RobLe

If you’ve got pre 2000 uPVC windows that are in good nick with internal beads, I think it’s worth replacing the U=2.7 glass units. We did this 15 years ago - our frames are 35 years old now, but with newer planitherm one glass and diy foam filled frames(make sure that water can drain). It’s a tough decision to not change the whole window for a more expensive ‘eco’ one; we could have afforded it - I just hate skipping perfectly good stuff. Plastic is a problem - throwing a window frame away doesn’t help.

That COP is so low, that I think something odd is happening, as opposed to the usual high flow temperature at the ashp. I think ostensibly it suggests an average COP of (124+38)/124 =1.3. This may or may not include the weekly immersion use. You don’t complain of a cold house; in this weather if the COP were actually that low I’m surprised the unit can actually warm the house sufficiently. Has it always been like this? Suggestions to try: I note that 124kWh elec in a cold week is not insane - can you calculate your heat loss in kW, see how it compares to an hour of the generate+consumption at the design external temperature? Fit an independent mid-certified elec meter to the ashp, and also a heat meter - more effort this, but will give a believable COP. Eg opentherm.

You can figure it out from a thermal conductivity (U value) ratios, based on a known value: Indoors, in still air, a vertical boundry air layer has a U value of 8W/m^2/degC. That is from the centre pane glass surface, to the air temperature in the middle of the room. The U value of the glass represents the heatflow per m^2/degC through the glass, based on the air temperature in the middle of the room and the outdoor air temperature. So, if you measure the values below with the same IR thermometer pointed at a bit of tape stuck on the glass etc: Tindoor = the indoor air temp Toutdoor = the outdoor air temp Tglass = indoor centre pane glass temp Uglass = 8* (Tindoor - Tglass)/(Tindoor - Toutdoor) I can generally get the answer within +/-20% of the expected value when it's cold outside. That's good enough to tell if the units are just glass, or have a coating on them.

Your walls likely don’t have CWI (cavity wall insulation); you can usually see the pattern drilled into them if it’s been done. Typically 25mm dia holes into morter, every 500mm or so over the entire wall. Suggest taking a photo of the brickwork (regular, not IR), and we can likely tell.

It’s most likely that your DG windows are U=2.7W/m2/degC, or thereabouts - that’s what you get from two panes of glass with a 16mm air gap between, as happened until around 20 years ago. It’s possible to measure the U value of a window. Now there’s a metal low emission film applied, a warm edge spacer, and ‘heavier than air’ gas inside the cavity. The obvious thing to check first is if you e had your cavity filled yet? You can generally tell from the brick pattern outside. I say this as the wall looks cold too.

While I agree you can draw over 3.6kW in a home, in practice I believe it’s sufficient for most people. We have 2+2 people here, and don’t make any attempt to limit our combined appliance useage - kettle/cooker/washing/gshp(v small) whatever. Cars charge and dhw heats at night on the cheap rate. Our 3.6kW sunsynk +15kWh Seplos batt system charges at night in the winter, and gives us 97.3% cheap rate elec useage. The most significant aspect I found for improving the cheap rate useage has been getting the CT coil noise related offset working well. pic below ignoring export:

Yes. eg a sunsynk 3.6kW Ecco, will allow 2 strings of 425V*13A max. It can charge a 16cell LiFePo4 batt at 90A while sending 3.6kW to grid. Or get an even bigger unit that is g100 certified, and get the overall grid export limited to 3.6kW.

There might be an issue with a heatpump being force fed super hot water into its heat exchangers, which a gas boiler could do. Doubt an issue in practice, but let’s say you pushed in water at 90C then the refrigerant pressure might end up above the heat exchanger design pressure. Maybe. We ran with a heatpump and a gas boiler for a year and a half, until Mrs RobL was convinced the new fangled (diy) heatpump was ok. The electrics totally switched one or the other in with a 4 way rotary switch, the water flow was magnetic filter->gas boiler->heatpump, and the heatpump flow & return could be closed off and bypassed. The gas boiler was eventually taken out, as it was never used, gas meter gone, no standing charge.

How is this unit one year on? Does it cope ok with heating during cold weather ?

I’ve been looking out for R290 based regular split system air conditioners now that regs have relaxed and allow them - there’s one by electriQ, and it seems like more are on the way: https://naturalrefrigerants.com/the-rise-of-propane-based-residential-ac-in-europe/

Consider upvc 3G instead of 2G. Supply only is £170(3G) instead of £120(2G) for an opening 1m x 1m window*, and it’ll get the u from 1.2 down to 0.7. It’ll be a bit quieter, and even less cold feeling nearby. Likely installation actually costs more than the units themselves anyway. When they’re being fitted, get some decent air tightness foam in the gap between frame and brick; usually there’s sealant on the outside and just a bit of trim glued inside. *https://www.modernupvcwindows.co.uk

We had a trippy circuit for ages, it got worse and worse - the gap between tripping of 6 months, 3 months, 1 month, every week. Tried a current clamp around the combined tails that some internet bod recommended, all the circuits had some, that one didn’t stand out. By the time it was tripping every week I was switching off a few double sockets for a week at a time, trying to isolate what it was - nothing I did made any difference. We had some stuff running on long extensions from sockets on other circuits, as part of this elimination process. In the end I swapped out the RCBO with an identical one. That fixed it, no trips in over a year now. YMMV.

We’ve never really bothered staggering loads to avoid peaks. I have however trimmed the inverter ‘zero’, and that makes a lot of difference. That is, the CT clamp feeding the inverter to tell it what the power draw is, that the inverter is trying to null, and gives a small signal that is easy to swamp with mains hum giving an offset. With most inverters this can be tweaked out such that the smart meter shows 0W or maybe slightly -ve (exporting) rather than slightly importing all the time - this simple change makes a significant difference to the daily peak import. I think I optimised this offset mid April, the pic below shows the peak consumption happening every day before that. There’s still sometimes a high daily peak import - I think that’s mostly the Ohme charging the car at odd times of the day - octo-aid doesn’t understand it’s actually octopus giving cheap hours.

We have a 3.6kW sunsynk hybrid inverter, with 3.6kW of PV on one string, which works fine. I intend to add another 5kW onto its other string next year😁. The sunsynk can output 3.6kW to mains whilst charging a battery at 3.6kW - which seems perfect - the difficulty is finding a way to ensure that the battery has enough unused capacity so that it’s got space during the peak few hours of the day. Has anyone got this inverter with over 3.6kW of PV, and got a good way of managing summer battery use so that peak daytime PV isn’t wasted ?

We have that 3.6kW sunsynk; it’s enough power for us. Attached is our Octopus days use for last Saturday, which is the heaviest use it’s ever had in the year we’ve had it, as Mrs Robl cooked thanksgiving turkey dinner for 18 people! We have an all elec house(car, gshp,etc) - in the morning the tesla and home batt were charging. By 8pm the 15kWh home batt was flat 🤬 - but this was a very unusual day, I think it goes ‘flat’ 5 times a year roughly. You can see that even with loads of heavy cooking - and Mrs robl makes no attempt to stagger elec devices when cooking - there’s hardly any elec drawn during the peak period (Oct intelligent go):

Apols, yes, capacitors. V acrms means it’s a root-mean-square-voltage measurement of an alternating current signal. It’s volts not amps we’re measuring, just the AC is how we describe the waveform (weirdly). Certainly our house sees very high voltages when it’s a sunny day. There’s a few houses round here exporting, and all that export increases the grid voltage.

Those caps that fail are thin metal film, plastic insulation ones. The thin film is a safety feature, it means that internal insulation breakdown faults cause the film to erode away gracefully rather than abruptly, making the capacitance drop over time. I think that in PV type households this type of component used to give a constant current source giving non isolated low voltage suffers quite badly - I say that as we regularly get 252V acrms, and these caps start to degrade 😞. We had this with our old solic, also with the Honeywell heating controller (on cap#3).

A V2G car charger can import or crucially export, so "uses up" some of your export allowance. A hybrid inverter does the same; but usually you can and do add PV to the hybrid inverter rather getting another inverter, so this doesn't take up extra export allowance. Ours could be software limited (not by me) for G99 purposes; the software at the time was such though that this limited import and export identically, so this wouldn't work. As far as I know they have never been actually installed as V2H only - ie. with "never export" hard coded in - obviously it could be, it's "only software" 🙂 Maybe you could get an install with a G100 restriction and an extra PV inverter, again only software 🙂. I've not come across a 3phase V2G unit; single phase ones are rare enough already - I can't imagine demand for a 3phase unit being that strong, so it would cost more. The electronics inside a V2G box are similar to inside a hybrid inverter, but low volume production makes the reported 1P V2G product price around £3000, versus actual 1P hybrid price of around £700. Price is an ethereal concept for objects that have never actually been sold - I think V2G has always been in trial, never sold without strings attached. There's nothing wrong with the concept of V2G; the devils in the detail for consumers though. Having gone through it, I think a hybrid inverter + home batt is a better thing to actually spend real money on right now, as they just work. If V2G achieves mass adoption things will change; it's not obvious to me this will happen before battery prices drop such that nobody cares. Indra made a great summary of their V2G trial, here: https://www.cenex.co.uk/app/uploads/2021/05/Sciurus-Trial-Insights.pdf

We were on a v2g trial for 2 years with a Nissan leaf. We kept it another year after that, then gave it away (post install it has very low value as the company refuses to allow extra units to be connected to their servers). It worked great for us - largely because it was thru lockdown, so the car was just sat there! Things to consider before doing this: It uses up some of your grid tie ‘allowance’, making adding PV more difficult (this is why we stopped). You’ve got a car on yr drive/somewhere all the time. If you have such a spare drive space, you could rent it out on ‘park on my drive’ or some such. The tech is so rare it’s expansive - for the suggested cost of new inverter+install you can get a hybrid inv+15kWh batt It’s all still trial stuff - nothing actually for sale yet. And it won’t work without internet connectivity. In my experience it was more prone to faults than regular non-internet connected inverters. You don’t get the full use out of the car batt - our 30kWh Nissan leaf+v2g had the same useable capacity as our much smaller 15kWh lifepo4 battery - the chademo protocol doesn’t allow the bottom 30% to be used, and an old 30kWh leaf won’t actually be 30kWh anyway. I think there’s a very niche place for v2g ev campervans one day. Maybe tesla will add a connection between car and their power wall.

Vaguely reminiscent of my vent axis sentinel kinetic BH plus - it was installed 12 years ago ish, and some 6 or so years ago it started leaking. When I took off the front panel, the plastic base had a cm of water in it (it’s right under where the mains elec pcb is!). It has a drain, that had been working absolutely fine for years, with lots of condensate coming out of it. In confusion I ended up tee-ing in the other condensate drain which fixed it. On my model the drains both come from the same void at the base of the unit, so it can’t be an issue using both. Maybe I should have some sort of routine dettol type drains maintainance ? I didn’t actually discover what was wrong, just bodged another pipe on🥴

We got Octopus to remove our gas meter a year and an half ago, they do it for free, can’t say fairer than that! They took the meter itself, capped off the plastic pipe inside the meter box outside our house, turned off the pipe shutoff and took the lever away for good measure. I think a year after it had gone we got a letter from Cadent saying they’d come and inspect - and that if they thought it was unsafe they’d charge to take out the pipe all the way to the road - and that would be the pricey bit(£1500 ish I think). The pipe was still nicely protected in the meter box, so they went away happy - no charge 😇

When propane is cold and/or compressed, it will be a liquid. When it’s hot and or at low pressure it will be a gas. The chart shows the transition from gas to liquid, which is at a certain pressure at every temperature. Tbh it’s all quite non intuitive, but that phase transition is where most of the energy is stored enabling heatpumps to operate. Generally heatpumps are refilled at the cold low pressure side, when they are on, so that the low pressure there allows the fluid to be pushed into the system.

I suggest to liberally spray suspect joints with a water/washing up liquid mix. If there is a leak, you generally get a bubble forming. Our R290 system generally runs at around 13bar (hot), 3bar(cold) - of course the actual values depend on the temperatures involved - the temperatures should be slightly above the saturated values in the snip below at the measured pressures - the compressor in and out always has gas in it, never liquid. Snip from a handy free danfoss ‘ref tools’ app. My gut feeling from what you’ve said though is that the compressor is suspect or the metering device is too open - I would expect pressure post compressor(guess that’s what you measured- normally you would measure both ports simultaneously, and also temperatures at the compressor) to increase rapidly after switch on. 5bar at 9C ambient is I think the expected pressure….I am at the limit of my knowledge I’m afraid though. I assume you know r290=propane…. so don’t even think of getting a spark / flame anywhere near it.

I found exactly the same. That is, anecdotal use involved an osb sandwich, while the actual bba certification calls for the waterproof layer to be directly on top of tr27. I think osb under the tr27 is in a danger zone, should anything go wrong with the vapour barrier below or the waterproof layer above, and will rot. However practically not doing that makes for a springy roof surface that can easily get damaged by foot traffic. I went with osb under epdm, put lots of solar panels on it, walk about on it, I’m happy with it. I especially liked the osb tg4 interlocking as it made for a nice flat surface which I might not have got with tr27 directly as I was doing a refurb, and not starting with a perfectly flat surface. The only thing I’d do differently is choose a lighter colour membrane, what little of the epdm is visible around the PV gets hot in the summer sun.

Short cycling of heat pumps was a real issue, back in the bad old pre-inverter days. The rule of thumb was that greater than 6 start-stop cycles an hour was likely to lead to fatigue failure of copper joints near the compressor due to the start up impulse. I would expect that particular issue isn't much of a concern with a modern ASHP inverter drive, which I would expect to soft start. GSHP are often still on/off relay driven simpler units, I think as noise is less of a concern for them, so they are more likely to still suffer fatigue failure with excessive cycling.