Sunamp failure - NOW FIXED!

[Jeremy Harris]

53 minutes ago, Nickfromwales said:

The main issue is (was?) that the case is manufactured as 2 parts, and the break is where they have to be plastic welded together. I’d throw a £20 on that being where these fail, as it’s about 9/10th’s of the way up the case not at the top; ergo they’re in the dynamic section where movement is continuous.

 

These things began to look like they were 3 months pregnant after installing and cycling a few times, where they heat up, get loose, and slump into their final resting position. A bit like when you get out of work and take your tie off and loosen your top button. Just too much weight in a flexible plastic box, particularly if that’s been bean-counted too; more reasonable argument for how these early failures have occurred.

 

 

Very many thanks for that gem of information, Nick.  It ties in very well with what I now think most probably happened to our unit.  If this welded joint had been stressed by transit damage then it seems possible, perhaps even likely, that the cell in our old unit was just waiting to fail.  It also ties in with the installer mentioning that some failures were down to transit damage.  Our's had suffered a hefty knock near the top on the side where the pipes connect, which coincidentally seems to be where I think the leak may have been.

 

From what little I could see of the top of our old unit it looked similar.  Different elbows for the pipes but the rest looks the same.  I suspect there may have been several different iterations of the design, some driven by value engineering, perhaps some driven by the findings from product failures.  Seems quite likely that the weld process along that seam could be one such potential failure point, perhaps exacerbated by transit damage.  Easy to see how a knock might start a crack, that then gradually grows over the years until the point where the PCM leaks out.

 

Just found this photo that I remember sending to Sunamp when I found the transit damage back in 2018.  I've added some annotations to it.  The leak seemed to be around that end to the right where the knock seems to have been.  That knock was enough to cause the lid to move left and rip out the rivnut at the lower left corner.  Seems possible that this could have caused hidden damage to the weld.  The plates where the pipe holes are were both bent inwards, presumably by the force applied to the right hand top of the unit when it fell over in transit (and I strongly suspect from the pallet damage and marks on the cardboard that it may have fallen over).

 

[Nickfromwales]

29 minutes ago, Jeremy Harris said:

 

 

Very many thanks for that gem of information, Nick.  It ties in very well with what I now think most probably happened to our unit.  If this welded joint had been stressed by transit damage too then it seems possible, perhaps even likely, that the cell in our old unit was just waiting to fail.  It also ties in with the installer mentioning that some failures were down to transit damage.  Our's had suffered a hefty knock near the top on the side where the pipes connect, which coincidentally seems to be where I think the leak may have been.

 

From what little I could see of the top of our old unit it looked a bit different.  Different elbows for the pipes and no gland on the sensor dip tube.  I suspect there may have been several different iterations of the design, some driven by value engineering, perhaps some driven by the findings from product failures.  Seems quite likely that the weld process along that seam could be one such potential failure point, perhaps exacerbated by transit damage.  Easy to see how a knock might start a crack, that then gradually grows over the years until the point where the PCM leaks out.

A weak point for sure, and one has to wonder how the integrity and shape of the hot base section, then bloated and misshapen vs the cold lid, when brought together, are aligned for correct / sufficient purchase to be practically achieved during their 'welding' together.

 

Lets remember that a proper fusion weld of 2 'plastic' gas mains is acceptable for large underground gas mains (high and low pressure) and these are extremely robust and almost impossible to break apart.

 

My guess is that their process doesn't allow for the weld to be as good, given that (I assume) there is only access allowed from the outside of the pre-filled cell. Perhaps there is a lack of deep penetration so the full thickness of the cell isn't all part of the finished welded join. Who knows...I mean I have genuinely tried to think of how this could be done differently, today, but I am drawing a blank. Production-line solutions would probably be the answer, if this is now being outsourced for eg to a manufacturing giant.

[Jeremy Harris]

46 minutes ago, Nickfromwales said:

A weak point for sure, and one has to wonder how the integrity and shape of the hot base section, then bloated and misshapen vs the cold lid, when brought together, are aligned for correct / sufficient purchase to be practically achieved during their 'welding' together.

 

Lets remember that a proper fusion weld of 2 'plastic' gas mains is acceptable for large underground gas mains (high and low pressure) and these are extremely robust and almost impossible to break apart.

 

My guess is that their process doesn't allow for the weld to be as good, given that (I assume) there is only access allowed from the outside of the pre-filled cell. Perhaps there is a lack of deep penetration so the full thickness of the cell isn't all part of the finished welded join. Who knows...I mean I have genuinely tried to think of how this could be done differently, today, but I am drawing a blank. Production-line solutions would probably be the answer, if this is now being outsourced for eg to a manufacturing giant.

 

 

If failure around the weld is the cause (and right now it's right up at the top of the list in terms of the reason for our unit failing) then that isn't a hard thing to fix during production.  There have been quite a few failures involving leaking PCM, particularly from older units, and, given that Sunamp have improved the design a fair bit over the years it seems probable that they have looked at improving this aspect, especially if it was costing them a lot of money in warranty returns. 

 

May well be that improvements need more than just one change.  Could be that they've improved the weld process as much as possible, but have also made other changes to reduce the stresses, like that deliberate ramp of power on start up.  I know the old unit we had didn't do this, as I remember hearing the contactor clacking in and out as it pulsed power the elements during the first 20 minutes or so.  Making that a linear ramp up will have added cost, so must have been done for a good reason.

 

The emphasis on careful delivery and handling from manufacture to installation seems to be another change that may have been done for good reason.  The setup the installer had yesterday, where the delivery van and lift crew were organised to arrive ten minutes before the installer, so the new unit could be upstairs ready to install and the old unit taken away as soon as it was out, looked to me to have been a process these guys used all the time, and it had a very clear focus on handling the thing very carefully.  The installer even made a point of checking the packaging carefully for any sign of damage.  Goodness knows what he'd have made of the state our old one arrived here in!

 

Fingers crossed that the various improvements made over the past 6 years or so have made these units more reliable.

[Nickfromwales]

8 minutes ago, Jeremy Harris said:

 

 

If failure around the weld is the cause (and right now it's right up at the top of the list in terms of the reason for our unit failing) then that isn't a hard thing to fix during production.  There have been quite a few failures involving leaking PCM, particularly from older units, and, given that Sunamp have improved the design a fair bit over the years it seems probable that they have looked at improving this aspect. 

 

May well be that improvements need more than just one change.  Could be that they've improved the weld process as much as possible, but have also made other changes to reduce the stresses, like that deliberate ramp of power on start up.  I know the old unit we had didn't do this, as I remember hearing the contactor clacking in and out as it pulsed power the elements during the first 20 minutes or so.  Making that a linear ramp up will have added cost, so much have been done for a good reason.

 

The emphasis on careful delivery and handling from manufacture to installation seems to be another change that may have been done for good reason.  The setup the installer had yesterday, where the delivery van and lift crew were organised to arrive ten minutes before the installer, so the new unit could be upstairs ready to install and the old unit taken away as soon as it was out, looked to me to have been a process these guys used all the time, and it had a very clear focus on handling the thing very carefully.  The installer even made a point of checking the packaging carefully for any sign of damage.  Goodness knows what he'd have made of the state our old one arrived here in!

 

Fingers crossed that the various improvements made over the past 6 years or so have made these units more reliable.

🤞

[Onoff]

1 hour ago, Nickfromwales said:

and didn’t feel

much like bending over

 

😉

[SteamyTea]

4 hours ago, Jeremy Harris said:

As de Havilland tragically found out in the 1950's, squar'ish shapes do not tolerate fatigue cycling anywhere near as well as round'ish shapes

I seem to remember, from college, not the 1950s, that the surround was reinforced with mild steel that fails differently below 273K.

[Jeremy Harris]

8 minutes ago, SteamyTea said:

I seem to remember, from college, not the 1950s, that the surround was reinforced with mild steel that fails differently below 273K.

 

 

You might be right.  My memory is from a structures course at RAE Farnborough in the 1980s, where we were shown a cine film of the Comet fuselage being pressure cycled in a big water tank, until the corner of one of the windows catastrophically failed, blowing out a fair chunk of the fuselage structure.  It was a pretty dramatic bit of structural testing, the sort of thing that seemed to be commonplace back then, probably because doing manual stress calcs was such a laborious process, one that was very hard to apply to every element of every structure in any sort of fine detail.

[Onoff]

Why not use a grade 316 marine grade stainless box to contain the PCM? Could be TIG welded and internally braced to limit "swelling". 

 

Thinking 316 (marine grade) stainless would be ok with the "salt" that is the SAT. 

 

OK this is ali but I know some mustard TIGists:

 

 

 

[Jeremy Harris]

Might there be a risk of the PCM getting overheated during the welding process?   From what I remember from discussions around the time when the Sunamp first came out, the PCM was said to breakdown at a fairly low temperature, something like 120°C to 130°C sticks in my mind for some reason.  I'd imagine that there would be a significant risk that the PCM towards the top might get to that sort of temperature whilst TIG/MIG welding a stainless cell closed.

 

If I've understood Nick's description correctly, they fill the lower half of the cell up with hot (so liquid) PCM, then they add the lid and weld the two parts together.  I'm guessing that they must have a good reason for doing it this way, perhaps relating to the viscosity of the molten PCM, rather than pouring the PCM in through a filling hole.  Before I read Nick's description I'd assumed that they filled the cell up via the red cap, but it seems that is just a vent.

 

It might also be made in two parts so the heat exchanger(s) can be fitted.  Looks like this a secured to the upper section before the two parts are welded up.  Presumably the top and heat exchanger(s) assembly is lowered into the still molten PCM, before it cools and sets solid around the thing.  I can't see this being the only reason for welding the cell up after the heat exchanger(s) have been fitted, though.  If the PCM could be poured in through a port of some kind then it makes sense to do the welding of the two halves before pouring the PCM in, if only because then all the welding can be done on horizontal surfaces, by putting the cell on its side and rotating it as it's welded.

 

It seems so obvious to use a fully sealed container and just fill through some sort of fill port that I can only assume they must have a very good reason not to do it this way.

[SteamyTea]

1 hour ago, Jeremy Harris said:

structures course at RAE Farnborough in the 1980

Mine was at Dorset Institute of Higher Education in the early 1980s, my mind was often on other things back then, but it did explain the design of the hydraulic ram clamps we made, and why, on old steel structures you often see a neatly drilled hole at the end of a jagged crack.

[SteamyTea]

53 minutes ago, Onoff said:

Why not use a grade 316 marine grade stainless box to contain the PCM

A rotationally moulded bit of polyethylene is cheaper.  Why they make canoes out of it.

[Nickfromwales]

34 minutes ago, Jeremy Harris said:

 

It seems so obvious to use a fully sealed container and just fill through some sort of fill port that I can only assume they must have a very good reason not to do it this way.

This is what I saw / heard back in the earlier days, so maybe watching these getting cobbled together then bears no relevance now, for completeness.

 

Perhaps filling via that hole then inserting the ‘vent’ is the way now, as you say it would make much more sense. Not sure how that downstream high pressure could be held back by such a retro-fit fitting, but I’m unsure of whether the residual gap at the top was intentional / functional and allowed expansion within the cell, which it may not have if it was fully filled perhaps?

[Nickfromwales]

How do salt and certain metals ‘get along’? Hex is copper.

[Nickfromwales]

1 hour ago, Onoff said:

Thinking 316 (marine grade) stainless would be ok with the "salt" that is the SAT

 Bean-counters prob won’t like the materials or manufacturing costs of that solution methinks, but for north of £6k for supply and fit, there’s should be enough frigging meat on their bones…

[Jeremy Harris]

31 minutes ago, SteamyTea said:

Mine was at Dorset Institute of Higher Education in the early 1980s, my mind was often on other things back then, but it did explain the design of the hydraulic ram clamps we made, and why, on old steel structures you often see a neatly drilled hole at the end of a jagged crack.

 

A stop hole.  They do the same with cracks in aircraft structures . . .

[Jeremy Harris]

5 minutes ago, Nickfromwales said:

This is what I saw / heard back in the earlier days, so maybe watching these getting cobbled together then bears no relevance now, for completeness.

 

Perhaps filling via that hole then inserting the ‘vent’ is the way now, as you say it would make much more sense. Not sure how that downstream high pressure could be held back by such a retro-fit fitting, but I’m unsure of whether the residual gap at the top was intentional / functional and allowed expansion within the cell, which it may not have if it was fully filled perhaps?

 

They must have to leave a gap at the top for expansion, I'm sure, but they could do that if the cell was filled via a port, after assembly, which is what makes me think that they might need a very large opening in order to safely pour the molten PCM into the thing.

 

[Onoff]

35 minutes ago, Jeremy Harris said:

 

They must have to leave a gap at the top for expansion, I'm sure, but they could do that if the cell was filled via a port, after assembly, which is what makes me think that they might need a very large opening in order to safely pour the molten PCM into the thing.

 

 

I was thinking fill the container with the PCM post welding. Tbh they could have a fine thread filler cap. I'd hazard make the containment well and strong enough out of stainless and it would nigh on be immortal. You could even periodically swap out the PCM.

 

I wonder if there's any mileage if someone had an old unit with failed containment to do just that. 

[Jeremy Harris]

3 minutes ago, Nickfromwales said:

 Bean-counters prob won’t like the materials or manufacturing costs of that solution methinks, but for north of £6k for supply and fit, there’s should be enough frigging meat on their bones…

 

The supply and fit price looks exhorbitant to me.  Before Sunamp agreed to replace the failed unit I looked around and found that I could buy the 9kWh/210 litre unit for £2,035.60, including VAT.  Add on another £82.63 inc. VAT for the PV02 key, plus shipping at another £75 and the total comes to £2,193.23.

 

The guys arrived to install the replacement unit just before 15:00 yesterday afternoon.  The delivery guy was off and away with the old unit by 15:25.  The new unit was in and switched on just after 16:00.  The installer finished up and left at 16:35, so all told a couple of people for half an hour and one on his own for another hour and ten minutes.

 

Clearly there was travel time, cost of vans, equipment, other overheads etc on top of these guys pay, but I'm struggling to see how that could add up to the silly price I was quoted.  I could have hired a stair climber from the local hire place for £235/day, including delivery and collection by them.  I'd happily pay someone's day rate for a job that took under two hours, given they might have travelling time etc.  The most expensive trade I've used recently (chap that serviced and certified our treatment plant) charges £600 a day, including 50 miles travel, waste disposal etc. 

 

Someone should be able to swap over a duff Sunamp for a replacement for no more than £600 to £800 in labour, unless I'm seriously out of touch with prices.  The total price should have been under £3k in total, rather than double that, IMHO.

 

On a separate point, this experience of not having hot water has caused me to think about a back up system.  Doesn't need to be super capable or anything, just needs to supply 120 to 150 litres at a mixed down 40°C (the "MAX40" number I think), so we could have a couple of quick showers.  I've no floor space at all, but there is space up on one wall, I've got a space about 500mm wide and 1.4m high on a wall I could use, with a maximum projection off the wall of about 500mm or so. 

 

I've spotted a cheap wall-hung UVC that looks like it would fit in the space.  The idea would be to plumb it to the hot and cold supplies on the wall beneath, with an easy to access isolator on the cold feed, plus a DP changeover switch to allow either the Sunamp or the UVC immersion to be heated.   I'd need to add another 22mm discharge from the tundish, out through the wall and down to ground level, but that looks doable.  I now know a keen local plumber, too, who wants to come over and talk about the Sunamp, so could get him to install the pipework and sign it off.  Looking at the weight I'm sure I could get it up on the wall, ready to plumb in, easily enough, as there is already a big lifting eye by the ceiling right where it needs to go from when I lifted the old thermal store into place, years ago.

 

My gut feeling is that for not too much money it would be useful to have a standby hot water system, even if it only has a modest capacity.  Even a 3 or 4 minute hot shower would be a heck of a lot better than no shower!

[Jeremy Harris]

6 minutes ago, Onoff said:

 

I was thinking fill the container with the PCM post welding. Tbh they could have a fine thread filler cap. I'd hazard make the containment well and strong enough out of stainless and it would nigh on be immortal. You could even periodically swap out the PCM.

 

I wonder if there's any mileage if someone had an old unit with failed containment to do just that. 

 

 

That thought in your last sentence ran through my head earlier, looking at those photo's that @Nickfromwales posted.  Looked to me almost like the beginnings of a DIY kit!  I don't think that much PCM actually leaked from our unit, perhaps only 3 or 4% at most.

[Onoff]

1 hour ago, Nickfromwales said:

How do salt and certain metals ‘get along’? Hex is copper.

 

https://onlinelibrary.wiley.com/doi/abs/10.1002/maco.200290004

[SteamyTea]

9 minutes ago, Jeremy Harris said:

On a separate point, this experience of not having hot water has caused me to think about a back up system

Why not an inline heater, with a manual change over valves.

(I thought you had one in the past)

8 minutes ago, Jeremy Harris said:

don't think that much PCM actually leaked from our unit

Can you get it analysed to find out what is in it. We have a lazer spectroscope thinking at work, not sure if it does all chemicals or just metals, never tried it out.

[Onoff]

1 hour ago, SteamyTea said:

A rotationally moulded bit of polyethylene is cheaper.  Why they make canoes out of it.

 

Yeah but this a box with sharp corners and dodgy welds. They probably need to take a leaf out of sealed lead acid battery book. That plastic enclosure looks less structural than a washing up bowl. If you took the appliance white shell as the biscuit tin, the containment the translucent flimsy insert that holds the biscuits! 

 

Is the thermistor string in its own tube or direct in the PCM?

 

 

[SteamyTea]

8 minutes ago, Onoff said:

If you took the appliance white shell as the biscuit tin, the containment the translucent flimsy insert that holds the biscuits

Not sure I want to play the biscuit game, I am still scared from a Public School education.

 

[Jeremy Harris]

27 minutes ago, SteamyTea said:

Why not an inline heater, with a manual change over valves.

(I thought you had one in the past)

Can you get it analysed to find out what is in it. We have a lazer spectroscope thinking at work, not sure if it does all chemicals or just metals, never tried it out.

 

 

I did have an inline heater, but it was quite limited, so wouldn't run the shower.  A tank that I can heat up when needed and provides the same flow rate as the Sunamp would just make for an easier to use back up I think.  The one I've been looking at can supply about 117 litres of water at a mixed down 40°C and is actually cheaper than buying an inline unit, plus it only has a 2kW heating element, so can just be changeover switched from the Eddi supply, which is nice and easy to do.

[Jeremy Harris]

11 minutes ago, Onoff said:

 

Is the thermistor string in its own tube or direct in the PCM?

 

 

 

The thermistor string is in a dip tube that is sealed at the lower end and comes through a gland in the lid.