Sunamp failure - NOW FIXED!

[Nickfromwales]

[ToughButterCup]

4 hours ago, Nickfromwales said:

 

...

The build quality was horrific, cable management was not just rough as a badgers arse but also (IMO) dangerous, with single PVC 230v cables not in containment or grommeted and more. I have MANY pictures should anyone think I am just a raging buffoon or a liar. No wonder they didn’t last very long…..

....

 

Morning.

I wonder if you would be kind enough to post the images referred to above on here this thread please?

Thanks.

[Jeremy Harris]

4 hours ago, Nickfromwales said:

Hi @Jeremy Harris.

 

I wonder if you could clarify if it was after this thread was posted that you got your offer of the replacement? Feel free to ignore this question, it is purely for my own curiosity, but it would be refreshing to hear that this was a genuine olive branch that you were offered with some signs of sincerity…

 

 

Very many thanks for that marathon response, Nick, full of tremendously helpful information, not least that it seems that my theory about over-pressure may well be completely awry!  I've tried to trim down the key elements with replies, below

 

 

4 hours ago, Nickfromwales said:

SA PV’s just choked up with limescale due to the tiny micro bore pipework in the HEx, and the cheap option “water conditioners” recommended to the customers were not even touching the sides of combatting scale etc. If you live in anything resembling hard water, you will need a full-on water softener to stand a chance here, but all I found were plug-in bits of crap with a clamp on the pipe performing miracles, or not. The amount of crud that I got out of one SA PV in the Forest of Dean was clearly the cause of that lady having no hot water (or resolution) for 13 weeks or so, which is how long she was given either excuses or zero communication for.

 

 

Ours is fed from a Harvey water softener, always has been, and I've never seen any signs of scale in the pipes.  Last time I did some work on the plumbing was to install a "bum washer" toilet a couple of years or so ago, and the pipes I cut into to plumb that in were as clean as a whistle, thankfully.  I can understand how hard water would play havoc with something with small bore pipe work, though, but I doubt that has contributed to our failure.

 

4 hours ago, Nickfromwales said:

Things seemed to worsen with the advent of the UniQ, IMO, where these would often swell up with overheating of the PCM to the point where the lids were bulging 75mm upwards, and the forces thereof were enough to snap the M5 Allen head bolts off that held the lids on. That and cheap overheat stats keeping tripping out rendering the unit dead. Pics of those still provoke gasps, as you need to see it to actually believe it.

Most would jump to the assumption that the incoming cold mains pressure had compromised the SA HEx, but no. To this date, I know of not a single unit that had a failure that had potable water constantly leaking from the HEx. Let’s be Sherlock Holmes for a moment, “Fetch my pipe, my dear Watson, we may be here a while…”.

Mains pressure leaking from an instantaneous water heaters HEx (heat exchanger) would result in a flooding of a property where you needed wellies. This type of failure would result in a constant flow of water coming out at mains pressure, until discovered, and then the mains being switched off to prevent the 10’s or 100’s or 1000’s of litres of water, or more, being lost. In a nutshell, If it was a cold mains related failure, you’d bloody well know about it.

These instances seem to recount my experiences, eg where the units puke out the excess PCM due to it ‘going nuclear’ and the only damage / evidence is the small-ish volume of molten salty yogurt making its way through the fabric of the building (vs constant running cold water flooding out everywhere). Jeremy says in his instance this may be because the PCM sealed the leak, but me, personally, I doubt the PCM could ever hold back static cold mains pressure, and even less where there would be a path already created from the hypothetical hole in the HEx, carved by the molten PCM, to atmosphere.

 

 

Matches our experience, I actually left the water supply on to the Sunamp after the failure, just because it didn't occur to me that this could be related to a water leak, given that the leakage had stopped.  We both showered with reasonably hot water from it the following morning, both because this was going to be our last chance to use hot water for a while and because I felt it made sense to get the heat out of the unit so the PCM went solid and stopped leaking.

 

It did puzzle me that the leakage just stopped, TBH, as you rightly say, if water at 3 bar was entering the PCM cell then it seems unlikely that solid PCM should stop it.

 

4 hours ago, Nickfromwales said:

The issue actually was / maybe still is (admitted to me directly by an ex Technical rep at SA) that the thermistor string that resides vertically in the core was ‘on the move’. Curious of this explanation, whilst stood in front of another trio of failed units in one home and yanking my hair out in despair, I checked the validity of this information / theory. It seemed kosher as the thermistor string was only held in place by gravity plus an ill-fitting compression cable entry gland (I’ll post the pics of these parts when I can find my other iPad) which seemed completely incapable of doing its job; the gland couldn’t close down tight enough at its max excursion so was pointless / useless. This chap was clearly a clever lad and had looked into this with vigour, arriving at this conclusion from (I assume) multiples of other similar instances of failure which forced him to surmise. Saved me a lot of looking / testing / time wasting because he was spot on! Shame he’s not there any more, but SA had difficulty holding onto good people, particularly those who had an opinion.

So, to explain what was happening, the PCM would ‘flex’ slightly during heat / cool cycling and this created physical movement within the core of the cell. There was a flimsy tube set into the centre of the PCM core, which allowed the thermistor string to be pushed downwards into it, and each heat / cool event seemed to be creating a little movement which somehow squeezed or otherwise manipulated the tube which translated into the upwards migration of the lowest thermistor of the string. This could very well have been happening maybe by only 10th’s or 100th’s of a mm a time, but that marked the beginning of the end. As the immersion heater is installed at the very bottom of these things, and the electronics that switch 230v power to the immersion on / off rely on feedback from the thermistor string, as the lowest thermistor creeps away from the heat source that provides the temp reference the longer the immersion stays on and the hotter the PCM becomes. This continues from bad, to worse, right through to ‘salty ceiling mode’ kicking in.

Basically the immersion was then staying on way past the point that the PCM melt-point required, which then led to the PCM58 effectively boiling (I think it did that at around 85º) and there’s the source of your salt-puking problem. The PCM cell is a sealed unit and comes factory fitted with its own internal ‘PCM PRV’ which opens to allow the PCM to escape when over pressure is terminal, eg to allow it to go ‘POP’ and stop it going ‘BANG’.

 

 

This seems to make perfect sense to me.  I haven't had the lid off (simply because I didn't want to interfere with anything Sunamp might request) but the PCM has mostly leaked out around/through the grommets where the pipes enter and exit at the rear.  Looks to me as if the stuff may have come out of the top of the PCM cell, perhaps.  It's then run down the back and formed that big icicle of solid PCM.

 

Based on what you've suggested I've just had a look back at the energy data for the past couple of weeks (I run Home Assistant and have a separate OM115 meter on the Sunamp supply).  What I tried to look for was an indication that it had drawn more energy than normal, in other words that the power had stayed on for longer, in the days leading up to the failure.  There were two days when the energy used was much higher than I would have expected, first on 2nd January, then again two days before the failure, on 9th January.  The night of the failure, 11th January, seems normal.  TBH I can't see how we'd have used 8kWh of hot water on 9th January, there was nothing out of the ordinary the day before to have used a lot more hot water than usual, nor on 2nd January.

 

 

Regarding the new requirement for a cold water PRV at 6 bar, in our case this is a bit of a nonsense requirement if it is intended to protect against the static cold water pressure exceeding 6 bar, simply because our well pump could never achieve that, even if the pressure switch were to fail and keep it turned on.  I'm sure they will fit one, given that it's now a requirement, but I'm not going to be overly worried about it ever operating, TBH.  I still have the drain and tundish from when the Sunamp PV was installed, so that can be re-used.  It's easily visible, so I can check it regularly to see if there is ever a leak.

 

Thanks again for taking the time to write this Nick, it's really helpful.  I will try and take a look at the position of the sensor string when the unit is taken out, to see if it has worked its way up within the dip tube.  If that may be the root cause, then it seems simple to better retain it.  I had to change the sensor string when our original control box was replaced, and well remember it just being a loose drop fit down that dip tube.  This was over 5 years ago now, though, and it has never occurred to me to check it.

[Nickfromwales]

44 minutes ago, ToughButterCup said:

 

Morning.

I wonder if you would be kind enough to post the images referred to above on here this thread please?

Thanks.

For the benefit of others, yes.

[Nickfromwales]

[Nickfromwales]

 

Looks like Edward Scissorhands did this when fecking hungover and riding on a waltzer. Horrific discovery was made when I was connecting the wiring up on the replacement units.

 

If you follow the red arrow, the larger supply cable is the main feed at full 20a potential from the 20a DP switch. Then you see that tapped off by the small grey link between L and 1, without a fuse, to power up the PCB, relays, and circuitry. This means the rest of the live feed to the PCB has the full 20a potential but with smaller cabling and copper printed tracks / components etc which all have far lesser current rating.

 

Then feast your eyes on the blue and brown single PVC cables exiting the box via the top right-most exit; these are the switched 20a feeds to the immersion at the bottom of the unit which are without the second PVC sheath and appear to be without containment all the way down and through the metal chassis. Even my crap Ariston combi has PVC/PVC 230v cabling throughout, outside the containment of the box that the PCB resides within. Note the incoming cable for the 20a mains feed is PVC/PVC, white outer sheathed one, for comparison.

 

Further note the lack of any cable entry glands etc, just shocking tbf….    

 

Then on the two spade connectors on the relay, the ones feeding the immersion, they’ve not been pushed onto the relay terminals like the others. I thought this was because those two terminals may have been different and stuck out further, but no. Just installed with the same fineness as the rest of it. See pic for clarity.

 

 

These were replacement SA units provided to swap out 3 failed units (which left the client without hot water in a £4m home) and to thus (attempt) to allow us to restore the customers faith in the product. One would think these would have got some white-glove treatment at the least(!) but instead they gave them to the 1st year apprentices to slap together…. Faith was not restored after I left and issues began to arise with the replacements….

 

[Jeremy Harris]

I've just been back through over a years worth of energy data for the Sunamp, back to 1st January 2024, to see if I could spot a pattern of the energy use having increased, perhaps due to the heating element staying on longer, as per the sensor string displacement theory.  As well as just pulling graphical data from Home Assistant I've also downloaded the data as a csv file to play around with the numbers in more detail, to try and see if there is a trend that's developed.  Despite my earlier excitement when I found two higher energy usage days this month, it seems they were not at all unusual, the pattern of energy use seems pretty consistent and it just so happens that on some days more charge is needed than others, almost certainly just down to the vagaries of putting the data into 24 hour hour daily bins, from 00:00 to 23:59 each day. 

 

The reason for this seems to be down to how much solar charge there has been.  On days when there's been solar charge, but when there has also been an overnight off-peak charge which was partially used by the morning showers,  the hot water energy seems higher, because the Sunamp is getting two lots of charging in that 24 hour period.  This is definitely what happened on 9th January, the Sunamp was charged overnight from off-peak, partially discharged by two showers that morning, then partially re-charged from solar during the day.  This is the whole year energy per day plot:

 

[Jeremy Harris]

@Nickfromwales,

 

One thing I've spotted in the Thermino MIs is that there is no over current protection mandated for the 0.75mm² always-on cable for the electronics, nor is there any cable length limit specified.  They show the 0.75mm² cable fed directly from a 16A, perhaps 20A, protected supply, which is not a great idea.  If there is an internal fuse then this might be allowable under the "3m back fusing" rule, but if there is no internal fuse, or if that 0.75mm² cable is over 3m long, then it's a clear breach of the regs.

 

In our case I chose to feed the 0.75mm² cable to the UniQ control box via a 13A plug fitted with a 3A fuse, specifically to protect this cable, only because there was no mention of cable protection in the early draft of the MIs I had.  This same cable can be used for its replacement, as it plugs into an outlet that's fed from the HW DP isolator and 16A RCBO.

[Nickfromwales]

1 minute ago, Jeremy Harris said:

@Nickfromwales,

 

One thing I've spotted in the Thermino MIs is that there is no over current protection mandated for the 0.75mm² always-on cable for the electronics, nor is there any cable length limit specified.  They show the 0.75mm² cable fed directly from a 16A, perhaps 20A, protected supply, which is not a great idea.  If there is an internal fuse then this might be allowable under the "3m back fusing" rule, but if there is no internal fuse, or if that 0.75mm² cable is over 3m long, then it's a clear breach of the regs.

 

In our case I chose to feed the 0.75mm² cable to the UniQ control box via a 13A plug fitted with a 3A fuse, specifically to protect this cable, only because there was no mention of cable protection in the early draft of the MIs I had.  This same cable can be used for its replacement, as it plugs into an outlet that's fed from the HW DP isolator and 16A RCBO.

Completely understand that their own MI’s can be in the gutter, no argument from me there.

 

The MI’s may state 16a and that may be the case, thank you for pointing this out. 16a on printed copper PCB is not what I’d want in my product though. 

 

For comparison, every combi boiler I’ve ever worked on has had a fuse before the on board electronics, to de-rate the current available via the always-on electronics etc. Usually this is a 1a quick blow fuse.

 

Clearly you felt this was inadequate, as you say you took it upon yourself to introduced your own DIy fuse protection.

[Jeremy Harris]

13 minutes ago, Nickfromwales said:

Completely understand that their own MI’s can be in the gutter, no argument from me there.

 

The MI’s may state 16a and that may be the case, thank you for pointing this out. 16a on printed copper PCB is not what I’d want in my product though. 

 

For comparison, every combi boiler I’ve ever worked on has had a fuse before the on board electronics, to de-rate the current available via the always-on electronics etc. Usually this is a 1a quick blow fuse.

 

Clearly you felt this was inadequate, as you say you took it upon yourself to introduced your own DIy fuse protection.

 

 

I agree, Nick, IMHO it's not enough to assume that the installer will realise that the requirements of BS7671:2018 take precedence over the MIs, the MIs should really stipulate that this 0.75mm² must have over-current protection with a rating of no more than 6A (the normal max rating for this size flex).  It's no problem if there's a switched fuse or 13A outlet available to just fit a 3A fuse to protect this cable, just in case.  I did this simply because common sense mandated that protection was needed.

[Nickfromwales]

10 minutes ago, Jeremy Harris said:

 

 

I agree, Nick, IMHO it's not enough to assume that the installer will realise that the requirements of BS7671:2018 take precedence over the MIs, the MIs should really stipulate that this 0.75mm² must have over-current protection with a rating of no more than 6A (the normal max rating for this size flex).  It's no problem if there's a switched fuse or 13A outlet available to just fit a 3A fuse to protect this cable, just in case.  I did this simply because common sense mandated that protection was needed.

Agreed, and ironic that Sunamp had these specifically set out in the MI’s for the UniQ’s with external wall-mounted controls! They were all installed with one 16a supply and one 3a (IIRC) supply, separating the current availability accordingly on the predecessor to the Thermino. Same PCB’s give or take, but just seems this safety consideration has just been deleted. Bean counting strikes again, I expect.

 

6a is fine, as long as all the PCB and associated components are all rated at 6a min or above, just strange that these were deemed only suitable to be connected to a reduced current supply by SA previously….

 

Bean counting ahead of safety seems insanity, at least to me anyways, and I doubt if I went to Mercedes to change my car they’d say that there was no need for an airbag anymore so don’t expect to see one, even though my current Merc has one…... “You’ll be just fine sir, just don’t crash” lol.

[Barney12]

I’ve just posted on the “Introduce Yourself Forum” for those that don’t know me (or have chosen to forget 😁).

 

You can add my name to the list of Sunamp casualties. Somewhere on this forum is a picture of one of my Sunamps (I had a pair) bursting its way out of the steel casing. It was replaced under warranty. 
 

What you won’t know is the second unit also failed. 4 weeks after I sold the house! I don’t know if the buyer eventually got any redress from Sunamp but from a couple of emails they sent me I don’t think so.
 

In my latest build (actually a renovation) I fitted ground source with a lovely shiny cylinder full of water. The whole system works like a dream and has never missed a beat. I intentionally went for a German brand with a proven track record in the market. 
 

I would never go near bleeding edge, emerging technology in a property again. It’s just not worth the grief. 

[Jeremy Harris]

23 minutes ago, Nickfromwales said:

Agreed, and ironic that Sunamp had these specifically set out in the MI’s for the UniQ’s with external wall-mounted controls! They were all installed with one 16a supply and one 3a (IIRC) supply, separating the current availability accordingly on the predecessor to the Thermino. Same PCB’s give or take, but just seems this safety consideration has just been deleted. Bean counting strikes again, I expect.

 

6a is fine, as long as all the PCB and associated components are all rated at 6a min or above, just strange that these were deemed only suitable to be connected to a reduced current supply by SA previously….

 

Bean counting ahead of safety seems insanity, at least to me anyways, and I doubt if I went to Mercedes to change my car they’d say that there was no need for an airbag anymore so don’t expect to see one, even though my current Merc has one…... “You’ll be just fine sir, just don’t crash” lol.

 

Interesting that, as the first copy of the UniQ MIs I have don't mention the fuse for the always-on supply, but the second version sort of does (right at the very end)!  Mind you, the documents I had for the UniQ were pretty much all drafts, I believe, rather than the final version that accompanied the product.   Given it was such a simple thing to install I never really gave it much thought, TBH, especially as I'd spent a fair bit of time on the phone with their (very helpful) technical chap, discussing the installation.

[Nickfromwales]

There were so many different manuals flying around back then even SA were sending the wrong MI’s out with units delivered to customers homes. I found one email with a historic manual:

 

 

Top image shows the originally featured (obligatory) 3a fuse integral of the original units, titled “constant supply”, the one servicing the PCB etc, with other versions showing an external secondary supply dedicated to “control circuit power supply”, but at 6a. Here’s 2 shovels, take your pick…

[Jeremy Harris]

That's two versions later than the last PDF I had.  The version I had (just after I had installed the thing) was V2.0.  I think the one I worked to initially may have been V1.7, maybe V1.8, can't be 100% sure as I made a point of deleting all the earlier drafts in case I opened the wrong one by mistake!

[Jeremy Harris]

I've had a look inside our UniQ control box and it does have a 2A fuse on the incoming supply to the PCB, as does the board in your photo, @Nickfromwales.  Not obvious unless you know what these Buss PCB fuses look like though.  From one of your photos, it's marked "F1" on the board screen print:

 

 

This would be fine for back-protecting up to a 3m 0.75mm² cable run, even though it's perhaps not ideal, in that it's generally best to protect cables at the supply end.  It would prevent the PCB tracks from blowing though, I'm sure, and is probably good enough to protect the cable in the majority of installations, perhaps.

[Nickfromwales]

I scanned all the MI’s but no mention, and not seen that type before. A long way away from the incoming terminal though! I suppose without any mention, a-la the PRV discharge details etc, then who’s to know? 
Shame the original docs took time to mention / demonstrate such provision, so another example of folk left looking for their own answers, but good to see it there and good for folks to know, thanks. 

[Jeremy Harris]

24 minutes ago, Nickfromwales said:

I scanned all the MI’s but no mention, and not seen that type before. A long way away from the incoming terminal though! I suppose without any mention, a-la the PRV discharge details etc, then who’s to know? 
Shame the original docs took time to mention / demonstrate such provision, so another example of folk left looking for their own answers, but good to see it there and good for folks to know, thanks. 

 

 

Very easy to miss, Nick, if it wasn't for the fact that I'd used this exact PCB fuse before, back when I built the first iteration of our house environmental monitoring system (the very old Picaxe microcontroller based one) then I'd have missed it, too, TBH.  I wasn't actually looking for it, either.  I just wanted to compare the board inside our box with the one in your photo's, just out of curiosity, to see what had changed (not very much, it seems)!

 

 

[Jeremy Harris]

Out of the blue I had a phone call just after lunch from the installer, who was on his way to us, about 2 hours away.  Two vans arrived just before 15:00, one with the lifting crew and a stock of Sunamps, the other with the plumber.  Within ten minutes the new Sunamp was up the stairs and ready to install, ten minutes later the old one was out, loaded in the delivery guys van and they were off.  The new unit has been in for about half an hour or so and is now charging, the installer's happy all's well and has just left.  Total time from the first van arriving to last van leaving was 1 hour 40 minutes, so not a difficult job.  Does make me wonder how the chap from Bristol had the brass neck to want several thousands pounds labour for a job that took less than two hours . . .

 

Anyway, hats off to Sunamp, I really can't fault the speed with which they've fixed this.  Impressed by both the handling crew who did the delivery and lifting and by the installer.  Poor lad had driven all the way down from Nuneaton to do this job. 

 

 

[Pocster]

3 minutes ago, Jeremy Harris said:

Does make me wonder how the chap from Bristol had the brass neck to want several thousands pounds labour for a job that took less than two hours . .

It’s the Bristol way to vastly over charge .

[Onoff]

On 13/01/2025 at 22:32, Jeremy Harris said:

I'm not sure I have the strength of will to do battle with them over it.

 

I was worried you'd been ill when I read that! 

[Barney12]

34 minutes ago, Jeremy Harris said:

Anyway, hats off to Sunamp, I really can't fault the speed with which they've fixed this.  Impressed by both the handling crew who did the delivery and lifting and by the installer.  Poor lad had driven all the way down from Nuneaton to do this job. 

That's one hell of a result, nice work. You’ve just put another dent in their impressive trading losses……
 

Without wishing to derail your post I happened to glance at Companies House during my current jet lag malaise.
 

Those guys are burning cash at an impressive rate. In the two years to March 2013 a cool c£7.8m. In the same period they list new loans of £15.8m.

 

But it also looks like they’ve landed another £4m in loan notes from the Scottish National Investment Bank in the last couple of months. 
 

Bleeding edge is an expensive game and they’re clearly not playing at it! 
 

I don’t really understand the financials of the market but with a long tail of investors and some hefty debt there clearly must be the view that there is a pot of gold at the end of the rainbow! 

Edited Thursday at 17:12 by Barney12
Typo

[TerryE]

33 minutes ago, Jeremy Harris said:

Within ten minutes the new Sunamp was up the stairs and ready to install, ten minutes later the old one was out, loaded in the delivery guys van and they were off.  The new unit has been in for about half an hour or so and is now charging, the installer's happy all's well and has just left. 

 

What happened to your idea of anti-PCM bund?

[Onoff]

7 minutes ago, TerryE said:

 

What happened to your idea of anti-PCM bund?

 

If the floors solid you could likely fix down 4 sides with CT1 or something. 

[Russdl]

@Jeremy Harris that’s excellent news, let’s hope the current model has some subtle modifications to prevent a repeat failure!
 

We’ve got a few years left on our warranty and I’ll be paying closer attention to our Sunamp in the future. It’s kind of easy to forget it’s there as it quietly goes about doing its thing (or quietly spilling its guts). 

Out of interest are they going to fix the damage caused to the house or have they swerved that one?