SunAmp : Snog, Marry, Avoid?

[Dreadnaught]

14 minutes ago, JSHarris said:

Interestingly, for anyone looking at using an electrically heated Sunamp, charged either by off-peak electricity on a time switch, or excess PV generation, then there is a very good case for not over-sizing at all.  If I had opted for a UniQ 6 as a replacement for our Sunamp PV we would definitely have not had any problems with the controller 50% threshold, as every day we would have used at least 65% of the full capacity.  It was my decision (and it was my decision, Sunamp suggested I buy the UniQ 6 as a replacement for the 5 kWh Sunamp PV) to choose to over-specify, by opting for the 9 kWh model, when I knew that our requirement was usually around 4 to 4.5 kWh per day.

 

Except presumably in the case with the following sequence: one person away (so usage less than 50% of capacity on a given day), person returns and two shows the following day. Cold water on head at end of second shower?

[Jeremy Harris]

6 minutes ago, ProDave said:

Interesting. The plot thickens.

 

I'm pretty sure that Sunamp were completely unaware of the potential problem if trying to "under-run" a higher capacity unit at the time I bought ours.  Everything about the relationship I've had with Sunamp tells me that they would have been upfront and open with me had they suspected there might be a problem.  They know that I've been a bit of an ambassador for their technology, and that I firmly believe that there are really worthwhile benefits from switching over from hot water storage to phase change material heat storage, so they would not have wished to risk that by not letting me know if they thought that the 50% discharge threshold issue was likely to cause a problem.

 

I'm convinced that our use-case, where we normally discharge the Sunamp to about 45% or so, with that increasing to around 65% or more on some days, just had not be considered when the controller code was signed off for production.  I know they test everything very thoroughly, so can only guess that our use-case was one that slipped through the net and just wasn't picked up on as a potential problem.

 

 

[Jeremy Harris]

Just now, Dreadnaught said:

 

Except presumably in the case with the following sequence: one person away (so usage less than 50% of capacity on a given day), person returns and two shows the following day. Cold water on head at end of second shower?

 

Yes, that would probably cause the same problem that we've been seeing. 

 

The key thing here is that running out of hot water is seen as being a major problem.  My wife was exceptionally unhappy when her shower ran cold because of this issue, and no amount of persuasion by me is going to convince her that spending all this money on this new box of tricks was a good move!

[Nickfromwales]

15 minutes ago, JSHarris said:

Interestingly, for anyone looking at using an electrically heated Sunamp, charged either by off-peak electricity on a time switch, or excess PV generation, then there is a very good case for not over-sizing at all.  If I had opted for a UniQ 6 as a replacement for our Sunamp PV we would definitely have not had any problems with the controller 50% threshold, as every day we would have used at least 65% of the full capacity.  It was my decision (and it was my decision, Sunamp suggested I buy the UniQ 6 as a replacement for the 5 kWh Sunamp PV) to choose to over-specify, by opting for the 9 kWh model, when I knew that our requirement was usually around 4 to 4.5 kWh per day.

This refers to my 'perfect storm' scenario I'm afraid, and reiterates the need for installer to specify not Mr Smith. ( Jeremy, please set aside the fact that you're a knowledgable chap and permit me this as a generic response or we'll be here all day ).

For pre-heating eg excess PV > SA > gas combi I would always go one size under if I thought a bigger unit would be part redundant. I oversize if they're the ONLY means of getting DHW, therefore Jeremy I would have suggested a 12 for you based on 24 hours usage not the morning and evening events as I know it would have had enough charge to get you through 24 hours and there would have been no need to explain my reasoning as you'd be the customer and I would be the designer and installer. Even though you can discipline yourself to use the same amount of DHW each and every day, I would never specify that close to the wind as when guests come to stay you'd be outside the bathroom door with stopwatch in your hand. No customers of mine want to buy into that, but admittedly it 'can' be 'overcome' by hitting a "guest" / "boost" button as and when required. That however is not 'fit and forget' and is only done by prior agreement as a caveat, eg I cant get my arse chewed over it then. 

 

[joe90]

From a Luddite perspective, having to overspecify by a large margin or add your own timer is not the sort of answer you would expect from a company making what should be such a good product . I am glad I stuck with an unvented cylinder (but I have no PV.)

[dpmiller]

And all this, folks, is why I'll be working out my own controls on a well-insulated TS.

[Nickfromwales]

3 minutes ago, joe90 said:

From a Luddite perspective, having to overspecify by a large margin or add your own timer is not the sort of answer you would expect from a company making what should be such a good product . I am glad I stuck with an unvented cylinder (but I have no PV.)

Would be exactly the same for a cylinder All these things either need sizing for the worst case scenario or a means for boost under duress incorporated accordingly.

[joe90]

2 minutes ago, Nickfromwales said:

Would be exactly the same for a cylinder 

 

I know, the cylinder you recommended for me is huge, with dual immersions, proper job?.

[Jeremy Harris]

It's worth reiterating why I chose to buy a 9 kWh model, as an "upgrade" for our previous 5 kWh model, as I think it's a key issue that needs to be addressed.

 

The situation we had with the Sunamp PV was that it was the only size unit available when we fitted it.  Sunamp didn't make a bigger unit at that time, although they were developing the Sunamp Stack, that did offer the ability to increase the heat storage capacity.  As it happens, 5 kWh was a good match for our daily needs, and also meant that the unit would reliably charge whenever there was any excess PV generation, or if the overnight off-peak boost was used.

 

I chose to increase our heat storage capacity not to provide us with more hot water, but to be able to better utilise self-generation, by being able to store more self-generated energy on sunny days, so that this could be used if the following day was cloudy.  Being able to store ~ 48 hours worth of heat means that we could ride out cloudy days, and not use the grid to boost the charge.

 

Maximising the benefit from our PV generation was the aim, and not just my aim, either, I know of at least one other Sunamp customer who has made much the same decision as I did, for the same reason. 

 

At no time was I advised that increasing the capacity of the unit would possibly create a problem, because under-use could lead to it failing to charge.  I'm convinced that, at the time I bought our unit, Sunamp were not aware of the potential problem that variable, or under, utilisation could create.  I had all the documentation before I made a decision to buy, and there is nothing in any of it to highlight the "50% problem".

[Ed Davies]

6 minutes ago, JSHarris said:

I chose to increase our heat storage capacity not to provide us with more hot water, but to be able to better utilise self-generation, by being able to store more self-generated energy on sunny days, so that this could be used if the following day was cloudy.  Being able to store ~ 48 hours worth of heat means that we could ride out cloudy days, and not use the grid to boost the charge.

 

I think this is the key point this thread has been dancing around. If you want to make good use of PV you need significantly more than one day's storage so you need to oversize by the standards of systems specified for heating from E7 or a boiler, so the first day or so's discharge will likely leave the store at over 50% SoC.

 

On the previous page @ProDave said:

 

Quote

… to overcome a known problem that SA don't seem willing to properly deal with.

 

Maybe they're perfectly willing to deal with it but are unable for some reason, e.g, because the original firmware developers are unavailable. Maybe they don't even have the source code, who knows? In my experience it's not unknown for small companies focussed on other issues to not give sufficient attention to software as an ongoing development process. Maybe fixing such issues is part of the reason for the work on a new controller.

[Barney12]

One thing first; I have nothing but praise for @Nickfromwales  His commitment to detail (his OCD almost matches mine!), work quality and ethic are right up my street. In fact if it wasn't for his input and intervention in my personal Sunamp journey (much of which is not written on this forum) I would have personally driven from Devon to Scotland  and poured the salty goop through Sunamp's letterbox!

 

Now that's said, I'll continue:

 

We appear again to be having endless string of conversations about "workarounds" when actually we should be focusing on what Sunamp are marketing and what you get. My focus is on the electric units (as that's my only reference point)

 

What I believe to be factually correct based on the information provided thus far is:

 

They are not a direct replacement for cylinders or thermal stores

The usable capacity (and thus comparisons with the above) are not as advertised

They do not and cannot optimise PV. That is simply an untrue statement when they have to be 50% depleted before they call for power. Also; the soft start function throws many PV diverters into a spin!

 

Even THE MOST BASIC of DHW (and space heating) system allows you to "boost" your energy requirement to cope with scenario's where "normal" demand is exceeded. With the electric Sunamp unit you simply cannot do it. I'm sorry but that has to be a fundamental design flaw. End of chat!

 

 

 

[Barney12]

1 hour ago, Nickfromwales said:

This refers to my 'perfect storm' scenario I'm afraid, and reiterates the need for installer to specify not Mr Smith. ( Jeremy, please set aside the fact that you're a knowledgable chap and permit me this as a generic response or we'll be here all day ).

 

Sorry Nick but this is just wrong. ANY system should be able to cope with scenarios that are "not normal" its called a bloody boost or override switch!

 

To give you just one of numerous examples I could think up:

 

So I'm Mr Smith who has no knowledge of how his system works and has NO boost or override switch..........."Dear family, please don't go away for a few days and leave me on my own. I'll have to have three showers in the morning otherwise the water will go cold tonight when I have a bath and I can't do ANYTHING about it"

 

Add the complexity I have which is where my dual 12kwh Sunamps provide DHW and Undefloor heating (220m2 passive standard house) with NO other form of back up and I can can really screw your head with "not normal" scenarios. For the avoidance of doubt the system was designed by Sunamp originally as two 9kwh units but after they both completely failed they were upgraded (FOC) to 12kwh.

[Barney12]

On 27/02/2019 at 10:43, recoveringacademic said:

I've read and re-read this thread now many times. The thing that worries me is that - as written - the problem statement (one instance of the electric charge version of Uniq doesn't work as advertised) appears to be based on a very small sample size : 1.  

 

That sample size is most definitely 2!

More importantly the 50/90% "feature" is confirmed by the manufacturer.

The problem statement is about the effects of that implementation on its advertised performance NOT that the statement is incorrect.

[Jeremy Harris]

The sample size is three that I know of.

 

Anyone who uses timed charge and has variable hot water use may well encounter this problem, it's just a matter of when.  All it takes is a day of low hot water use followed by a day of higher hot water use for the problem to manifest.

 

We just happen to be on the cusp of the critical usage pattern every day, as our usage fluctuates between about 4 kWh some days to around 6 or 7 kWh on other days.  If the Sunamp refuses to charge during the timed charge period after 4 kWh has been drawn off (which it will around 30% of the time, by my estimate) then it's touch and go as to whether we'll get as much as 5 kWh the following day before the hot water runs out.  When the hot water runs out the only option is to manually boost charge the unit at peak rate so that we can get limited hot water back within about 40 minutes to an hour later.

[PeterW]

So my thoughts are ... if this is an issue with the state of charge, and it’s based on the “level” in the green goo tanks, then can you not just lift the sensor array up by a small amount and that will then mean 50% moves to 40%..? I’d expect that these things overcharge anyway to ensure full liquidity phase change so it would be a quick fix. 

[Onoff]

1 minute ago, PeterW said:

So my thoughts are ... if this is an issue with the state of charge, and it’s based on the “level” in the green goo tanks, then can you not just lift the sensor array up by a small amount and that will then mean 50% moves to 40%..? I’d expect that these things overcharge anyway to ensure full liquidity phase change so it would be a quick fix. 

 

Don't you then chance overcooking the PCM below the raised sensor array? Isn't it that it doesn't change state all at once but relies on convection currents that can't be "rushed"?

[Jeremy Harris]

3 minutes ago, PeterW said:

So my thoughts are ... if this is an issue with the state of charge, and it’s based on the “level” in the green goo tanks, then can you not just lift the sensor array up by a small amount and that will then mean 50% moves to 40%..? I’d expect that these things overcharge anyway to ensure full liquidity phase change so it would be a quick fix. 

 

It's not a bad idea, but it's not as simple as lifting all three thermistors up, as the lowest one is cold start safety critical (I believe) and the top one is full charge safety critical (I believe).

 

Disconnecting the centre thermistor and moving it up might work, but we're really second guessing the way the controller interprets these sensors, and I'm not 100% sure our guesswork is accurate.

 

What's really needed is an explanation as to why the hysteresis between fully charged sensing and able to accept a charge sensing is so wide.  A ~50% SoC difference seems an awfully wide band to me, and ideally I'd like to see that reduced to about 5% to 10%.  The hysteresis only really needs to be about the same as the 24 hour heat loss if the unit is not used, which for a 9 kWh unit would be around 8%.

[paulmb]

While I can understand why it might be desirable to defer re-charging until cheap or free electricity is available, I don't understand why it is necessary from a PCM technical perspective.  I mean, is it always safe (from an over-cooking the PCM perspective) to power-off and on again to potentially trigger a re-charge when there's no spare capacity?  i.e. if the Uniq battery was fully charged, and then power-cycled as @JSHarris is doing daily to trigger re-charging, is there any danger of over-cooking it?

[Jeremy Harris]

17 minutes ago, paulmb said:

While I can understand why it might be desirable to defer re-charging until cheap or free electricity is available, I don't understand why it is necessary from a PCM technical perspective.  I mean, is it always safe (from an over-cooking the PCM perspective) to power-off and on again to potentially trigger a re-charge when there's no spare capacity?  i.e. if the Uniq battery was fully charged, and then power-cycled as @JSHarris is doing daily to trigger re-charging, is there any danger of over-cooking it?

 

I doubt that the intentional hysteresis is anything to do with off-peak or free charging, as those are the very circumstances that cause the problem.  If power was available 24/7 to charge the unit then the present ~50% SoC hysteresis probably wouldn't cause any noticeable issues at all.

 

As I understand it (and this is based only on what I have learned about the PCM from informal discussions, some limited research plus experience) there are two issues that relate to the safety of the PCM, in terms of reducing or eliminating any overheating risk.

 

1.  When the PCM is solid, under the cold start condition, there is a risk that the heating element could locally overheat the PCM, as heat conduction through it is relatively slow, compared with the rate of temperature rise of the heating element.  The way this is safely controlled is to sense this condition using the lowest temperature sensor, which then triggers the control unit to enter cold start mode, where it pulses power into the heating element, to reduce/eliminate the local overheating risk.  Once the lowest temperature sensor, plus, perhaps, the centre temperature sensor, starts to indicate that there is liquid PCM around the heating element full power is applied.

 

2.  When the PCM is all liquid, all three sensors are probably used, so that the controller can detect that the entire volume of the PCM is above the phase change temperature.  The temperature is then restricted from being able to rise above an upper threshold (probably around 65°C to 70°C at a guess). The upper sensor may well be the most significant one when it comes to detecting the upper temperature limit, and when this condition is reached the controller turns off the power to the heating element.

 

Those are the two safety critical sensing functions, the remaining one is the one that seems to be causing the problem, and that's detecting when the PCM has cooled enough such that it's reasonable to allow the heating element to receive power.  At the moment I believe that it's the centre temperature sensor that may be the one that's controlling this, when the 50% option is selected, with the lowest temperature sensor probably determining the 90% discharged point.  This is guesswork, but it would seem to fit with what we know, that there are three sensors in a string down the middle of the cell, spaced just below the top, near the centre and just above the bottom.

[Nick1c]

Would it therefore be possible to use any source of electricity (inc. peak if needed) to keep the pcm liquid locally around the heating element and then Pv or e7 until fully charged?

[Jeremy Harris]

4 minutes ago, Nick1c said:

Would it therefore be possible to use any source of electricity (inc. peak if needed) to keep the pcm liquid locally around the heating element and then Pv or e7 until fully charged?

 

 

If you knew how to sense that condition, then yes, but keeping the PCM liquid around the heating element isn't a problem and doesn't contribute to the snag that we're seeing.  What I believe is needed is a way to sense the other end more accurately, up near the top, so that the first onset of crystallisation can be detected somehow and the unit allowed to accept charge.

[Onoff]

Just now, JSHarris said:

 

 

If you knew how to sense that condition, then yes, but keeping the PCM liquid around the heating element isn't a problem and doesn't contribute to the snag that we're seeing.  What I believe is needed is a way to sense the other end more accurately, up near the top, so that the first onset of crystallisation can be detected somehow and the unit allowed to accept charge.

 

So a vertical string of multiple sensors in series that trip out a little before the crystallization temperature?

[Jeremy Harris]

23 minutes ago, Onoff said:

 

So a vertical string of multiple sensors in series that trip out a little before the crystallization temperature?

 

 

 I think so, yes.  Having more sensors should allow for better resolution, and the limited ability to determine (or more accurately estimate) the state of the PCM at any given depth in the cell.

 

I've been tempted to drop a cable with a lot of DS18B20 one-wire temperature sensors into the central pipe where the thermistor chain sits, and hook them up to a logger to see what happens at various depths in the cell, really just to get a better feel for how the thing really works and try to remove some of the guesswork.

[Onoff]

4 minutes ago, JSHarris said:

 

 

 I think so, yes.  Having more sensors should allow for better resolution, and the limited ability to determine (or more accurately estimate) the state of the PCM at any given depth in the cell.

 

I've been tempted to drop a cable with a lot of DS18B20 one-wire temperature sensors into the central pipe where the thermistor chain sits, and hook them up to a logger to see what happens at various depths in the cell, really just to get a better feel for how the thing really works and try to remove some of the guesswork.

 

Do it please. As I started reading your post I was wondering if there was a vertical access "tube". Is the existing thermistor chain removable? I wonder if SA would consider a retrofit sensor array a better controller could make use of? 

 

Is it one immersion heater in these units or 3 x 1kW or am I thinking of something else?

[Jeremy Harris]

Yes, there is a tube down the middle and the sensors are just a bit of wires with three thermistors, covered in heat shrink and dropped down the tube.  It's dead easy to pull this out, as it just sort of dangles down the tube from the top.

 

I believe that it's just a single 2.8 kW (at 230 VAC) heating element right at the base of the cell.  I've not had the lower cover plate off to look, but I think that @Barney12 has, as he had to reset the trip on his a few times I believe.