) \(.t)"), (.value_exc_vat|tostring)] | join (" ")' | \ sort > $W/agile-tariff-$MON.lst done Aggregate Tariffs #! /bin/bash W=~/Desktop/agile for ((i=0; i<$2; i = i+1)) ; do m=$(date -d "$1-01 $i months" +%y-%m) perl -ne ' my ($h,$m,$v) = /.. (..):(..) (.*)/; my $i = 2*$h+$m/30; $tot[$i] +=$v; $cnt[$i]++; END {foreach (0..47) {printf("%.2f,",$tot[$_]/$cnt[$_]);}; print "\n";}' \ $W/agile-tariff-$m.lst done

That at ~ -20°C, you'll need about 65 kWh of heat input per day to keep your house at design temperature.

Where do you live? The English average temperature for the winter months is closer to 4°C with the odds spell a few degrees below zero. -20 °C would produce Texas style strain on the Grid; however assuming continuity of supply, the heating could always be supplemented when the temperatures do plummet - for example I use an oil-filled electric radiator in the coldest month to add a little bit of extra heat for my top floors. IMO, you need to do the sums (perhaps assuming that you can do most heating overnight on E7 rates) and work out what your annualised eating building will be at a CoP of 1. An ASHP will give a CoP between 3 and 4, so you can calculate the annual saving if you add an ASHP, and this will give you the investment case to see whether it's work while installing one.

@jfb, my point was partly that 1 part Portland to 2 lime, gives most of the advantages, but the mortar still goes off in 24hrs.

IMO, there are cases where pure lime mortars are recommended e.g. when you are matching on a listed property, but the making and execution of lime mortar pointing is quite specialised. Our last property was a farmhouse that was a mix of 1700s - 1900 stonework, and we repointed all the stonework. An old boy helped us in the early days and we followed his advice. The main thing is that the mortar must always be softer than the stonework being pointed. We kept to a 1:3 cement: soft sand mix, but the cement was a mix of fresh lime and Portland. We found a 2:1 mix of these worked best for our stone, hence it was a 2:1:9 mix lime:Portland:sand. You could work this for up to 4-8 hrs depending on temperature, and brush out excess after maybe 6-12, again temp / direct sun dependent. It would start to harden properly after maybe a day. You will need to experiment with technique: rake out, wet, point, brush out; all spaced out according to hardening rates. Work from top to bottom. The colour is a light grey rather than the paler pure lime mortar. This job was spread over 33-25 years ago. We still live next door to this house, and the stonework / pointing still shows zero sign of degradation.

It was pumped cellulose. It was amazing watching the guys do this. The same crew were also responsible for signing off air-tightness. They first boarded out the inner surfaces of the cassettes with a special OSB that includes an air-tightness surface. They taped all of the joints, including redoing the Internorm seals; their logic is that they want to pass the test first time, so don't trust anyone else's seals. They also battened out the walls for the service cavities. They filled ½ the kitchen and ½ the living room with bales of filler and started at the first floor working up into the roof. In each vertical between the battens, they cut a circular hole at the top for filling access and pumped filler until full, then taped up the hole. As I said we have a warm roof because the under-roof space is a third living floor (that my son effectively uses as his own apartment). Because the rafters were lined with a semi-transparent membrane you could see the filling in action. Amazing. The cellulose was packed in really solid. The independent air-tightness test was done one the last day. The guy was gobsmacked when the house comfortably passed the 0.6 ACH target first time without any tweaks or leak searching; he said that was the first time he'd seen this done. We added most of the through-cassette access channels (e.g. for power for external lights, etc.) afterwards at our leisure. The trick was that I used a 60cm × 15mm masonry bit (because I already had one, but any 60cm drill would do) to drill through inside to out then used the appropriate hole cutter on both sides using the 15mm pilot as centre; and then twisted a PVC pipe through and cut it to length leaving ~ 1cm proud enough to Sikaflex and tape to ensure a good seal. (We then foamed and taped the tubes after the wires or whatever were run.) A good trick here is to use a mutlitool to cut a couple of saw teeth in the end of the pipe that you twisting in, and this way it will cut its way through the filler; it's too solid just to push through. BTW, you do need to think about wind lifting tiles. IMO, it's just too easy to loose chunk of tiles along unsupported eaves and verges in a severe storm. Google Kytun dry verge; they also do eaves systems. Also search the forum. I used Kytun for our eaves, verges, and verge trims; we were extremely pleased with the products. They also do systems for tiles as well as slate and fold custom profiles on request -- which is what we did. I am sure that there other vendors offering similar products if your want to shop around. You might also consider notching the bottoms of your rafters so that you can run a 150mm plank flush along the base of the rafter line to support your tile eaves properly. Again this can be painted as needed or aluminium-wrapped for zero-maintenance.

@Dominic, I assume that MBC are sharing the AutoCAD DRW files. I assume that you have an AutoCAD vewing tool. I used ViewCAD, but there are a few available, and I recommend using one to review and print drawings. Is your roof racked? I can't remember whether this optional or required. Our's was racked as we had a 3rd story in our warm loft space and the roof racking was a part of the shear strengthening design. We also had our frame erected in early Nov and the slater couldn't start until Feb, so we had almost 4 months with the breathable membrane over racking as the whether tight roof. No probs though. But because of this the erection crew just racked over the roof light holes, and I didn't actually cut the light holes in the racking until I agreed the timing and the exact vertical setting with my slater (to align the light positioning to the slate coursing). The main drivers in all this are: Ventilation requirements, both for under slate and any behind the slips. In our case the roof was counter battened and battened so we had under tile ventilation which employed a continuous ripple gap under the ridge tiles, and above soffit ventilation strips. I am not sure that the approach with tiles is, but for slated roofs, you usually have a slight kick on the last course, so this strip doubled to act as the spacer for this kick. We has a 50mm airgap beyond the cassette outer Panelvent, and then a stone course, but this is different for slips. I believe that you uses some form of papered board over vertical battens and fix the slips to the board, so in this case the airgap is directly under the slip layer. Zero maintenance construction. We used black powder-coated folded aluminum sections for our eaves, verges, slate edgings, etc. Rot proof and won't need painting in our lifetimes. Likewise all ties, ventilation strips, etc, where these were rust-proof andhad a min 25-year design life. Detailing. Whatever the approach is, you will need detailed drawings of the cross sections including ventilation mechanisms at top and bottom including any coverings. You need your SE, builder and BCO to OK all of this before you start the build. The last thing that you want to happen is someone crying foul and stating "it's not going to work". In our case we were not exactly sure what the final depth of the airgap and stone skin would work out at, so I got the MBC erection crew to leave the rafter overhand slightly long and boarded out the sarking about 10cm further than needed. I later agreed the exact profile with my builder, and it was a quick job cut off the excess with circular saw -- a lot easier than adding a bit. As well as tops and bases, you also need to detail window positioning and framing, cills, lintels, etc., remembering that you must achieve three goals (i) a decent thermal design with no material thermal breaks; (ii) airtightness; (iii) weather-proofing. I am a little surprised that you don't have soffits or at least a decent overhang. What about eave detailing on your gables? You need to think about the ventilation and the risks of windward side rain ingress. We used std mat back plastic guttering and downpipes but with metal gutter carriers that could be set both in depth and height. This allowed us to position the guttering pretty precisely for best run off.

I have an MBC frame and stone skin. I'll post back later today with some hints, when I have more time.

As an example of this sort of vulnerability:

What I want to do is to update my CH scheduling algorithm after I switch to the ToU tariff. The current algo has inputs terms: (i) based on external temperature, and (ii) a day-to-day feedback to prevent creep over or under-heat. This has a number of limitations: Solar gain. Because the orientation of the principle elevation (SE) and our relatively small front windows we get little solar gain in the winter. The one time we do on clear sunny mornings in heating days where this gain can still amount to 5-10 kWh -- occasional but when it happens this offset has a clear impact (maybe a °C or so) on our internal temperatures. Visitors and visits. Not an issue this last year, but adding say 2 kids and 2 grandkids to the house plus all of the extra cooking etc., is still another 5-10 kWh heat, and a similar step down when they leave. The oil-filled electric radiator. This is an unknown addition from the perspective of the control algo. Perhaps I need to stick to a fixed heat output and used a smart plug to control the heat input. E7 boundary issues. The control system computes a daily input and puts up to 7 hrs in during the off-peak window. Any extra is input only when the hall temperature falls below a threshold, so looking at the event log, (1) and (3) mean that that the computed total is usually an over estimate and this drip-feed of prime rate heating means that the actual total is usually less that the planned total. However outside of core winter, the planned total is significantly less than seven hours, so this is all dumped in during the E7 window. If the morning is sunny, that I might get an overrun of 1°C or so. OK, the daily feedback takes this out of the next day's heating, but you still get an avoidable blip. A ToU tariff softens the E7 peak/non-peak tariff. Yes, I need to adjust for (1) and (3), but having done do it still makes sense to plan say 80% of the required load over the nighttime slump and then top-up as needed any remaining heat in the moderate rate windows. I'll do a blog post on all of this once I have it working.

I don't have any issue with IoT integration. I do a lot of this in my Home Automation setup; I just don'y use any cloud services here. My concern is the additional security and 3rd party service DoS vulnerabilities using cloud services for critical internal services. I use OK Google a lot to save typing, but its no great loss if it doesn't work; the same could be said for Alexa -- so long as I have a mobile app connecting to a local LAN based service such as Home Assistant. I don't want my lights or heating to stop working if the Internet is down, or if some vendor decides to retire support for some 4 year old devices.

You want to investigate whether a Time of Use tariff might give you a better overall rate. A simple strategy on moving as much use outside the 16:30 - 19:30 window when the rates pretty much treble would give you a far better average rate.

I've just done the query. The daily averages (rounded to nearest kWh) for Sep..May are 2, 16, 17, 23, 25, 22, 19, 9, 5 kWh or 4,148 kWh total. I also run a small oil-filled electric heater on an overnight timer in my first floor study Nov-Feb with the hours per night depending roughly on the average external temperature. I estimate that this adds another 480 kWh so the total heating load is around 4,700 kWh with about 90% at E7 tariff. I just had a look at my "JSH style" initial estimate. The main error that I made in this was due to my assumption that all other electricity use ends up as waste heat which for heating days also warms the house: our new house electrics are more energy efficient so the base load is less than I estimated, and therefore we need more top-up. I also used a temperature set-point of 21°C rather than our currently preferred average which is around 22.8°C. Adjusting for both of these, the new estimate (to my amazement) was also 4,700 kWh . On reflection, it looks like my 10kWh mentioned above was just a measure of the thermal banking in the main ring foundations within the warm slab. It looks like the mitigation did its job. TL;DR: our as-built thermal performance is a good ballpark of the as-designed. I am currently on an OVO 2 year fixed tariff at 9.19p and 15.81p per kWh + VAT, so totals roughly £490 for my annual heating cost. DHW, cooking and all other use is on top. Given that I use little or no heating for the 4 summer months, I can cross check my total annual bill against the annualised cost for these summer months and this is consistent with this figure. Incidentally I estimate that switching to use an ASHP would only save perhaps 50-60% of this (at an average CoP of 3 say, but more peak rate tariff), say £290 p.a. That's why I find it difficult to justify the cost of installation of an ASHP. I am going to switch to an Octopus ToU tariff but the process is that you have to switch to Octopus on an interim single rate tariff until they switch your smart meter, so I am waiting until April because I want to take advantage of the OVO off-peak tariff in the meantime. I also have no components in the system requiring annual professional maintenance so I have no annual maintenance bills.

@Stones Jason, I'll need to run a SQL query to give my annualised actuals. I'll post back later. My target setpoint is 22.3 °C, but the way my algo work this is more of a typical lower bound and the house cycles between about 23.3 °C at peak which is around 11:00 to a minimum at 01:00. Of course we could go lower but Jan like to wander around barefoot in sleeveless tops. However on the actuals, one thing that we do know is that a thermal bridge in out slab (I discussed its mitigation in this blog post) adds maybe 10 kWh to our daily slab losses for the ~6 months of net heating -- say £150 p.a. on our energy bills. We can estimate that by switching form UFH to space heating for a few days and letting the slab slowly chill down maybe a degree or so during this window. Annoying but we can live with it.

IMO, never use an external vendors web service to control any internal infrastructure devices.

Ah, the tragedy of the commons. It seems to me that there are two broad classes of self-builder (i) those you are using their personal efforts to maximise the house they can get for a cash constrained investment; (ii) those who are self-building primarily because they don't like the pre-built choices on offer and want to have design control over the house they want to live in. Whereas the first category might want to repeat (and improve) the experience, the second will tend to want to make the most of their "dream house" and so will have different payback trade-offs to make. In our case, we wanted a warm, cheap to run, pretty much zero maintenance house that would last us through our "3rd age" (as the French call it). BTW, in our experience designing for energy efficiency didn't add materially to the build cost. The big premium was as a result of LPA imposed aspects like using locally quarried hand dressed stone for the skin and a slate roof because they wanted the house to fit well in the local context in the village (mostly 1700s and 1800s stone houses). However the fact that our plot was a "freebie" calved from our previous large garden more than compensated for this and the result was a fine looking house that is a pleasure to live in.

The price would be an issue for me as there would be no way to generate a realistic cost-benefit payback period at our level of use. In our case, it is easier and cheaper to take the hit for the resistive heating costs which we can optimise on a green ToU tariff.