TerryE

That would both strain the pump and tend to cause the heater to trip. I already have a Y-strainer on my rising main to prevent this sort of crap getting into the house potable supply in the first place. (I have checked this once about a year after we moved in but it's probably due for a recheck). I thought that these have been standard in new builds for years. Didn't your plumber fit one? I will check mine as part of the "5 year service" that I mentioned about, but you need to take the side panel off to get at it. (Loop back to previous caveats on doing this.)

And here is the Heating Curve for last night's heating cycle. Note that the funny ripple is because I have 2 SunAmps, but I decided only to have one being reheated at any time (as the Willis is also typically running as well), so they alternate getting 15 mins each for 3 hours heating. Note that we'd had a big bath last night so the SunAmps were pretty low on stored heat. Also the SunAmp reached the cutoff temperature at about 6 mins into the last 15 min cycle.

@Cooeyswell, BTW, this is easy to diagnose with the lid off if you have a multimeter with decent probes. The cut-out has 240V live in & out push connector terminals on it. The resistance to ground should be around 18Ω (Ohm's law for a ~3kW heater at 240V) when measured at both sides. If one side is open-circuit, then the thermal cut-out needs reset. However, if this happens then you should really get to the root cause. PS. Turn off power to the unit before you start poking around if you don't want any flashes or jolting surprises.

@Cooeyswell, do you use something like a Harvey water softener on your potable water supply? If you don't on the HW, then pipe furring can become a real issue over time and cause TMVs etc. to fail as the fur builds up in such devices and pipework. This is why the SunAmp people say that you should using a water softener for all potable water if you have anything but a very soft supply. This is especially the case as heat cycling causes calcium salts to precipitate out on the surfaces inside the SA. Unlike the classic CH pipework which runs that water in a closed cycle and you add scaling inhibitors to it, the SAs use open-cycle potable water so scaling will become a problem. (Think of what happens inside a kettle, and hint: if you need to descale your kettle even once a year then your water is too hard and the Sunamp units will fail over time.) If your circulating pump is getting noisy, then you do have a furring problem. If you are in this situation, then the unit might need a complete strip down and descale. However, note my above high-jinks: the SA unit needs to be cold before you attempt any panel removal for disassembly otherwise the cells will sag and bulge, and you won't be able to reassemble. I suspect that this a 'once every X years' sort of service, where X depends on residual hardness. (In my case I am at 6 years, and X is probably 5. 🤣) This is all a pretty straightforward job, if you have any plumbing experience as the pipework has demountable joints at all the necessary places. Just make sure you know how to demount standard compression, Pegler Tectile and John Guest Push-fit fittings before you begin. (Watch a Youtube tutorial.) Also take a load of photos during each stage of disassembly, so you can reverse it to reassemble. One of my big bugbears with the SunAmp factory assembly is that they didn't use PTFE tape on the standard compression couplings so many of mine have wept over time until being sealed by the weap furring; these will all need clean-up before reassembly. I guess that this is a job that I am going to have to schedule for next spring, if I want my SunAmps to last another 6 years. In my case I have 2 units, so I can do one at a time without loss of hot water. My concern here is that the thermometer on the heat outflow is an analog one that is clipped to the output pipe: any scaling in this pipe section will cause this thermo to start to read low so the SA control algorithm will tend to run hotter over time and eventually the independent safety cut-out (which trips at ~80°C) will start to trip. I have now lagged around these thermometers, so hopefully they don't read low now. In my case, I already have a load of DS18B20 digital thermometers wired into my CH system (which runs using Node-RED on an RPi). So it is relatively straightforward to tape another DS18B20 to the pipe coming out of the heater (next to the analog thermometer that the CS control uses and to gate the SunAmp demand on that, so that the SunAmp never overheats. @Nickfromwales, you are probably the most experienced member with installing and using SA units. Have you any comments or advice?

One of the things that I really like about Octopus is their pricing transparency. They publish their pricing formula on their customer portal: they add a 100% markup on the wholesale spot price for their distribution costs / overheads + a 14p premium from 4-7pm UTC to incentivise Agile customers to avoid the peak demand window. There is also government mandated 100p / kWh price cap. 100% might seem a bit steep but overall this is still working out at ~30% less than the OVO E7 tariff.

I ran the HW (a bath for visiting grandkids) to cold using this SunAmp for its salt to go through phase change to solid, and I then left it overnight to thoroughly cool, before disassembling / removing the OSB pressure plate and retaining frame. The cells were now solid and nicely cuboid, so reassembly was now straightforward. I will do my commissioning tests at midnight (when the Octopus price drops). I recall that some of our members have discussed issues with their UniQ series SunAmps having issues with sagging or split cells. The lesson is quite clear: Do not remove a side panel unless the unit is well below the phase change as the cells do not have internal bracing to retain their shape at operating temperature without the steel casing support in place. If you need to bring the SunAmp up to operating temperature with a side panel removed then you will need use some bracing formwork to retain the cell side shape.

After some debate, Jan and I decided that we could press out the sag in the cells if I made a suitable jig, so I made up an H frame that is held against the side of the SunAmp by some tape strapping, and which could hold a pressing plate (a square of OSB3, and with two bars that allowed a set of folding wedges to apply the pressure. I then brought the SA to temperature and used the wedges to load up the pressing plate and push the cells back true. All that remains is to allow the cells to cool back to solid phase and then I can reassemble the SA. A bit Heath Robinson, but simple, very cheap and effective. Job done. Here is a pic (note that the wedges are pretty much at the end of their travel as the cells are back in shape and position :

@ProDave Dave, it is 6 years too late to cry about that. 🤣 BTW, The two PCM cells in each unit are roughly the size and shape of a Jerrycan. The internal plumbing is a mix of end-feed, standard compression, Pegler Tectile and John Guest Push fit. I can't understand the rationale here for this .

@Nickfromwales, I have just done a 101 cockup. 😪 The heater thermo hat tripped, so I took one side off to get at the heater and reset the trip. I then did a quick trial heat to see if everything was hunky-dorey. What I didn't do was to put the side back on first. Unfortunately the cells keep their shape because they are retained by the steel outer container, and with one side missing they started to slump slightly. By the time that I noticed the two cells had belled about 5mm -- not enough to cause permanent damage, but enough to pop the unsupported vacuum panel off and to stop me reassembling the unit. I am an idiot. So I now have a catch-22: the cells are still integral, but have sagged slightly, and I need to get them back to their constrained cuboid shape in order for me to be able to reassemble the unit. To do this i need to bring enough of the PCM salt past its phase change (at least 50%) of the contents into the liquid phase with the cell is a correctly sized box formwork for it to reset into correct shape. This is all entirely doable but just a total PITA to do. The unit is currently on a shelf ~1m off the desk and with each cell at ~40Kg weight, this is enough of a lifting hazard for an old fart like me to need a lifting aid(e.g. a pulley) but again thanks to my time in the Royal Engineers, this is something that I am familiar with. I will try chatting to a SunAmp engineer, but I am not hopeful. I suspect the answer will be: "sorry, but it's an out of warranty obsolete unit", with maybe an offer of a small discount on a replacement unit -- which would need me to replumb by HW runs and move my control panel 😞 So it's going to be a case of disconnecting and dismounting the SA (the front unit in the pic below, which as taken 6 years ago during commissioning), and then working on it in slow time. I still need to diagnose the root cause for the thermo tripping in the first place and fix that before I can recommission the unit. The moral of this story is that using a constrained space like this for services is a bit fraught with gotchas. I would have been better having the heavy kit on the deck and moving the manifolds up to the top shelf to leave room for the heavies below -- but this would have required me to have left 2m tails on the UFH loops before the slab pour -- something that I didn't even consider at the time. At least I have the isolation valves with compression fittings so breaking out the unit itself is straightforward.

Nah, it's more a Q of physics, and the algo isn't that too dissimilar to how I heat my slab, but in the context my " issue is with redoing it" comment, I was really talking about SunAmp doing it.

@MikeSharp01 that's why I like 2 off units. The PV units are quite compact (D x W x U): 530 x 300 x 740 mm so two side-by-side are a small footprint and one does use a day's HW so long as we don't want a big bath. They weigh 80 kg, so you really need a hoist and trolley to be one-man manoeuvrable. The Uniq series have a bigger footprint (W x D): 365 x 575 with height depending on capacity, so if we did decide to replace the PVs, I'd probably go for the 210e model (which is 870mm high), but I am loathed to replace the PVs after only 6 years use. And it the circuit board that is the Achilles' heel, and I really can't see what the issue is with redoing it.

@Nickfromwales I have just had one of my 2 PV units fail. In my case the heater is still measuring 17.8Ω which is what I'd expect for a 2.8kW heater. The obvious failure is that the thermal cutout on the heater has tripped, as I read zero continuity over it. What I am not sure about is why the heater overheated, but I need to reset it to diagnose why, and isolate the root cause. This is the next step. (All potable water goes through a Harvey Water Softener so we have almost no scaling in our pipework.) Overall, I like these PV modules, and because I can isolate either if it fails, and I still have HW even if one fails. They have two major design flaws: The control board is a crap design: there is no decent isolation of the 240V power circuits and their track dimensions are far too small. The main power relays are also too small. How got this CE certification amazes me. There is no customer / support engineer diagnostics facility. BTW, the same board is used on their Uniq product range. The gubbins as shown in @Novice pic above is squeezed into a tiny volume with no thought about access for maintainability. The package looks nice, but so what? I can now buy cheap ESP32 based boards with the necessary PWM output, and other I/Os, relays, etc. so my fallback alternative to replacement will be to do a complete refurb and possible replace the control with a stock alternative.

@SteamyTea Nick, your point is valid, but I can only take small steps. I'll update this in a week or so, when I have reworked by optimisation algos and got some initial results.

@S2D2 Thanks but since I have to use the Octopus API for future calculation and monitoring this will be a 1-day issue. I will get the cost of that day when I reconcile my last and first days. I'll just average the previous and next days then scale to balance the cost, which is a one-time spreadsheet exercise. I've got to do my data models anyway to support half-hourly costing.

I've collected half-hourly meter readings from OVO since we had our smart meter fitted in Jan 2018. OVO stopped recording these on the 6th but Octopus only started on the 8th, so it looks like I am going to have to backfill the lost data somehow at some point if I want a complete bumpless data set. 😱 More to the point, I now need to expedite my new algo though TBH even my old "do as much as possible 0-7AM" will still be over 20% cheaper than OVO at current prices.

This comparison might be useful:

@Alan Ambrose Alan, the human-in-the-loop responses seem to take about 5 working days, and are a bit hit and miss, ATM. For example, I have asked for an account on the Octopus forum and am still waiting on a reply. In this case, I logged onto my Octopus Account and navigated to the request Agile form, and submitted a request to switch there. I got an "Oops something went wrong response, but also an automated email which included a custom link asking me to read and accept the Agile T&Cs which I then did. I then got the confirmation of the switch by return -- but note that I already have an Agile-supported smart meter (in my case a Secure Liberty 100 SMETS1 meter that OVO had already done the necessary firmware upgrade to).

Well I've just switched to Octopus Agile. It proved relatively painless and far quicker than I expected: 4th Nov. I had to request an E7 tariff quote and accept this to get the process rolling. 5th Nov. Both OVO and Octopus confirmed that the switch was under way. I had to do an old fashion eyeball meter reading 8th Nov. First day of Octopus supply. 9th Nov. Octopus API access to my Secure SMETS1 meter confirmed, and half-hourly data being captured 9th Nov. Requested cut-over to Agile tariff. Confirmation of the switch by email return. My account details confirm that I am on the Agile tariff. So now I've got to make the changes to my heating algo to make use of the new ToD tariff, but this merits its own blog post.

The HA automation language is quite rich and NodeRED even more so. I can chat offline. I've PMed you.

Yes, but why the scepticism? This is partly conservatism, that is with a small c, but an awful lot is because of the FUD propaganda in many sectors of the press and news. OK this is partly klick-bait driven, but a lot is it is funded by fossil fuel money.

I was just looking at the blog entries and realised that I had never published this one, so I've unhidden it just in case anyone is interested.

@SteamyTea, I like your first-cut analysis. It tells a tale. So ~£3Bn for 60 GWp generation and 500 GWh storage at ~£160Bn. I note that current LFP pack prices are around £100 / kWh at scale, or £100M / GWh, or £50bn for 0.5 TWh. Though at those sorts of volumes we'd probably be looking at Na based chemistries which are projecting to be nearer £50 / kWh, so maybe nearer £50Bn. There are also going to be other Engineering curve economies of scale at those sorts of volumes. We also need to factor in other storage technologies such as pumped hydro, ultra high temp heat batteries; and the flexibility of other non-combustion generation options: windfarm scaling, etc. and geographic spread using interconnectors. There are also games that you can play with overabundance (e.g. go for 3× or 180 GWp generation at an extra £6Bn which will create all sorts of business options for what you can do with the excess (e.g. green hydrogen 🤢). So yes, there are lots of cheaper options that we could develop with the international political will; ones that would make society better for our old age and our children's. The challenge is that we've got a lot of old and powerful men whose wealth is tied up in extraction/ mining, refining and burning stuff and who don't mind investing a little of it in lobbying, 'think tanks', etc.

We've got lots of suitable offshore for large wind farms, and just think of all of those acres used to grow biofuels (The last published ONC figure was 120,000 hectares in 2020) and could convert this to PV or agroPV (e.g. grazing livestock under the PV cover). With some battery and extra pumped hydro we could easily get over 80%.

I've also loaded the following into github as Gists: octopus_agile_download.pl octopus_price_comparison.pl

Yes, the market is complicated. There are different rates, e.g. next day, and various N month rates where the distributor is offering fixed price contracts, say 1 year term, and it wants to lay off this risk onto the generators. There is also a half-hourly price, which is fixed 1 day ahead at Midnight UTC+1 (CET). Individual distributors and generators can also enter into supply contract which are priced by negotiation (e.g. if OVO wants to buy types of green electricity, though these negotiations are heavily influenced by the prevailing spot rate. What this does show is that the Octopus Agile prices track the spot rate far more than OVO does. My E7 peak rate averaged over this year from OVO has been 35p / kWh when the average spot rate is less than a third of that.