Posted by: @sjn

↑

The pump is a 8kW Samsung Gen 6 connected to a Joule Cyclone Smart Plumb (including buffer tank

A few questions might help point us in the right direction:

Is the buffer tank plumbed 2 port or 4 port?

Is there a secondary pump on the emitter side of the buffer tank and how is it set up?

What is the indoor temperature?

I note that there is a peak in output just before 1600 each day, is that when your DHW heats up?

A 100 sq m new build shouldn't need anything like 8kW, more like 4kW, so that may be a factor.  To my surprise my 7-8kW house (measured figure)  has been running all day at 4kW and is toasty, not sure why (perhaps because the exterior of the fabric is only slowly cooling down??).

Posted by: @jamespa

↑

A few questions might help point us in the right direction:

Is the buffer tank plumbed 2 port or 4 port?

It is a 4 port buffer.

Is there a secondary pump on the emitter side of the buffer tank and how is it set up?

Yes, there is a secondary pump on the UFH return to the buffer.

The initial install included yet another pump on the secondary at the UFH manifold as directed by the Joule design. This was always on when there was any call for heat, even if the secondary circuit valves were closed for a DHW cycle so it was disconnected from power at the last service. It is still freewheeling in the secondary circuit but the service people are here again on Monday to remove it totally.

What is the indoor temperature?

We were getting a nice 22C but it has dropped to just above 20C since this issue started.

I note that there is a peak in output just before 1600 each day, is that when your DHW heats up?

Yes. This gives us enough DHW for the full day.

A 100 sq m new build shouldn't need anything like 8kW, more like 4kW, so that may be a factor.  To my surprise my 7-8kW house has been running all day at 4kW and is toasty, not sure why (perhaps because the exterior of the fabric is only slowly cooling down??).

Agreed. I think the system was over specified at 8kW. It certainly struggled in the shoulder months as it can not modulate below around 2kW which seems to be too much for when we are anything over OAT of 8C. It short cycles if I try a flow temperature of anything below 32C.

I was expecting everything to be fine when the heat demand increased but with it not holding the target temperature I don't think the UFH can emit enough energy.

OK.  It does sound like its possibly the system not the heat pump, thats limiting the output, and since room temp has dropped then there quite possibly is a problem.  The plot of compressor frequency makes me think its trying to heat up the FT but getting stopped by something

There are a few things to consider

Is the UFH able to emit enough heat?  Hopefully it was designed to emit enough, but at what flow temperature?  Do you know?  Do you know what temperature the flow to the UFH is actually at?  Unless the buffer tank is well set up (which statistically is very unlikely indeed) the flow to the UFH is likely to be ~5C below the flow from the heat pump (which BTW is costing you 10%). 

Is the secondary pump running fast enough to take the heat away (can you measure the temperatures on any of the emitter side ports?).  

Have you got any secondary controllers (thermostats, homely etc) limiting output on the emitter side.  Also I think (but am not sure) you may have selected a value for 2091 which means that the WL controls the flow temp but the internal sensor in the controller will switch it off once it reaches set temp - can you confirm the value of 2091.  What is the target room temp set to on the Samsung controller and where is the controller located?

These are possible explanations based on what you have said, although the oddity is that its not holding the target flow temperature, so maybe there is a genuine fault in the unit itself (presumably its still under warranty?)

Its very likely the case that you ideally want the buffer out, or replumbed as 2 port either in the return or the flow.  Any idea why it was fitted?

If its not obvious and its still under warranty then I would probably get the installers/Samsung to have a look.  The fact it isn't reaching target FT is a bit odd and I dont have an immediate explanation (others might).

@jamespa - thank you for your comments.

The plot of the compressor frequency is very interesting. Before I had the instrumentation sorted out I was always curious about the continual revving up and down sound as if the pump was "hunting" for a stable point so it was good to see it in the data. It made me think of badly tuned PID controllers that I have worked on in control systems in the past. I was pleased to see the stable state that the system achieved over the weekend with just the occasional change from around 16Hz.

The buffer and secondary pump set up does appear to be reasonable with around a 1C difference between the reported TW2 from the pump and UFH flow into the manifold (measured using a ds18b20 sensor strapped on the outside of the flow pipe). For example, early this morning:

In this next plot the dT between the UFH flow and return for the same period shows the difference between when the TW2 is at target and when it is low. My rough calculations are that this equates to the UFH emitting around 2kW at target and 1.4kW when low.

Unfortunately I do not have any detailed design data.

The 2091 setting on the Wired remote controller is "Use(Signal ON/OFF) or WL Interlink OFF( Water Pump2)" - I think that's the value 2 described in the manual. 

The system is set to use an external thermostat. There are individual stats in each room which go to a wiring centre. When a thermostat calls for heat, the manifold valve for the room is opened and the external thermostat input to the Samsung controller is activated. Based on advice from this site, I have removed the manifold valve actuators for all but the bedrooms meaning that the UFH for the rest of the house is one open zone. I can then call for heat using one of the public room thermostats. The one I am using is set to 26C so is continually calling for heat. Sorry if that's all a bit complicated but I am in mid trial stage here.

Like you, I think that it would be better if the buffer was configured as a 2-port. My biggest problem is getting anyone around here to work on it. The visit on Monday is them just getting back to remove the redundant pump that was identified as an issue during the service way back in May.

The pump is still under warranty so that may be the way to go. It is good to have some ideas to talk to the service people about.

Many thanks again for your thoughts.

OK 

I must admit Im a bit stumped.

Reducing it to its basics, the heat pump does not appear to be reaching the flow temperature you have told it to reach ,and there is no particular evidence from the return temperature that the 'system' is telling it not to.  If it were consistently reaching it then cutting back when the return temp was rising, I would say its the feedback from 'the system', but that's not what I observe in your plots.  If the Gen 6 had an 'adaptive' mode (where the system learns your house response) then I might say it was that, but I don't think it does.  If you had Homely, I would say it was that, but you presumably don't.  Basically I think your system has a controller which is supposed to achieve a FT and doesn't, but doesn't overshoot either and clearly isn't maxed out.  Your 'poorly tuned PID controller...hunting' does indeed seem relevant.

I would be tempted to present it to the service people in these simple terms (leaving out any discussion of PID controllers, which they likely as not won't understand).  Perhaps, if you have time, do some experiments where you somehow force the target FT to be different, in particular higher, and see if you observe the same behaviour.

One thing to check, although I would be surprised if its the issue based on what you say, I think that the Samsung has an easily accessible 'FT Offset' setting, I presume that is set to zero.  

Others, familiar with Samsung, may have more insight 

@jamespa thank you for your helpful comments. It has been very useful for me to write all this down and for someone else to sanity check it in case I had missed something obvious.

You are correct that I have no smart controller like a Homely and that the Gen 6 appears to be a straight, FT target control system. Yes, the FT Offset is set to zero.

I will try some more experiments over the next few days and discuss it with the service people coming on Monday. We can take it to Samsung support if they have no further thoughts.

Many thanks again

Stuart

Posted by: @sjn

↑

I will try some more experiments over the next few days and discuss it with the service people coming on Monday. We can take it to Samsung support if they have no further thoughts.

Sensible IMHO, varying the target FT might give a clue!