@batalto - I think they may have photographed the board, they certainly photographed other parts of the system. I have to be mindful that if something does go wrong, Freedom are all but certain to say I have blown the warranty. It may well come to me needing to take the initiative, but I want at the minimum to give Freedom (and my installer, as Freedom will only deal with him, not me) the option of doing things in a cooperative way, and that means waiting until my installer is back.

@derek-m - I've just noticed you said take reading every minute over a suitable period of time. I'll do that tomorrow am. This morning's data is above, current 20:00 room temps are: bedroom 18.5, bathroom 20.0, kitchen 16.8 degrees, ie all still below design temps. The last time the rooms were at design temps was Wednesday evening, so for the last three days the house has continuously been below design temps.  

@cathoderay Yeah you would think a 16kw and a smaller one would be the way to go but the installer normally installs only Vaillant but couldnt get them at them time so that may have something to do with the 2 X 16kw, but in terms of cost it was the cheapest quote we had (I think Midea are probably one of the cheaper brands) so thats kind of why we went for it with the hope we would never struggle for heat and if one failed we could still cope.

Hi @cathoderay

I suggested taking readings at 1 minute intervals to get a better idea of how the temperatures are changing and how rapidly. You may find later that you can extend the time interval.

Whilst measuring temperatures may shine some light onto the problem, it may not identify the true cause. Looking at the data that you collected, taking the data from 08:40 to 09:00 the Primary In ranged from 45C to 47C and the Primary Out was in the range 40C to 41C, giving a delta T of 5C to 6C, so there is some transfer of heat energy taking place. The delta T is normally controlled by the heat pump, to the value set during configuration, by varying the flow rate to meet the energy demand. This is normally in the 5C to 6C range. If the heat demand reduces, the RWT will start to increase, and hence the delta T will reduce. This reduction will be sensed and the water pump speed will be lowered so that less heat energy is being transferred. So the delta T could still be 5C, whether the heat pump is producing 5kW of heat energy or 12kW. If possible measure the water flow rate and add this to the recorded data.

The same thermodynamics are at work on the secondary side of the heat exchanger. During the time period detailed above, the Secondary In was in the range 30C to 33C, and the Secondary Out ranged from 43C to 46C. So this indeed shows that heat energy is being transferred from the primary to the secondary, but the fact that there is a temperature increase of between 11C and 16C would indicate that the flow rate in the secondary circuit is more than half of that in the primary circuit. I have attached a document for Lowara PHE's, which as shown in the table on page 5, a difference in temperature should be present between the Primary In and the Secondary Out when the PHE is working normally.

I therefore suspect that the PHE is inadequately sized, or the flow rate around the secondary circuit is too low, or a combination of both. If possible try varying the speed of the secondary water pump and obtain a set of readings at each setting. Also it may be useful during the testing to set a fixed LWT of 50C or 55C.

A further test I would suggest is that you measure the temperature of the inlet pipes to, and return pipes from, each radiator. This may indicate if the flow rate through the radiators is adequate.

That should keep you busy for a few hours.

@mookyfoo - Midea promote themselves as being value for money, based on economies of scale, they are a huge company. They also promote themselves on their long warranty, 10 years if certain conditions are met, and their everything in one box approach, which means compact installations, and installers like them because they are simple to install. Although not a well known brand in the UK, there is no reason to suppose they are not a reasonable brand to go with. In my case, their >10kW units had the definite advantage of being low profile, ie wider than they are tall, which mean they have less visual impact for listed building consent and planning applications purposes. In effect, a Midea unit got baked into these applications, and became part of the consents.

My installation was a bit of a rush, caused by the time wasted by the LAD grant preferred installers, who in the end I had to fire, and the fact my preferred installer, from the earlier GHG Voucher Scheme which was superseded by the LAD scheme, was going to be away for six weeks during the necessary installation period (Jan 2022 to 31st March 2022), to meet the LAD deadline. These short interval grants are an absolute PITA - they cause all sorts of problems, panics and troubles. In my case, the installation had to happen at the beginning of February, or not at all, because it would miss the LAD deadline, so no grant, and without the grant it wasn't going to happen.

The final twist was the combined effects of the Mitsubishi warehouse fire and the global shortage of semiconductors, which meant there were no Mitsubishi heat pumps available, one of the most common brands used in the UK, so installers started installing whatever they could lay their hands on, under the pressure created by grant deadlines (LAD and of course the RHI). That meant a rush on Midea units, meaning that in late Jan/early Feb there were no 16kW Midea units available, the unit in my quote, but there were 14kW units available, and Freedom's heat pump tool v2.4 which my installer and I were using at the time, tweaked to expose the 14kW unit, showed the 14kW was just OK, by a tiny margin, 1%. Here's the chart from the 'Your System' worksheet: 

On this basis, and knowing that if we waited for a 16kW unit, lead time 2 months or more, we were going to miss the LAD deadline, meaning the installation wouldn't happen, we went with a 14kW unit. The problem is, Freedom's Midea data over-estimates the 14kW output at lower ambient temperature. We now know from @derek-m's diligent research that the actual Midea 14kW unit output at low ambient temps is less, around 11.3 kW at -2 ambient, 55 flow (the design temp),rather than the 12.4kW figure in Freedom's heat pump tool. This difference is critical, given the estimated heat loss of the property is 12.3kW at -2 ambient... 

@derek-m - will take some i minute measurements shortly and report back. Baseline room temps at 0800 this am: bedroom 17.5, bathroom 20.0, kitchen 16.0, all below design temps of 19, 22 and 19, ambient from Midea controller 3 degrees. 

@cathoderay 

Do the Midea capacity figures include an allowance for defrosts? Mitsubishi's data includes this for its 2 deg C figures.  In that link I posted a few posts up, Graham H suggested 10% contingency should be allowed for defrosts.  Also, 3% allowed for DHW seems low.  I know max and total output aren't the same thing but most people use 10-15% of total ASHP output on DHW. 

@derek-m - OK, here are some minute by minute readings, col head abbreviations as before, csv text format. Ambient throughout was 5 degrees:

Time,LWT,RWT,Prim in,Prim out,Sec In,Sec out
09:10,42,37,40,35,29,39
09:11,42,37,41,36,29,39
09:12,43,38,41,34,30,42
09:13,44,38,41,34,30,40
09:14,44,38,40,33,30,42
09:15,44,38,42,37,31,40
09:16,44,38,41,34,31,42
09:17,44,39,44,38,31,44
09:18,45,39,43,38,31,43
09:19,46,39,42,35,31,45
09:20,46,40,45,37,31,45
09:21,46,40,44,38,32,43
09:22,46,40,45,37,31,43
09:23,46,40,42,38,32,45
09:24,47,40,45,38,32,45
09:25,47,40,44,37,32,46
09:26,47,41,44,39,32,44
09:27,47,41,46,40,31,46
09:28,47,41,46,39,31,46
09:29,48,41,45,40,31,45
09:30,48,41,46,39,32,45
09:31,48,41,44,40,32,46
09:32,48,41,46,40,32,44
09:33,48,44,45,38,38,43
09:34,48,46,43,40,39,42
09:35,46,45,44,39,41,43

and this is what they look like charted:

By 09:37 the Midea controller showed the heating had switched to DHW, so the movements from 09:32 onward are probably a reflection of the fact the system was changing over. What we see from 09:10 to 09:30 is a steady rise in all temps, some a bit more volatile than others. However, the LWT is below what it should be (at 5 degrees, it should be ~48, and it only reaches that at the end of the period), and the average of the secondary in/out, ie a proxy for all the rads averaged out, is only in the mid to upper 30s, giving a room delta t of around 15 - 18. On these readings, the PHE doesn't fare to badly, most of the time the primary in (yellow line) is close to the secondary out line (pale blue), and both the primary and secondary circuits show in/out temp diffs consistent with heat transfer taking place.

The bottom line is the rads are averaging around 37 degrees, when they should be 45 degrees or more (with weather comp and perfect heat transfer, around 48 degrees). The question is why aren't they getting up to temperature: inadequate heat pump output? Inadequate PHE heat transfer? A combinations of these factors? Or something else altogether?

@kev-m - I think the Freedom Midea figures are supposed to include defrost effects, see title to the chart in the screen grab above, and perhaps they show up as the dip in output at around zero degrees ambient, but the data is wrong. Midea's data, I don't know, but think they might, as there is a step change downwards in output between 5 degrees ambient and zero degrees ambient (seen in the charted Midea sourced data I posted earlier).

I've just taken a one off set of readings at 1035h    

Ambient now 7, LWT 48, RWT 41, Primary in 46, Primary out 37, Secondary in 31, Secondary out 42, ie close to where they were before the DHW cut in. Kitchen is 15.9 degrees, 3.1 degrees below where it should be.

Hi @cathoderay

Did you see the message that I posted yesterday evening?

Hi @cathoderay

What is the make and model of your secondary water pump, and what is the size and type of your radiators?

Do you have a 3-way diverter valve for changeover to DHW or two 2-way valves?

@derke-m - Argghhh!!!! The forum has just deleted my reply. very frustrating. 'Something went wrong with your submission' or words to that effect, and it doesn't seem to be saved in draft anywhere or caught in a cache, as far as I can see.

In brief, yes, thanks, I did see your reply yesterday and that's why I did the minute interval readings this morning focusing on the PHE, which on the basis of those readings doesn't seem to be doing too badly at the time the readings were taken. 

Posted by: @derek-m

What is the make and model of your secondary water pump, and what is the size and type of your radiators?

Do you have a 3-way diverter valve for changeover to DHW or two 2-way valves?

The pump is a Grundfoss UPM3 Auto, currently on max constant pressure setting. The rads are mostly K3s (7 K3s and 2 K2s), all new, fitted to existing 28mm/22mm/15mm pipework via lockshield valves and passive TRVs. All TRVs are currently fully open, lockshields have some semblance of balancing adjustment (its very volatile). The rads are all sized to exceed each rooms heat loss as design temps with a flow temp of 55 degrees, a compromise to get slightly smaller rads that would fit in the available space, which is often constrained, small old cottage etc. Typically, the oversizing on each room/rad is between 10-20%. Often there had to be some sort of compromise, so this is a combination of getting an available rad that will both fit in the available space and give out enough heat.   

The heating/DHW changeover is done with a bog standard 3-way valve. I recall reading elsewhere on the forum (a post from @heacol I think) that these throttle ASHP systems, so maybe that is part of the problem, which is I guess why you asking. Maybe the volts ('potential difference', temps) are getting across the PHE, but somehow the amps (current/flow) aren't happening, to use a dubious analogy? The Midea controller does report water flow rate, typically it seems to be something under 1 cubic metre/hour (it's currently showing 0.83M3/H, other parameters much as they were this morning), if that is of any relevance/use.  

Hi @cathoderay

When you have the available time, this is what I suggest that you try.

Set the LWT to a fixed 55C. When the system has stabilised, take a set of readings including on the pipes to and from each radiator.

Set your secondary pump to a fixed speed setting of III, and take a further set of readings.

Perform the above with a pump speed of II and then I.

Post the results for analysis.

What are the physical dimensions of each radiator? It should be possible to assess the heat energy being emitted by each radiator and compare the total against the estimated output from the heat pump.

Posted by: @derek-m

It should be possible to assess the heat energy being emitted by each radiator

This gets complicated. Firstly there is the nominal output, as listed in the catalogues, usually at a delta t of 50. A correction factor then has to be applied to allow for a lower ASHP delta t. These conversion factors are widely available, and range from 1 at delta t 50, is no change, to, for example, this is from Stelrad's conversion factor table, 0.513 at a delta t of 30, ie at this delta T, the rad output just over half what it would output at a delta t of 50 (or to use oversize methodology, it would need to be 1.95 times larger to maintain the same output). But what is the correct delta t? There is the design delta t, but in practice this never happens. Instead, the rads are constantly changing temperature, constantly at a lower temp than they should be, and there is the variable panel warming up problem previously referred to, so the delta t is different for different panels on the same radiator. 

I guess the idea is that by measuring the rad input and output pipe temps, and taking an average, the idea is to get the average delta t for that rad. The problem here is getting an accurate readings. Many of the clip on pipe thermometers have a hopeless +/-10 degrees error margin, and even the better ones quote an error margin of +/-5 degrees, still no where near accurate enough. The IR thermometer has significant problems with shiny copper pipe readings (wrong emissivity), though this can be mitigated to an (unknown) extent, and the rad tails are small (15mm), a difficult target. The white painted surface of the rad adjacent to the tails is another possible measuring point, but they are subject to the wide variations in surface temps.

Example: kitchen rad a short while ago (kitchen rad is the largest in the house, 600 x 1800 K3):

tails (with black tape hack): 19/28 degrees

front panel adjacent to tails: 18/18 degrees

rear panel adjacent to tails: 21/23 degrees

top middle rear panel: 30 degrees

Then 10 minutes later       

tails (with black tape hack): 18/31 degrees

front panel adjacent to tails: 17/18 degrees

rear panel adjacent to tails: 20/23 degrees

top middle rear panel: 30 degrees

During this period, outside ambient was 6 degrees, kitchen was 16.5 degrees (vs 19 degrees design) and the room stat which is in the kitchen was calling for heat.

The rad is nowhere near the temp it should be. The relative drop across the tails is sort of OK (9 to 13 degrees), and if we accept the tail values are reasonably accurate for overall average temp (and ignore the actual top middle rear panel reading), then the rad is at around 25 degrees, given a delta t in the kitchen of around 6 degrees, when it should be of the order of 25-30 degrees! No wonder the room is not getting up to design temperature. Even with the top rear panel reading (which is a very local reading, not representative of the whole rad), the delta t is still only around 11, half what it should be.

The kitchen's calculated heat loss is 2200W (three solid stone outside walls, three windows, glazed back door, high air change count). At a delta t of of 50 degrees, a 600 x 1800 K3 should output a whopping 4525W (it is a huge rad). At the design delta t of 33, it should output 2629W, making it 19.5% oversize. At the current delta t of 6, it is off the Stelrad conversion factor table (lowest quoted delta t is 15). Assuming the delta t/conversion factor relationship is linear (it is very close to linear over the quoted range), and extrapolating to a delta t of 6, the conversion factor approaches zero. My 4525W at delta t 50 effectively becomes a dead weight hanging on the wall at a delta t of 6.    

The bedroom and bathroom are also still below design temp, 17.5 vs 19 and 19.0 vs 22, making this the fifth day in a row when design temps have not been met. The house also feels a little chilly in places this morning. For this privilege, I used 334kWh over the last week for the heat pump (readings from its dedicated meter) which at the new current average price cap price of around 28.34p/kWh or unit = £94.66 plus standing charge plus VAT, ie pretty much £100pw.

Current (0915h) flow/return temps and top middle rear panel rad temps:

Midea controller: Ambient: 7; LWT 47; RWT 41

PHE pipework: Primary in 45; primary out 40; secondary in 30; secondary out 44 

Rads 22/18/26/41/39/27/34/30/32, crude (not weighted for rad size) avg = 30 degrees. The first 2 low readings are the reluctant bedroom rads.

Thoughts: LWT is OK, about where it should be with weather comp; PHE secondary out is OK at only 2 degrees less than primary in, however secondary in/out average is 37 degrees, but the rads are only averaging 30 degrees. The 44 degrees coming out on the PHE secondary circuit is not appearing at the rads.

At 7 degrees ambient, 50 degrees flow, the Midea 14kW unit should be able to meet the load (Midea claim it can output almost 15kW in these conditions, vs a heat load or around 12kW), so at this ambient temp the heat pump should be adequate (unlike around zero degrees ambient and below, when it is definitely inadequate) but the rads/rooms are not getting up to temperature. Not enough watts are getting through from the heat pump to the rads. Is this a flow problem: the circulating water in one or both circuits isn't carrying enough heat (watts)? How do we test this? How do we correct this (if it is the problem)?

I will put the secondary circulating pump onto max constant flow (I think Grundfos call this 'constant curve': 'Constant curve: The pump runs on a constant curve, which means that it runs at a constant speed or power'. They need to get their plain English sorted out. The other two options are Proportional pressure and Constant pressure), was on max constant pressure, and see what happens.   

@derek-m - here are the rad sizes I forgot to add to my last post:

Radiator Type and Size (H x W)

K3 750 x 1400
K3 600 x 1200
K3 600 x 1800
K2 750 x 600
K3 500 x 1600
K3 500 x 800
K3 500 x 1000
K2 500 x 1600
K3 900 x 400