@derek-m - I will have another close look. The only other connection to the rad circuit is the connection to pressurise it (it is a closed system) and that does have a pressure relief valve but that vents to externally (covers the inside of the airing cupboard with water when it opens, somehow the plumber managed to pressurise the system to the point where it opened when he was pressure flushing the system - but he was only using mains pressure, but it looks as though the pressure relief valve is set to 3 bar (it is difficult to read the figures on it as it is upside down in an awkward to get at place), so maybe the mains  pressure was enough to do it). In the meantime, I have just had a look at the pipe runs, and I can't see a bypass, deliberate or otherwise. The new PHE is connected to the old boiler feed and return pipes to and from the old HWC in the upstairs airing cupboard which route the water to and from downstairs where the old boiler was and tees there route the water to and from both the downstairs and upstairs rad circuits, if that makes sense. The only thing I am not sure about is two pre-existing 15mm pipes teed off the PHE flow and return pipes that disappear under the PHE and don't appear to emerge again ie they are probably capped off. I think they may have been part of a very old bypass when I had a solid fuel boiler, which was 'always on' and so had to have an 'always open' bypass. They are not very warm either, if they were an open bypass they should be warm (at PHE outlet water temp). I may be able find some old photos that reveal more.

I will also set the circulating pump to the lowest 'proportional pressure' setting and see what happens.

@cathoderay

Are you ready for the maths?

The Specific Heat Capacity of water is approximately 4200 Joules (J), which means that it takes approximately 4200J of energy to heat 1kg of water by 1C, in the temperature range of a central heating system.

Since 1kg of water in this temperature range has a volume of approximately 1 Litre, then we can substitute Litre for kg.

1W = 1J/sec, therefore the amount of energy supplied by 1W over a 1 hour period is 3600J. 1kWh is therefore equivalent to 3600000J or 3.6MJ.

Applying the above to your system, which has an antifreeze mixture in the primary circuit, lets assume that the mixture has a specific heat capacity of approximately 4000J.

So if your heat pump is producing 10kWh of heat energy and raising the mixture temperature by 5C, what will be the required flow rate?

10kWh = 36MJ of energy, and if we assume that it takes 4000J to raise the temperature of 1 Litre of mixture by 1C, then to raise the temperature by 5C will require 20000J of energy.

10kWh of energy can therefore heat 36MJ/20000J = 1800 Litres of mixture.

But this is over a 1 hour period, so the flow rate in Litres/min would be 1800/60 = 30l/min.

If your controller claims that the flow rate when the heat pump is running is 1.4 cubic metres per hour or 1400 l/hr, then the heat energy produced when it is operating will be approximately 1400l/1800l x 10kWh = 7.78kWh.

But if your heat pump is running for only 50% of the time, then the heat energy produced will more likely be in the region of 3.89kWh.

If the heat demand of your home is greater than this under the present operating conditions, then it will probably fail to reach the desired temperature. I would suggest checking against the heat loss calculations for your home.

@cathoderay

Here are a few things to check and confirm.

Is the pump in the secondary circuit running continuously or switching on and off with the heat pump?

What is the size and type of each radiator, and their total heating capacity at a DeltaT of 50C between water temperature and indoor temperature? How does this compare with the total heat loss calculation at the specified outdoor air temperature?

Try measuring the temperature of the flow and return pipes on each radiator with the heat pump running, to try to identify where the heat energy is being dissipated.

Posted by: @derek-m

Are you ready for the maths?

Yes! But I'll come back to that if I may. Answers to the questions in your next and post:

Posted by: @derek-m

Is the pump in the secondary circuit running continuously or switching on and off with the heat pump?

Continuously, as long as the room stat is calling for heat, which it is almost all the time. Sort of makes sense, otherwise the system would stagnate.

Posted by: @derek-m

What is the size and type of each radiator, and their total heating capacity at a DeltaT of 50C between water temperature and indoor temperature? How does this compare with the total heat loss calculation at the specified outdoor air temperature?

They are all K3s except for two K2s, one of which is in the downstairs loo, the other on a long narrow landing where a K3 would intrude on the walking space. They are sized to provide >100% (next largest size up that achieved this) of the heat loss for each room at -2 ambient at the design delta T of 30 or just above (ie LWT 55 (@-2) rooms at 18-21 degrees). The thinking was a high-ish LWT meant see could use slightly smaller rads, to fit in the limited space (small old cottage, not a lot of wall space). 

Their total heating capacity at a delta T of around 30 (ie the design delta t, around because not all rooms have exactly the same design temp) is 13.03kW (the spreadsheet uses ASHP delta Ts, not fossil fuel delta Ts). If I adjust the delta T to nearer 50, the total heating capacity goes up to 23.27kW. This seems about right to me, very roughly going from fossil fuel delta Ts to ASHP delta Ts halves the output (which is why rads sized for fossil fuel systems are almost always never large enough, unless they were huge), so going the other way would very roughly double it. More precisely, the output correction factor (from Stelrad's tables) used in my spreadsheet, assuming delta T is 1, is 0.513 for a delta t of 30 (ie multiply the delta T 50 output of the rad by 0.513 to get the delta T 30 output. By the same token, the oversize factor is 1.95, the rad needs to be about twice as large (in area, I think, and this can be further tweaked by going from K2s to K3s, but get complicated, that's why I stuck to using the output correction factor to adjust the output in kW.

The design heat loss calculations at -2 ambient show a total loss of 12.3kW, this should in the ideal world of design calculations be met by a total radiator output of 13.03kW (as above, at a delta T of 30), making the house Micawber positive in heating terms. For those not familiar with Micawber positive, think of Micawber's advice about income to the eponymous Copperfield in David Copperfield.

But - the design delta T only really exists on the spreadsheet. Most of the time it is not -2 outside, so the weather comp lowers the LWT, and so the delta T drops. I sort of assumed this was balanced by a reduced heat loss (hypothetical eg 7 degrees ambient, total heat loss drops to say 7kW, matched by a heating system running as a delta T of 20 and putting out 7.2kW) - thought this was the whole idea behind ASHPs and weather comp. On top of that (the system is deigned to run at a lower delta t at higher ambients) there are the anomalies that crop up in my measurements, when the delta T is clearly even further reduced.       

Bypasses - I've had another look, and can't see anything that might be one. The pressure relief valve is presumably what is supposed to deal with excess pressure, but quite how an ASHP powered system can reach reach boiling point is beyond me.

Setting the circulating pump to lowest (proportional pressure) setting for the last five hours or so: rooms and rads a bit cooler, kitchen is now 16.5, was 17.0, top of rads high teens to low 20s. Midea LWT 39, PHE output 36, PHE input 33 ie same 'nonsensical' result, PHE input higher than rads, only all temps a bit lower. I'll put the pump back onto the higher setting.

Posted by: @derek-m

If the heat demand of your home is greater than this under the present operating conditions, then it will probably fail to reach the desired temperature.

Coming back to the calculations, this may well be the case. If we assume the heat loss is linear, ranging from 12.3kW at -2 ambient, and dropping to zero at say 20 ambient, then the loss at current ambients (8 to 10 to 12) is between 4.4 and 6.6kW. If the primary circuit is only delivering 3.89kWh...

I'm struggling a bit with our old enemies kW and kWh and assuming that in this case they are sort of interchangeable, the kW value being the 'instantaneous' loss/gain, and the kWh being that loss/gain over an hour. I thought I'd cracked these two interlopers: my 2kW heater is twice as powerful as my 1kW heater, but is also consumes twice as much energy over an hour (2kWh vs 1kWh), in the same way that my 20HP engine is twice as powerful as my 10kW engine, but it burns twice as much fuel (2L/h vs 1L/h). The kW value is the power, what can be achieved (or lost, in the case of heat loss), the kWh value is the amount of energy used/burnt at the respective power over a period of an hour.

Assuming the last paragraph is valid, ie we can match the 'power' of the heat loss (kW) to the energy supplied (kWh), then given the Midea pump seems stuck at 1.4 cubic metres or 1400L/h, then it is designed to fail at anything but the mildest temperatures. This assumes the 1.4M3/H figure is a real reading from a real sensor, not some app making it up. Basically, I think what we appear to be saying is regardless of whether the heat pump can produce enough heat, the delivery system is incapable of delivering it to the house. 

This seems absurd - all the heat loss/ASHP sizing calculations assume the heat pump can always deliver what it produces. Here we appear to have a situation where the bottleneck is not in the potential production of heat, but the actual delivery of the heat from the heat pump to the PHE.

The other major unknown is the flow rate in the secondary radiator circuit: can it deliver the required amount of heat from the PHE to the rads? I'd sort of assumed that was the likely bottleneck, but if the heat pump to PHE is inadequate...

And how does all this sit with the actual temperature readings? As noted earlier, although the circulating temps (primary and secondary) are cycling through large ranges, the average of the top and bottom of the primary range is about where it should be with current ambients, based on the weather comp curve (around 43). Is it possible but that the water reaches the right temp, but because the flow is too sluggish, not enough heat gets transferred?

It is all terribly confusing, apart from the most basic of all readings, which is that taken from the thermometer in my kitchen, which shows the room is too cold.    

@cathoderay

As you state, in theory it should all work correctly, but lets try to apply some 'real World' aspects.

The manufacturer's performance data will, I suspect, be based on using water in the heat pump. So using a water/antifreeze mixture will reduce its possible output capacity.

Your system is further complicated by having the PHE installed, which will require the heat pump to produce a higher LWT to achieve the correct temperature at the radiators. This could possibly require a LWT up to 5C higher than without the PHE, which of course will reduce the overall COP of the system. As a trial you could try raising the LWT on the weather compensation by 5C at the -2C end, and then monitor the temperatures around the system over a number of hours.

For the RWT on the secondary side of the PHE to be higher than the water temperature returning from the radiators, would tend to indicate that some warm water is finding its way in from somewhere and is mixing with the colder water. I would suggest that you check that the DHW changeover valve is fully closed, and is not allowing flow through that is mixing with the colder water. Have a feel round the pipework in and out of your DHW.

Kev bought some aquarium type temperature sensors that he used quite effectively to fault find on his system. I feel certain that he would provide the details should you decide to invest in some.

As a final resort you may wish to consider a clamp on style flowmeter, some of which are not unduly expensive.

Posted by: @derek-m

@cathoderay

Kev bought some aquarium type temperature sensors that he used quite effectively to fault find on his system. I feel certain that he would provide the details should you decide to invest in some.

Go to Amazon an search on "Thlevel 4x Digital LCD Thermometer". The little rectangular ones with the probes, 4 for £10.99.

https://amzn.to/3TmWajn

Calibration was via confirmation that my own body temperature is consistent with human life. They all read the same to within 0.5 degrees or so. Just tape the probe on the bare copper pipe and cover with some pipe insulation.   

@derek-m - yes, we have to stick with physics (Newtonian, not quantum) and the real world observations and data, and not forget the Holmesian dictum that when you have eliminated the impossible, whatever remains, however improbable, must be the truth.

Somehow, despite saying I was going to reset the circulating pump to PP ('proportional pressure') max yesterday evening, I forgot to do it, so the system has been running on PP minimum for getting on for 18 hours. Result: a house that feels cold, rads cold to the touch, kitchen 15.5 degrees (3.5 degrees below design temp). From this I think we can conclude the secondary flow at PP min is definitely inadequate, but that doesn't tell us anything about whether the secondary flow at PP max is adequate or not. It could be the case that PP min is 25% of what is needed, PP max is better but only 50% of what is needed.

I agree about the PHE. Freedom as usual have bodged over this, for example until recently their instructions made the PHE (or LLH) mandatory, but their heat loss/radiator calculator does not take into account the inevitable temperature drop across the PHE, instead it assumes the temp arriving at the rads (which determines the delta T) is the same as the temperature of the LWT leaving the heat pump. As my system was deliberately designed with a highish LWT (55@-2) to allow the use of slightly small rads, this means I will probably have to set the curve to 60@-2 or even more to get the intended design temps. Even then it may not happen. Bye bye COP, hello big bills. Greener energy? I don't think so. How is all that electricity I am going to use generated?

The higher secondary RWT temperature compared to the temperature of the rads is remains a mystery. I don't think the two way diverter valve can have anything to do with it, it is in the primary circuit, diverting the heat pump output to either the HWC or the PHE. There is no physical connection (conduit) between the primary and the secondary radiator circuits.      

I've just done a quick whizz round taking IR readings on the PHE outlet/inlet and all the radiator tails. The readings, bearing in mind this is a dynamic system where temps are constantly changing, were: PHE out 37, PHE in 34, rad ins range from 16 to 36 (the 16 was on a cold rad I can't get to warm up at all at the moment, and the 36 was an exception, most are around the low 20s, rad outs range from 13 to 21 (the 13 is the room with the cold rad, and the room is also at 13 degrees, 5 degrees below design). Given the IR thermometer and the independent thermocouple/multi-meter thermometer give readings within a degree of each other, I am reasonably confident these whizz round readings aren't that far out, and even if they are a bit out absolutely, they are almost certainly relatively meaningful, as in higher reading equals warmer pipe. Back of the hand estimates of temps concur.

Thus we still have the higher PHE in temperature than rad out temperature anomaly. None of the rads is sneaking hot water back to the PHE, in fact they are all at least 10 degrees cooler than the PHE in temperature. 

I think I have to say the only way this can happen is if there is some sort of invisible bypass. There is no way water in a pipe at X degrees can magically increase its temperature to X + 10 degrees unless somehow, in some shape of form, extra heat is added. This would appear to be happening downstream of the point in the old boiler space under the stairs where the tees to the upstairs and downstairs flow and returns are located. Here the IR readings are: flow and return from and to the PHE (before the tees): 26/26; flow and return to and from the upstairs rads (after the tees) 25/24; flow and return to and from the downstairs rads (again, after the tees) 27/18. The return to the PHE is warmer at 26 degrees than any of the rad out tails. The downstairs rads flow being hotter than the flow from the PHE (1 degree hotter) is probably measurement error, but the other oddity is the upstairs rad circuit returning 24 degrees vs the downstairs rads returning 18 degrees (which is more in line with the rad outlet tails, but it is still warmer than the downstairs rad outlet tails). Note all these boiler space readings are on dull copper pipe, no black insulating tape in place.

All in all this would seem to suggest the bypass is more likely happening upstairs (the rads + the mystery bypass) return water that is pretty much the same temp as the water sent to them, flow 25, return 24. The downstairs rad circuit does extract some heat, flow 37, return 18.

Luckily the upstairs pipe runs are marginally more accessible than the downstairs runs. The latter are all buried in the solid concrete flow. I will dig out the old photos I mentioned earlier and see what i can see before I start pulling up carpets and floor boards. In the meantime, can anyone think of a way of testing for a bypass in a circuit without doing radical surgery as in opening up the circuit at various places and connecting hoses and seeing what water goes where?

@kev-m - those Thlevel 4x Digital LCD Thermometers look very neat and are very affordable. I will order some if we establish a test they will perform that I can't already do with the kit I already have.                   

@cathoderay

Measuring the temperature of pipework isn't the most useful way of finding out where the heat energy is being distributed, since pipework even with a low flow rate will probably achieve the same temperature as pipework with a high flow rate.

As you are no doubt aware, the water in your system will take the route of least resistance, so will probably be more inclined to flow through the downstairs radiators rather than any upstairs ones. This is why the lockshield valves are used to balance the flow between radiators. I would suggest that you check the valves on the cold radiator to ensure that they are not fully closed.

I would also suggest tracing back through the pipework on the return water to the PHE.

Please try the following procedure. Set your heat pump controller to a fixed LWT of 50C and your secondary pump to maximum, and warm up your home to the desired temperature or slightly above. Starting at the return water inlet to the PHE, measure the temperature of the pipework. Then work your way back along the pipework wherever it is accessible, noting the temperature as you go. If possible sketch out the layout of your system. When you get to a point where there are a number of pipes forming a junction, measure the temperature of each pipe in turn, as far away from the junction as possible. You are obviously searching for the warmest pipework. Continue working your way back along the warmest pipework until you locate a radiator or hopefully a bypass valve. In particular look for pipework that appears to connect between the flow and return pipework, since a bypass valve could be hidden below insulation.

Please let me know what you find.

@cathoderay

For kW and kWh, kW is the rate at which energy is being used, whilst kWh is the quantity of energy used.

Think mph is kW, whilst miles traveled is kWh.

Posted by: @editor

I noticed something this autumn while running weather compensation. With flow temperatures of 32-37C, the rads aren't "hot" but they are doing a great job keeping rooms cosy and comfortable. What I've noticed is that the rads are warm on top and noticeably cooler at the bottom. This applies to all rads in the house. They've all been bled, so that's not a contributing factor. When running at a set point of 40C or 45C, the full rad is warm. Has anyone else noticed this? 

Hi Mars,

I have had a further thought as to what may have been occurring in your system when running on weather compensation during milder conditions.

Milder weather means lower heat loss, and hence lower heat demand. Whilst in weather compensation mode your heat pump will lower the LWT, so that the heat emitters supply less energy to your home, your heat pup will still be trying to maintain a DeltaT of approximately 5C, which means the same amount of heat energy would be transferred if the water flow rate was maintained. To accommodate the lower heat energy requirement, the internal water pump within the heat pump unit will slow down. This in turn means that the water flow rate through the heat emitters will also be slower, so giving the water more chance to cool as it flows through the radiators.

Posted by: @derek-m

Think mph is kW, whilst miles traveled is kWh.

Yes, I find that analogy works for me too, though I add fuel burnt:

Think mph is kW, whilst miles travelled and fuel burnt is kWh.

I find it further helps to remember fuel and kWh are what we pay for, ie the energy, kW (and HP) and mph are something that we enjoy/benefit from.

I have found some photos of the upstairs pipework when it was visible as the floor boards were up, and I really can't see anything that even approaches being a bypass. Not every inch of pipe is visible but the bits that aren't are very unlikely to have any thing except straight pipe eg where they pass through a wall. 

The upstairs pipework starts in the middle of the house, and has one branch going to a bedroom, with a tee on the way to the bathroom, and another branch going the other way with a tee, one side of the tee supplying the landing and the other two bedrooms. That's it! 

Another possible oddity I have noticed is the primary feed from the diverter valve to the HWC is always hot, sometimes hotter that the feed from the diverter to the PHE, even when the system is definitely in heating mode. I did wonder whether the HWC was acting as a bypass on the primary circuit, and because it most of the time contains hot water, the primary circuit hot water stays hot, but on balance I think the pipe is hot because there is only about 8" or so between the diverter valve and the HWC causing the hot water in the tank to heat the pipe by conduction. I think it is probably a red herring.

The cold rad's lockshield was more closed than I remembered it being, opening it a bit has had that radiator warm up (to blood heat).

Kitchen is still 16.0 degrees, 3 degrees below design. Yet another thing I noticed is that the flow rate reported by the Midea controller is not a constant 1.4M3/H, during the off part of the cycling it drops down to 1.2 or thereabouts. I can't see how that makes sense, if it has cut out because it is too hot, then it should keep pumping to cool things down! But it does establish that the controller can vary the flow rate, at least downwards.   

I'll wait until Monday before I make any change to the weather comp curve because I am also monitoring weekly energy use, first thing Monday morning to first thing Monday morning, and if I make a change in between times then I lose track of what energy consumption happens with the varying setups. Last spring, I think a period of constant LWT and/or constant max flow on the rad circulating pump caused energy consumption to skyrocket.