Cliff, welcome to the forum.

You’re running into one of the classic debates in heat pump operation. Here’s my take. Others might disagree.

The fact that you’re seeing temperature overshoot with KISS isn’t surprising, and it highlights a key limitation of that approach, especially with some heat pump brands.

Grant is right in saying that if you rely solely on flow temperature without a stat, you don’t have a direct feedback loop for room temperature control. Weather compensation works on the assumption that the external temperature is the only variable affecting heat loss, but in reality, things like internal gains (appliances, people, solar gain) can throw that off, leading to overshooting. Your experience of a 3-degree overshoot suggests that either your system is responding too aggressively or your house has more thermal lag than expected, meaning heat keeps building up even after the system has backed off.

The issue with dialling down the heating curve too much is that, as you’ve found, you risk setting it so low that it won’t deliver enough heat when you actually need it. If you went down to 0.5 and still had overshoot, that suggests the system is struggling to modulate correctly at lower loads.

Grant’s recommendation of running “constantly” with a thermostat is a bit misleading, because as you noted, it’s not truly constant. It’s just ensuring that the system remains active, with the smart controller managing small adjustments. This approach can work if the thermostat has a wide enough hysteresis to prevent short cycling. If it’s too tight (e.g., calling for heat every half a degree), you’ll end up with inefficient on/off cycling, which defeats the purpose.

Since you’ve tested both approaches and found that Grant’s method also proved economical, it sounds like that’s the better fit for your setup. You might still be able to refine it, though. If your controller allows for it, setting a lower overnight setback and using a slightly less aggressive heating curve might help balance things out. It’s also worth checking if Grant’s controller has any form of adaptive learning, where it adjusts over time to reduce unnecessary swings.

At the end of the day, the best method is the one that keeps your home comfortable and efficient without excessive cycling. If KISS isn’t giving you the control you need and Grant’s method is working well, then that’s probably the way to go. That said, if you’re still fine-tuning, it would be interesting to see what happens if you use weather compensation but introduce a room stat with a very broad range; something that only kicks in if the room temperature strays too far from target rather than micromanaging every degree.

Im with @editor on this, but its definitely worth experimenting. 

Many (including me) do find that pure weather compensation with absolutely no room reference at all works very well.  However if there is a lot 'going on' in the house then it may not.  If its just a room or two which is off, try putting a TRV in that/those rooms only, using it as a limiter by setting it a couple of degrees higher than the desired temperature.  I, for example, have a room which gets lost of solar gain.  On recent days its got up to 24 (target 20.5) which is bit too warm.  It has a fancoil which shuts down its fan when it gets too hot, but doesn't stop the water flow so there is still heat coming out.  Closing the curtain helps, which is what I do in the height of summer anyway.

Just one more thing, there is no such thing as 'ridiculously low' for the heating curve.  The curve needs to be as low as you can get it whilst just heating the house sufficiently, whatever that is.  At this point ignore the 'design', its based on a heat loss estimate that could be way out, by as much as a factor of two,  At this stage all that matters is how the house responds.

One further thing, if you turn the heat curve down and it still overshoots when its milder but not when its really cold, its probably cycling or the heat pump is oversized.  

Thanks Mars

Some good stuff in there

The overshoot on KISS is probably nearer 2 degrees but I was trying to modulate at 19 degrees at the time in mild conditions.  My design temperature one end of the curve is -5 at 45 and shifting the curve too far meant this was not achievable.  Running it constantly even at night though doesn't feel right somehow and we don't need 20 degrees then even when its zero outside.  I am seeing this right am I?  Maybe I'm thinking too traditional.

The Grant method, which I was using anyway before I spoke to them, seems the way to go.  I can at least use setback at night of 18 e.g. when the day is 20.  The controller allows this. And the setbacks can alternated within reason with day rate to keep LWT up.

The hysteresis on the smart controller is adjustable and I've got it set to 0.3.  The system at zero outside is coming in about every 45min at mo so LWT is still warm.

Cliff

@cliffhanger @editor - I would be inclined to ask how and why is the overshoot happening. I'd start by looking at the estimated heat loss, and how the heat pump and emitters were matched to the estimated heat loss, bearing in mind the possibility that the estimated heat loss may well be considerably in excess of the actual heat loss. I would also look at the characteristics of the building, in particular the thermal mass. Overshooting by three degrees is really quite an overshoot. I would also at least consider the possibility that the Grant controller has a fault - such things do happen. @cliffhanger, can you post these details for your system, plus any data you may have, including if possible any historical charts of the overshoots, if Grant's controller makes that possible. By the way, I am sure you have spotted this, but in case you haven't, I wrote that ABC guide.  

Now we are all posting at the same time! One thing to add - I agree with @jamespa, there is no such thing as a ridiculously low curve, only a curve that best fits your requirements, home and lifestyle. 

Posted by: @cliffhanger

↑

Running it constantly even at night though doesn't feel right somehow and we don't need 20 degrees then even when its zero outside.  I am seeing this right am I?  Maybe I'm thinking too traditional.

You are being too traditional, heat pumps mostly should be run 24*7 as low as possible.  There are of course exceptions if you cant get this to work, but this is generally the starting point for minimum operating cost and maximum comfort (which, helpfully, happen at the same operating conditions).  A very modest nighttime setback may save some money (but may not, it may actually cost money) and may be necessary for comfort, but dont, at least initially, do a setback just because you somehow think its the right thing to do (based on boiler mentality), it quite probably isnt!

Posted by: @cliffhanger

↑

My design temperature one end of the curve is -5 at 45 and shifting the curve too far meant this was not achievable

Ignore the design temperature for the reasons set out in my previous post above.  Its history, what matters now is what the house does.

I see that @cathoderay has just posted, I agree with all he says.  Hopefully this gives you some food for thought please keep posting as you experiment and (as you can see) there are plenty of people willing to try to help.

Posted by: @jamespa

↑

Im with @editor on this, but its definitely worth experimenting. 

Many (including me) do find that pure weather compensation with absolutely no room reference at all works very well.  However if there is a lot 'going on' in the house then it may not.  If its just a room or two which is off, try putting a TRV in that/those rooms only, using it as a limiter by setting it a couple of degrees higher than the desired temperature.  I, for example, have a room which gets lost of solar gain.  On recent days its got up to 24 (target 20.5) which is bit too warm.  It has a fancoil which shuts down its fan when it gets too hot, but doesn't stop the water flow so there is still heat coming out.  Closing the curtain helps, which is what I do in the height of summer anyway.

Just one more thing, there is no such thing as 'ridiculously low' for the heating curve.  The curve needs to be as low as you can get it whilst just heating the house sufficiently, whatever that is.  At this point ignore the 'design', its based on a heat loss estimate that could be way out, by as much as a factor of two,  At this stage all that matters is how the house responds.

One further thing, if you turn the heat curve down and it still overshoots when its milder but not when its really cold, its probably cycling or the heat pump is oversized.  

Yep that's good to hear. The heat curve issue that is.   I thought the pump might be undersized but the company did a good amount of cals which I've read.   You're right though could be out on a 190 year old cottage!   Well insulated where possible of course.

Posted by: @cliffhanger

↑

You're right though could be out on a 190 year old cottage! 

Calculation based heat loss assessments are often particularly over the top for older buildings because the assumed U values for walls are just plain wrong. I have a similar old building, and my calculated heat loss is ~12.3kW whereas my actual (measured) heat loss is nearer 9kW. Put another way, the calculation based estimated heat loss is ~33% greater than the actual heat loss.

Posted by: @cathoderay

↑

there is no such thing as a ridiculously low curve, only a curve that best fits your requirements, home and lifestyle. 

This is an excellent point. Our heat curve is currently set at 0.3 or 0.4, which works well enough for us at this time of year. However, with the recent solar gain and mild temperatures, our heat pump still cycles regularly because it’s oversized for the current ambient conditions (8C and above). The on-off stats, which we hope to remove soon, exacerbate the issue by forcing cycling when areas reach 23C. We’ve set them to their maximum, but it’s not enough to prevent the problem. Fine from November to February, but not for the rest of the heating season.

It’s not ideal or automated, but I’ve found myself manually adjusting the heat curve based on ambient temperatures to compensate for our poorly installed system. On days like yesterday, I even turned the heat pump off for six hours because the house was warm enough, and I wanted to avoid unnecessary cycling.

It’s frustrating to see how poorly installed and oversized systems are killing efficiency. I genuinely can’t understand why so many systems are still being installed so badly. It’s a disservice to homeowners.

@editor Will get back.  Lots to try.  Thanks to everyone

Cliff

Posted by: @cathoderay

↑

@cliffhanger @editor - I would be inclined to ask how and why is the overshoot happening. I'd start by looking at the estimated heat loss, and how the heat pump and emitters were matched to the estimated heat loss, bearing in mind the possibility that the estimated heat loss may well be considerably in excess of the actual heat loss. I would also look at the characteristics of the building, in particular the thermal mass. Overshooting by three degrees is really quite an overshoot. I would also at least consider the possibility that the Grant controller has a fault - such things do happen. @cliffhanger, can you post these details for your system, plus any data you may have, including if possible any historical charts of the overshoots, if Grant's controller makes that possible. By the way, I am sure you have spotted this, but in case you haven't, I wrote that ABC guide.  

Hi cathodeRay

Can't pm you I haven't enough posts.   I'm attaching the docs I have on heat loss.  The controller is 3 weeks old so malfunction unlikely and no historical data, controller doesn't do this. 

Many thanks