It’s not evident that there’s any integration between the heat pump and the UFH controller - i.e. I don’t see any ‘Call for Heat’ signal to the ASHP from the UFH in the PDF Peter pasted with some plumbing and wiring diagrams 5 days up the thread. There’s also no mixing valve shown for the UFH.
I’ve formed a view that the generic NuHeat design isn’t expecting Passivhaus levels of heat demand so it isn’t especially well adapted to Peter’s use-case.
Peter’s currently trying his setup with the UFH target temp set higher that it’s ever likely to achieve, so the UFH pump runs all the time, circulating water at whatever temperature the Weather Compensation on the heat pump decides to generate, and the Flow and Return temps are being picked up by my monitoring, so we’ll see what that shows over the next few days.
@prwv - you might want to consider adding a Room sensor RTS 40 NIBE (I have no experience of that supplier but that’s a handy link to the right bit of kit; should be the same as your existing outdoor sensor) so the heat pump has visibility of room temperature which can be used to tweak its WC logic since - like me - you have the “problem” of passive solar gain on cold-but-sunny days which means much less heat is required than the outdoor temperature would imply, based on the WC curve alone.
The 2 parts (HP to buffer and buffer to UFH) should be independently controlled. UFH controlled by room temp and HP by buffer temp.
[edit]
I’d missed the PDF.
Fundamentally, if the UFH is running and the HP isn’t, there is nothing to stop the return from the UFH going through the HP. I’d suggest this is a flaw. If the HP is off, the circuit should isolate so the UFH water simply circulates through the buffer tank.
There should be a blending valve on the UFH that controls the temperature of the water in the slab. This then slowly drains the stored heat from the buffer until it hits roughly the target temperature of the UFH water temp. HP then comes on, charges up the buffer tank and switches off again when that is full.
There also seem to be multiple pumps working on the same fluid. One ‘charging’ the HP, one charging the top of the DHW tank and one charging the UFH - all operating on the same fluid stream. Crazy.
No it is not the same fluid. As far as I understand, it is working as a thermal store where the mass of the water harvests heat to the coil which is an extension of the DHW loop.
Now this was never envisaged as a space heating source because we did not think that was required. Granted all houses, even a PassivHaus has a heat demand, but we thought ours was minimal. This loop, which was bought by us directly, not a NuHeat purchase was put in as a “just in case” scenario before the slab was poured. I do know cases where people have installed multiple loops and never connected them. Because we were only seeing this as running occasionally and at a very low flow temp of say 25d, we made it a simple, single zone and did not install a TMV on the loop.
However, yes, the two systems are not “joined up” and we have clearly let this rumble on. No excuse, but I am away a lot and not tackled this earlier. Modern electricity prices have severely sharpened our mind! The drawings I posted were all I could find electronically on my laptop out here. They were more for the electrical first & second fix thus the plumbing elements are slightly cut down. The main omission is the second, parallel 200L DHW cylinder.
David has explained above what I am putting in practice short term and I concur that NuHeat did not particularly appreciate the PH demands or requirements.
I subsequently found your PDF I think says it must be.
To solve the issues, I think:
You need to have a valve on the UFH that limits the temperature of the water entering the slab - this blends the return water with a small amount of water from the buffer tank so returns a small amount of cooler water to the buffer tank.
A valve to prevent the return from the UFH flowing through the HP unless the buffer tank is below a set temperature
A couple of sensors to tell you that (the flow temp from the buffer tank should do).
A fail safe so the if valve at 1 above is closed the Shuttle valve is open to A
When the UFH is on, the water flows though at the HP flow temperature and returns not much colder to the HP and the HP tries to run if the UFH is on, but cycles as the DeltaT is too low.
(I am not a heating engineer - not that that is any guarantee either way!).
[edit]
It does presume you have a separate pump for the UFH as part of the manifold. If not probably change the manifold.
The best I can do in lieu of drawings is this photos that my partner took earlier today. I just phoned NuHeat and they confirmed it was a coil inside, but I see what you are highlighting and will check again
You need to understand exactly how it is setup. If it isn’t as per the diagram, you need to know that and get a diagram of the actual setup, else the rest is just guesswork.
[edit]
Are you sure you asked about the Buffer Tank? The DHW bit will be a coil.
[edit2]
This is the sort of UFH manifold I am suggesting. The valve at the bottom of the pump controls the temperatre of the water entering the UFH pipework
I very much did ask about the buffer tank. I agree you have spotted something that is not how I had always understood. I returned to NuHeat and the man in technical could not explain it either. He, like me , thought they were very much independent sources and the water did not meet, just like a DWH cylinder where the potable water is nowhere near the inhibitor filled water that warms it. I have always been of that view and apologise that I entered into this forum with a misguided idea, although my problem remains! It does look as if the return goes into the buffer as well as the return to the HP. I am awaiting an email to call back to a HP technician because he said, whilst not understanding the reason , " we have always done it in this way" I will update when I know more and I appreciate your ideas above.
A little difficult to tell if it plumbed as per the diagram above. If you can work that out, it would be a starting point.
Blue - looks like return from UFH, Red - Flow from HP via shuttle valve (circled) to charge buffer tank. Behind red line is flow to UFH Yellow - if as per diagram, this is return to HP.
Yes I absolutely agree with your arrows.In fact the yellow joins the return from the other cylinder - you can see it just below the UFH return. It looks exactly like the NuHeat drawings so the return from the UFH loop goes back into the buffer tank coil as well as combining with the returns of both HW cylinders. Well spotted, thank you. I am still waiting for an email from NuHeat to let me speak to a HP tech support person to explain this because it is absolutely contrary to how I ever understood it to be.
Is it possible that the charge pump on the return (in photo) is not running , via software command from SMO20 when HP is open to buffer by the red diverting valve, so that route is “closed” and opens/runs when the diverter valve is doing DHW and the buffer heating circuit would not be on so not pushing water out? if so, would the purpose of the loop is some sort of safety feature?
That’s interesting. This winter I have been using about 25kWh/day to run my NIBE (F1155 12kW) Ground Source Heat Pump; hot water and heating for a 327sq/m house. I thought it was running high. I have 3x Tesla Powerwalls (40.5kWh) and on Octopus Intelligent, so paying 10p/kW between 11.30pm and 5.30am. Not bought a single peak unit since having batteries installed in September 2022. I have 9kWp of solar PV plus another 8kWp going in this month. So looking forward to sunnier days at 56° north!
I’d anticipate that the charge pump not running, would still leave the fluid free to flow round the loop as well as through the buffer. Depends on what the fluid feels is the path of least resistance, but relying on that seems a bit off.
However, and this is a key question, you said the thermistor tells the HP to turn on. What temperature is that reading - which part of the system?
If you switch the UFH on, does the HP and Charge pump turn on at the same time? If not, at what point does the HP/Charge Pump come on - what triggers it?
If it doesn’t come on immediately, then fitting a new manifold to the UFH may solve a large part of your issues as it will limit the temperature of the water in the slab, and slow down the rate at which the heat in the buffer is drained.
Agreed. Looks like the buffer tank is the one from this NuHeat webpage Buffer Tanks - How do they work with heat pumps? which talks about “To increase volume” and “Hydraulic separation” - so it’s just a tank, with four tappings. I have a Buffer Tank on my system and it’s just adding volume, to stop the heat pump short-cycling. (I reckon they would call it a Thermal Store otherwise.)
@prwv - it’s not entirely clear from the photos whether the bottom-right (blue) tapping on the Buffer Tank is connected or blocked-off. One of the photos and the PDF implies blocked-off - is that indeed the case?
@borpin - I always understood it was undesirable to have a blending value on UFH fed from a heat pump - the logic being that you don’t want to make the heat pump work hard to generate a higher temperature just to blend it down again before the UFH (which is Peter’s only emitter) sees it. My own UFH is fed from my heat pump without any blending.
I have the benefit of seeing the monitoring from Peter’s system which is showing a Delta T of 5 degrees across his UFH loops and a return temp of about 25 degrees which I don’t see the heat pump having any problem with,.
Good morning @dMb and @borpin
Thank you both for your advice and ideas yesterday. I had to work late into the evening so was unable to respond until now.
Before the work, I eventually got to a HP “expert” at Nu-Heat after two previous conversations with others from the technical department. He confirmed Brian’s analysis of the installation. He also told me that when the outdoor sensor reads 18d over a 24 hour period, the controller puts the HP into a form of Summer bypass to save it cycling.
Until now, I have been under the impression that it was a closed coil feeding the UFH loop, much like a thermal store with the HP warming this fluid in the loop. One of three thermistors linked to the SMO20 control unit is attached to the buffer tank (you can see it in the photo, about midway vertically towards RHS of the tank). I presume this reports back buffer tank temp which is referenced to the WC curve to tell the HP when it needs to run to maintain the temp within the UFH loop. Brian, you are absolutely correct that the fluid within the UFH loop is combined with the fluid running through the HP/buffer tank. As we all previously discussed, there is no joined up writing between the systems and Nu-Heat have “designed” a system based on heat calcs they may have done for the size, layout etc of the house, although probably never quite appreciating the fabric and the PH element on sunny/cold days where the UFH loop is really not required. I thought that its overall use was minimal all the time, but with my partner on her own in the house for the 3-4 days we turned the heating off via the software earlier this week, the interior temp was clearly creeping downwards by perhaps as much as 0.4d per day, although had probably bottomed out by then.
I am now starting again in my head as the system is fundamentally different to how I thought it was. For those here who have gone through the process of a self build, you will appreciate there are an awful lot of things to take in and manage, whilst juggling work to pay for the build. I concentrated a lot on electrics and the automation as well as the internal plumbing of home runs from every destination, but missed or misunderstood this one completely. I apologise if I have wasted your valuable time and energy, but am still keen for any advice, as I still believe the running costs are excessive. My partner should have done another midnight HP meter reading to give past 24 hours, but she will not be awake yet to report back. Yesterday was 9KWh, up from3-4 when heating was off. The house is back up to normal temp I think, helped by mild weather outside.
The Nu-Heat technician also told me that his commission sheet (a non electronic version will be at home) shows they left it on WC curve 9. This would have been a target temperature of 41 at 0d outside. This would have been far too much for my house I believe. I recall calling them back because, whilst I was outside a lot on site clearing duties etc, I was surprised how much the compressor was running. I imagine they dropped the curve back, because it was at 3, but I do not now for certain. A one year service by a NIBE approved agent moved the curve to 6 (34d) because it was his way of reducing the amount of start ups the software showed, which I guess equates to short cycles. That is where it has stayed but currently it is back to 3.
There’s no need to apologise for any misunderstandings Peter. You’re far from alone in having out-sourced an important sub-system of a new-build to people you had good reason to believe were experts in their field - and now it’s evident that’s maybe not the case.
From your PDF and your partner’s recent photos (and the knowledge I and others have of how NIBE systems work) we can all understand what’s been installed, how it was intended to work and how it’s actually behaving.
I’ve got a broadly-comparable system to yours:
NIBE heat pump of a similar age with effectively the same controller
Mine is an F1145 GSHP with the controller built-in
DHW and Heating outputs
My Heating circuit has a Buffer Tank to add volume and the UFH uses the same inhibitor-protected water as the rest of the system; I do also have a separate Low-Loss-Header for ‘hydraulic separation’ whereas your Buffer Tank performs that function - or at least attempts to
I probably have similar heat loss - e.g. yesterday I used 8.2 kWh of electricity input across space heating and hot water, with 31 kWh of heat output for space heating
I’m on Curve 5 which is working well for me. As noted previously, I have the “problem” of passive solar gain on cold-but-sunny days so the Indoor temperature sensor effectively “lowers the curve” when the house goes over its target temperature, having the effect of pausing the heat pump cycles until the house cools back nearer to target. Sample graph from MyHeatpump app shown below:
This sensor must be what NIBE Uplink calls “external flow temp.” with parameter ID BT25. NIBE’s manual for the SMO20 confirms that BT25 is connected between terminals 8 and 10 on the X2 terminal block and the Nu-Heat PDF shows what they call “Temperature sensor external supply line” at X2: 8, 10 so that all matches up. (Their PDF shows it on the pipe beyond the Buffer Tank rather than on the tank itself, but hey.)
These NIBE controllers effectively work like this:
They read the Outdoor Temperature Sensor
They use the configured Weather Compensation ‘Curve’ to look up the Target Flow Temperature which is predicted to offset the heat loss at that Outdoor Temperature (and they report this via NIBE Uplink as parameter ID S1, the “calculated flow temp.”)
They then cycle the compressor On and Off so the actual flow temp (measured as BT25) averages out the same as the target flow temp (S1)
The actual flow temp will be lower than the target when the compressor is off, so the controller counts how much lower the temp is (how many degrees) and for how long (how many minutes) and keeps a running total of these “Degree Minutes” (reported as NIBE Uplink parameter DM). When this number goes too far negative (a configurable setting at location 4.9.3 in the controller) it starts the compressor.
Then, with the compressor running, the actual flow temp will be higher than the target, so the DM value gets less negative - and when it reaches zero the compressor shuts off. There’s still heat in the pipes from the heat pump so the DM value goes positive for a bit and then starts dropping again. When it gets back to the configured value the compressor starts again.
Too many start-stop cycles shortens the life of the compressor, so from a compressor lifetime standpoint you want fewer, longer cycles. That means having the 4.9.3 value set to a big, negative number; my system is set to -500 which today (+8 outside) is giving me a 45 minute-long ‘run’ about every 4 hours. For me that’s fine, since my house is very well insulated and has a high thermal mass so the internal temperature has been rock-steady at 21.1 all day. With less insulation and less thermal mass the temperature might have started to drop so shorter but more-frequent ‘runs’ might be preferred.
Edit: I’ve just spotted Peter’s actual heat pump shows up as an F2040-8 in NIBE Uplink as a ‘slave’ of the SMO20, and it does have inverter control (unlike my older GSHP unit, which is only ever On or Off). Time to amend by NIBE Uplink script to grab all the extra parameters…
That would be true if you were running it directly. As there is a buffer tank, that should act as,well, a buffer.
I also don’t see the logic at running the HP at such a low flow temp unless that is its most efficient state?
If this was fine, the HP would not cycle! For right now, but what about in the shoulder months? You could easily get a much reduced delta.
Personally, I suspect it simply cannot modulate down low enough.
I don’t monitor my UFH temps in all honesty. I am thinking of replacing the TRV with a webMQTT enabled TRV so I can control the flow temp - increase to heat quicker and decrease to prevent overshoot (which honestly I don’t get much of as I control to a 1/10th of a degree). I also have an automatic bivalent valve that effectively feeds the UFH from 2 different parts of the tank starting at the bottom of the tank and sets the flow temperature there as well.
If does demonstrates how difficult it is to design these systems and suppliers just roll out a standard setup, charges a fortune, then just walk away.
The more I read the more I think I was lucky with my system. However, I have recently worked out a major flaw in how it is installed that I need to correct. Currently the Boiler only heats the top 1/3. I’d be better heating the whole tank to a lower level. I’ll get it modified this year.
I installed the sensors as I completed all the wiring. I installed it at the pdf, but the engineer who commissioned it on behalf said words to the effect of " oh I know the manual says there but we have found it is far better here" and moved it. Hence why it does not look as tidy as my original placement!
Regarding your edit @dMb, I was confused because I believed it had inverter control. I look forward to the new data.
