@cathoderay 

I don't know much about PHEs but I wouldn't think they would work very well if the ASHP were cycling. I don't know though.  How are all these flow temperatures being measured?  It's quite hard to measure flow temp accurately and even the probes that come as part of the installation have to be installed properly (mine weren't!).  A badly installed one will read significantly lower than it should.

Posted by: @derek-m

Correct me if I am wrong, but did we not discuss the problems associated with your heat exchanger last Winter?

Posted by: @kev-m

How are all these flow temperatures being measured?  It's quite hard to measure flow temp accurately and even the probes that come as part of the installation have to be installed properly (mine weren't!)

@derek-m - yes, we did, but didn't get to the bottom of it, largely because there weren't cold enough ambients to test the system.

@kev-m - I agree - there is a lot of black box stuff going on here. I am hoping I can find a way of seeing in the dark, so if/when it gets cold, I can try and make some sense of what is going on.

The PHE is currently an enigma. 

@cathoderay

Let me see if I can explain PHE's in terms that most can understand.

Most people understand radiators, warm water flows through the inside, the metal gets warm and heats the surrounding air. Simples!

A PHE is like a radiator, but located inside an insulated container. It is actually a container, but with a plate (hence the name) down the center separating the two halves. The plate is corrugated to provide maximum surface area to the liquid on either side, and hence maximum heat transfer.

When warm water is initially pumped into the primary side, it will start to heat up the metal and cool the water, so the outlet flow will be cooler than the inlet. If there was only water flowing through the primary side, then eventually the metal and any liquid in the secondary side would approach the temperature of the primary water, and the outlet temperature would be very close to the inlet temperature.

If cooler water is now passed through the secondary side, this would absorb heat energy from the primary water via the plate, and in turn cause the temperature of the water at the primary outlet to fall, whilst at the same time the temperature of the water at the secondary outlet would start to rise. Heat energy is therefore being exchanged between the primary and secondary liquids, but the temperature of the secondary liquid can never exceed that of the primary liquid, and in most cases will be several degrees lower.

The rate at which heat energy can be transferred from the primary to the secondary of the PHE is very much dependent upon the temperature difference between the primary and secondary, and the physical size of the plate area. Just as it is with a radiator.

So looking at a practical situation, if an ASHP is capable of providing 10kW of heat energy, operating with a DeltaT of 5C, then it is a fairly simple process to calculate the required water flow rate to move that amount of heat energy from A to B.

On the secondary side of the heat exchanger the flow rate to absorb 10kW of heat energy at a reasonable DeltaT can also be calculated, but the 'exchange' can only take place if the heat exchanger is large enough to permit it to do so. Just as if a radiator is too small for a given room size, it will struggle to heat the room.

So what would probably be the case if 10kW of heat energy, at a WFT of 50C, is put into a PHE that is only capable of transferring 5kW of heat energy. The water entering the primary side would be at 50C, but the water leaving the primary side would not be at the required 45C, since only 5kW of heat energy are being absorbed, so the outlet temperature may be approximately 47.5C.

What would happen on the secondary side. The water flowing through the secondary side would only be able to absorb 5kW of heat energy, so the outlet temperature would be lower than anticipated. This in turn would be fed to heat emitters expecting 10kkW of heat energy, but only receiving 5kW, so they will run cooler which in turn lowers the temperature of the water going back to the PHE.

The end result would probably be that the heat pump cycles, because it is producing more energy than can be transferred, and the radiators and the rooms would be cooler, because they are not receiving the required amount of heat energy.

Please post any questions if there are sections that you do not fully understand.

@derek-m If this is the constraint, there are two ways in which it can happen:, because the PHE heat transfer is limiting or secondary flow rate is limiting.  In the first, I'd expect both secondary outlet and return temps will be sub-normal.  If it is the latter, secondary outlet temp will be normal (close to primary) but return temp low.

@derek-m - Thank you for the detailed explanation of the principles behind how a PHE works. That is indeed my understanding of the principles, though I would not have explained it so clearly. I also appreciate how an inadequate PHE (the analogy with a rad is helpful) or flow rate or both could compromise the system, and lead to inadequate heating.

The problem is the practicalities. Freedom being obtuse and unhelpful, there is no point in talking to them, they will tell me to get lost. My installer was under the impression they don't do individual PHE calculations, they just supply one big generic PHE (economies of simplicity and scale presumably) that should be big enough for most domestic installations. There is nothing on any of the paperwork to say what make or capacity the PHE is, though I have determined by looking at it that it is a Secespol/Hexonic one (the foam jacket has Secespol on it). Initially my installer's plumber left it with the secondary inlet/outlet the wrong way round, ie no contraflow (rather confirming that installing PHEs is unsurpisingly and understandably not routine stuff for a jobbing plumber) but that got corrected (by rotating the circulating pump 180 degrees). It also has its long axis horizontal when ideally it should be vertical, but Secespol/Hexonic have confirmed that is tolerable, even if not ideal. Beyond that, no one knows anything about it, except perhaps Freedom, but they ain't talking. The PHE itself sits in its foam jacket wedged in a tight space between joists and floor boards. I can't see any marking on it to indicate model number or capacity or anything else. I can't even measure its dimensions accurately enough to use them to determine what model and so capacity it might be. I'm not even sure which catalogue I would look in to try and find a matching model if I did have the dimensions which I don't.

Freedom presumably know what model/capacity it is, but won't say. I am utterly frustrated by their refusal to talk to end users. 

They could also do a lot to improve their dreadful service if they explained why PHEs/low loss headers were once ABSOLUTELY MANDATORY (no PHE/LLH, NO warranty), now they are optional. Why the change? For heavens sake, Freedom, pull the proverbial and tell us your reasoning. We have to live with the consequences of your decisions.

Posted by: @derek-m

if an ASHP is capable of providing 10kW of heat energy, operating with a DeltaT of 5C, then it is a fairly simple process to calculate the required water flow rate to move that amount of heat energy from A to B.

I am sure that is right, but I have not managed to find those fairly simple calculations. It has come up before (that they are fairly simple calculations) but it has never been explained how to do the calculations. I presume it starts with how much heat (energy) a litre of water can hold (already we hit complications - the fluid isn't water, its a water/anti-freeze mix) but presumably you also need to know how fast heat is being transferred in/out (as in fast transfer means water can flow faster, slow transfer means it has to flow slower) and it soon becomes even more complicated. The Midea controller reports the Water Flow as being invariably around 1.45 M3/H, presumably that is cubic metres per hour, whenever the room stat is calling for heat, ie it is pretty much running 24 hours a day in the winter. There is even less information about the secondary circuit. I suppose it would be possible to determine the volume of water in the rads, but how much is in the pipework is more like guesswork (some are buried in solid floors, no way of knowing the actual run). The Grundfos circulating pump is equally opaque. From the manual:

"Control mode explanation Proportional pressure [this is the mode recommended for ASHP/rads, set by 'smart' LEDs of the pump itself] The head (pressure) is reduced at falling heat demand and increased at rising heat demand. The duty point of the pump will move up or down on the selected proportional-pressure curve, depending on the heat demand in the system."

How on earth can the pump know what the heat demand is? It is just connected to the mains, via the thermostat, and so the heat demand is either on or off. There is no way it can know of this mysterious rising and falling heat demand.  

The curves themselves are even more ludicrous. Immediately under the above quote we have:

which is nice and clear (not - you can either have not very much, something in the middle or quite a lot of something, whatever that is) but the chart/curve which takes the biscuit is this one, which is the main performance curve chart):

It is one of the most outstanding examples I have ever seen of how not to do a chart. It's utterly meaningless. For starters, neither axis is labelled...

Getting back to the fairly simple calculations, the parameters we do know are:

The ASHP will be putting out something between around 5kW and maybe 10kW (who knows what they actually put out at any given time and set of conditions?). On a fair day with a following wind it might get to 14kW but it won't because on a fair day with temps in the teens it will be throttled back (by weather comp). It will also almost certainly be cycling, 5kW for 20 mins, 0kW for 20 mins etc. Does that mean we use 2.5kW as the output? More on cycling in a moment.   

At -2 ambient, the building heat loss will be about 12.5kW. At any other temperature it will be different. It is also affected by, and this is not factored in to standard heat loss calcs, weather eg the strong winds as we had earlier this week definitely cool the building faster. Perhaps we do the calculations for a range of heat demands? Or just the maximum demand, and then assume if the flow can cover that, then it will cover all lower demand situations?

We know pretty accurately what the rads can emit at various delta Ts. Unfortunately, the delta t goes up and down like the proverbial. Even if the room temp is reasonably constant, the rads are all over the place, because of cycling. Over the course of an hour, my rads might cycle between the low twenties and around 30 degrees 2 or 3 times. What is the delta T (given a room temp of 18 degrees)? 2? 4? 6? 8? 12?

Flow rates: primary reported to be 1.4 something M3/H. Secondary: no way of knowing. An the basis of the Grundfos literature and the LED settings, it is quite a lot (of something) (whatever that is).

PHE: we know where it is, what make it is, and its approximate dimensions, but that is about it.

In other words, we know almost nothing. The system is dynamic, nothing stays constant. Which forces me to ask the question: how practical is it to do a 'fairly simple calculation', given the many unknowns, and even the 'knowns' are constantly changing? Is there a way of doing a worst case scenario (say -2 ambient) calculation and then assuming that covers all other scenarios - even though all of the other parameters change?

Lastly cycling: I understand why this happens (heat pump can't throttle itself down enough, so its only option is to cycle) but doesn't this defeat the whole ASHP low flow temp always on Steady Eddy high COP design philosophy? In effect at many normal colder month ambients, our ASHPs are actually running in fossil fuel mode ie off - blast - off - blast -off and even worse, when they are in blast mode, the COP drops? They are absolutely not running in weather comp mode, the 'always on' thinking is myth. I suppose in a parallel universe you might say if the average between the off and blast LWT was around where it should be on the weather comp curve, then that is good enough. Or is it?

0700-0800 this morning:

ambient 9 -> 10 degrees

heat pump definitely in fossil fuel mode, LWT ranging between mid to high 30s and upper 40s

rads cool (not cold) to luke warm

room temp 17.0 -> 17.5 -> 17.0 which is 1.5 to 2 degrees below design temp (and room stat is set to 19.5, and is always calling for heat).

The system is not reaching design temps, even in relatively undemanding ambients.   

Posted by: @ronin92

@derek-m If this is the constraint, there are two ways in which it can happen:, because the PHE heat transfer is limiting or secondary flow rate is limiting.  In the first, I'd expect both secondary outlet and return temps will be sub-normal.  If it is the latter, secondary outlet temp will be normal (close to primary) but return temp low.

@ronin92 - these temps are not easy to measure accurately. The clip on thermometers are typically +/- 5 degrees so not discriminating enough. I use a point and click IR thermometer on black insulation tape on the copper pipe (to get the emissivity close to what the IR thermometer expects). The IR thermometer I have calibrated against objects of known temp and it seems to under-read by between 1 and 2 degrees. I can get reasonably repeatable measurements from the pipes but you have to be careful to keep the thermometer perpendicular to the pipe, if it is at an angle, the readings vary presumably because the beam gets deflected and or distorted. With these caveats in mind, the current readings with the heat pump compressor running are:

Heat pump LWT/RWT (from Midea controller):  43/41 (stayed at these values)

PHE (three IR readings taken one after the other going sequentially round the pipes over a period of about 1 minute)

primary in 38 36 39 (average 37.6) 

primary out 38 37 37 (average 37.3)

secondary out 36 35 34 (average 35)

secondary in 33 35 34 (average 34)

I'm not sure these readings match either of your scenarios. It might be possible to provisionally conclude the PHE is effective (secondary out is close to primary in). Now for the odd bit: the tops of the rads are typically in the mid 20s. About 10 mins after the above readings, on one where I have black tape on the tails, the inlet tail was 35, the centre top of the rad 23, and the outlet tail is 17 degrees. This and the secondary inlet temp doesn't make sense to me - if the rads are typically returning water in the high teens, how come the secondary inlet is at 34 degrees? Perhaps the pipe is warmed by being adjacent to the PHE primary circuit, ie it is a locally distorted reading? But this doesn't appear to be the case: the secondary returning water temp about 15in away from where it enters the PHE is pretty much what it is where it enters the PHE. 

@cathoderay There's some serious problem with the temp measurements.  Touch will also readily distinguish temps around blood heat which is what your readings claim.  Is touch consistent with temps around this level?  Secondary return should feel much cooler than secondary out if you are getting any heat supplied to the room.  Perhaps you may consider using a cheap LCD strip thermometer to verify temps are at least in the ball park?  I have an FLIR imager and I've generally distrusted it on any bare piping - copper is too reflective - I have attempted to paint the surface with a Sharpie pen but still distrust it.

@ronin92 - I do have a multi-meter with a thermometer scale using a thermocouple to take the readings. Putting that inside the insulation foam pipe wrap I get for the secondary circuit out 33-34, in 28 degrees, with the IR thermometer reading on exposed pipe with black insulating tape (my variation of your Sharpie treatment) within a degree or two. The back of the hand test also confirms these readings can't be that far out. About 5 mins later the readings were 40 degrees out, 33 degrees in and the IR thermometer on black tape gave mostly the same reading as the thermocouple. These readings are a bit more credible, there is a 5 - 6 - 7 degree difference, but it is still not a lot, and doesn't explain how the rads tops are in the mid 20s and the tails return water at around 18-20 degrees which somehow gets heated in the return pipe to around 30 (range 28 to 33 degrees in above measurements). The out/return pipes are often close but not touching, maybe an inch of air between them where visible.

About 10 minutes later - the secondary out is 36 on both thermocouple and IR, secondary in is 35, again on both devices. Aarrgh!     

@cathoderay

I will compile the information and demonstrate how to calculate the approximate flow rate in a later post.

Do all your radiators have TRV's fitted, and if so do you have a bypass valve located within your system?

It sounds like you need to get a plumber back and re-do this whole setup. No need for the plate, no need for the glycol...

Regarding your pump comments, I'm also at a loss to figure out how these so-called 'smart' pumps do what they do.

Under PWM control from an external controller, yes, but from a straight switched 230v input, no. Unless they have pressure and temperature sensors built in.

@derek-m - thanks, that would be very welcome, and I am sure others will welcome it too.

All the rads have TRVs and they are all definitely open (the heads are off, sometimes the final screwing down starts to close the valve, even when the heads is set to fully open). 

I don't think there is a bypass anywhere. The old system (oil fired) had one, in the form of a radiator (bathroom) with no TRV, with the valves always open so there was always a route water could flow through. Now the replacement has a TRV, but it, like all the others, is fully open.

The pipework is a bit all over the place, the PHE is in the airing cupboard upstairs, so the feed to the rad circuit starts upstairs, whereas in the old system the feed (boiler) was downstairs. I have wondered whether this change did something untoward, but I don't think it has, apart from making working out how the water routes itself through the pipework a bit more complex. I will have another look at it and see if there is an anomaly. It is a standard two pipe flow and return pipe setup.

The kitchen is still only at 17.0 degrees, ie two degrees below design. This may not seem like a lot, but (a) if the system is designed to achieve 19 degrees, it should achieve 19 degrees, especially when relatively mild outside and (b) 17-19 degrees is a critical range for me, 19 degrees, even 18 degrees, and I'm cosy, 17 degrees and I start to feel the chill, minor chilblains etc.    

Posted by: @hughf

It sounds like you need to get a plumber back and re-do this whole setup.

It may come to that, but I hope it won't! 

The Grundfos pump manual is as I mentioned before extremely opaque. It sort of claims it can sense the TRVs closing down (increased resistance) and adjust the flow - at least that is what I think their 'proportional pressure' claims to do. But that isn't measuring heat demand directly, and I am not even sure it (resistance) is a reliable proxy, and their graphs as to what 'proportional pressure' actually does are meaningless. This one claims to explain what is going on, but it looks more like astrology (Taurus/TRV in the ascendant perhaps?), at least to me:

After studying it for a long time, perhaps I am being a bit harsh. Perhaps 2 <- 1 is the is the increased resistance from the TRV closing, and A1 >> A2 is the proportional pressure reduction. Or maybe it is <DH2>? But there are no numbers on the axes, and so the whole diagram is arbitrary (we're still in not a lot/a bit more/quite a lot more territory) and without the pump knowing the system's actual fully open/fully closed resistance how does it know where a particular system is on that system's curve? I very much doubt all systems have the same resistance fully open and fully closed, one systems fully open may be the same as another half closed, or whatever. Or perhaps it is smart, it knows the range of pressures for a particular system, and so can set the fully open/fully closed parameters. 

The wiring to the pump is as you say a straight switched on off 240V supply, nothing 'smart' going in there, and no external sensors connected to the pump. All smartness, such as it is, has to be onboard and internal. There doesn't seem to be any way of knowing what the actual flow rate is (apart from possibly not very fast/a bit faster/much faster). But what 'a bit faster' is in L/min is anyone's guess.    

@cathoderay

I would suggest that you have a close look around your system for any form of pressure relief valve, since as far as I am aware one should not have a system with TRV's fitted to all radiators without having one installed. If you find something of which you are not certain then post details and/or a closeup photo.

Having the water pump running at high pressure could possibly be forcing open any pressure relief valve that may be installed, and hence reducing the flow through the radiators.

I would suggest setting the water pump on its lowest setting and see if this makes any difference to the temperatures around your system.