@johnmo I don't think I have a choice with the Cosy 9 - there's less control than with other manufacturers and it is designed to work with the Cosy pod.

@andrewj so this thermostat would normally go in a room that would be considered an average comfort, this stat is also only used to feed some info back to the ASHP, its not the overall master control. 

Assuming its in a room that averagely reflects all rooms then it just adds to the ASHP ability to fine control. The only reason you would ever have one room more than 1 deg diff to any other is down to an incorrect emitter sized compared to heat loss, the emitter is not working or the heat loss changed inside of the room after the calc, for example a window open so more ventilation.

The skill is the accurate measurement of heat loss to every room and the correct selection of emitter to match that room at any one given flow temperature, its about a balanced system.

For example, if you add an emitter that is to large to one room in the house and every other rooms emitter matches perfectly then that room would suffer over heating and all other rooms would per perfect all the time, the only way to overcome this is change the emitter or control (with a TRV for example) that emitter.   This is often why we open flow the ground floor and TRV (loosely controlled) the top floor, heat rises, solar effects the building so this way it helps control the higher rooms that may suffer from either.

@andrewj Ah, maybe, they also often design to 50 Deg C flow rate and do everything to reduce the amount of emitters changed. If you balance everything to run on the correct curve you may get away with setting the main stat higher than you need, say 23 and the curve will modulate down to 21 and it never reach.

@ashp-bobba Thanks - I think going back to my earlier comment it is about having a properly balanced system so that whilst that one room is being used to fine-tune the flow temperature the other rooms are following suit automatically because they've (a) been sized correctly to give balanced temperatures; and (b) the whole system is correctly balanced.

I think I get it now, thanks.  One more question: my wife asked me how long does it take to respond?  I think that might be irrelevant because it isn't really responding on a step change (e.g. dialling up from 21c to 22c).  Anticipating her next question: how long would it take to respond if I boosted it up from 21c to 22c?

@andrewj Oh, thats a good one, so the answer is subject to what power and recovery you have for the room/s assuming its typicly balanced and has the capacity to climb another deg as a lot of system are designed to 21/22 in my experience it would take around an hour.

Remember ASHP are low and slow, they are design to run 24/7 all be it you can use a setback at several points a day. So if you are running it this way to climb that last deg I would think an hour or so would be realistic.

Sorry I cant be more accurate as it is subject to a lot of variables. 

Fair enough, sort of what I was expecting really.  Thanks for your input, it's been very helpful.

@andrewj I see there have been a lot of replies already, and the video is attracting a lot of comments. My two cents worth:

Weather compensation adjusts the flow temperature of your heat pump based on the outside temperature. It follows a pre-set curve… the colder it gets outside, the hotter your flow temperature needs to be to keep the house warm. But it does this without knowing what is happening inside your house. 

Load compensation, on the other hand, responds to internal temperature changes. It uses an internal sensor to monitor how far the current room temperature is from your desired setpoint and adjusts the flow temperature accordingly. So if you’re only 0.5C off your target, it might lower the flow temp a little. If you’re 2C off, it may ramp it up more.

You’re absolutely right in saying that load compensation works off the internal temperature of just one space, and this is where things get tricky. If your home has uneven heat losses (which most do), the room with the load sensor might hit the setpoint perfectly, but other rooms could end up too hot or too cold depending on their insulation, size and solar gain.

This is why many installers recommend combining weather and load compensation. Weather comp sets a consistent baseline flow temperature in response to outdoor conditions, while load comp adds a layer of fine-tuning, tweaking that flow temp slightly based on real-time internal feedback. When both are set up correctly, the result is a more stable, comfortable environment with better efficiency.

But (and this is key) the positioning of the internal sensor or controller matters a lot. If it’s in a hallway, utility room or near a heat source, you’ll get poor readings. Ideally, it should be in a frequently used space that represents the average internal condition.

Let me know if that helps clarify.

Posted by: @andrewj

↑

I think I get it now, thanks.  One more question: my wife asked me how long does it take to respond?  I think that might be irrelevant because it isn't really responding on a step change (e.g. dialling up from 21c to 22c).  Anticipating her next question: how long would it take to respond if I boosted it up from 21c to 22c?

As @ashp-bobba says there are lots of variables.  My house takes 3-6hrs to change temperature by a degree.  I have one room with a fancoil that responds faster.  The question hardly ever arises, other than in unusual circumstances the whole house is comfortable 24x7 unlike with my gas boiler.

Posted by: @ashp-bobba

↑

For example, if you add an emitter that is to large to one room in the house and every other rooms emitter matches perfectly then that room would suffer over heating and all other rooms would per perfect all the time, the only way to overcome this is change the emitter or control (with a TRV for example) that emitter. 

Surely you can 'correct' a mildly oversized rad by turning down the lsv during the balancing process?  Certainly that's what I did with my installation.  Obviously not ideal but, given the scatter in loss estimates (certainly the ones that I have seen), perhaps likely?

Posted by: @editor

↑

But (and this is key) the positioning of the internal sensor or controller matters a lot. If it’s in a hallway, utility room or near a heat source, you’ll get poor readings. Ideally, it should be in a frequently used space that represents the average internal condition.

I think this is only part of the key information on this topic that is missing from the video conversation - I can only see this working in a system that is well balanced such that other rooms are able to follow the temperature being called for in the room with the sensor otherwise there will be temperature differences and that will feel uncomfortable.  As an example, with the boiler system now removed, the living room (with the thermostat) would be 21c and to all intents and purposes should be comfortable.  However, the hallway that leads to the living room would be 23c and upstairs just as warm.  So moving from an area at 23c to one at 21c felt cold and acclimatisation was required so the space felt uncomfortable.  Clearly that was all controlled by an on/off thermostat (in a system not very well balanced or sized) but it's the principle that I'm talking about: unless there is a consistent stable temperature across the house then movement across temperature differences is going to be felt.  The problem being, of course, there are multiple factors that can throw things off that can't be controlled, e.g. solar gain.

Having the sensor in an occupied room will surely affect the reading - e.g. in the living room with two of us plus the TV on with the evening sun shining through the windows, might be controlled by the sensor to 21c but that will inevitably be achieved by lowering the flow temperature and thus other rooms will be cooler.  So I would think it needs to be in a space that isn't subject to any form of heating/cooling except the emitter itself, so perhaps an internal hallway is a reasonably space?

I don't know: this feels like a topic that could be covered by your panellists by perhaps not digressing into why installers don't use WC (we know that there are poor installers out there so in reality answering that doesn't move us forward) or wireless controllers etc.  Not being critical here just throwing that into the middle for future videos.

Posted by: @editor

↑

Weather compensation adjusts the flow temperature of your heat pump based on the outside temperature. It follows a pre-set curve… the colder it gets outside, the hotter your flow temperature needs to be to keep the house warm. But it does this without knowing what is happening inside your house. 

Load compensation, on the other hand, responds to internal temperature changes. It uses an internal sensor to monitor how far the current room temperature is from your desired setpoint and adjusts the flow temperature accordingly. So if you’re only 0.5C off your target, it might lower the flow temp a little. If you’re 2C off, it may ramp it up more.

The other important point about wc is that the controller gets the signal to change what it is doing at the same time as the principal driving factor (OAT) changes.  This means that it changes in time to allow for the delay in the response of the house.  With load compensation, once the controller knows to do something, it's too late to do it in time to allow for the delay in the response of the house, frequently leading to oscillation.

This is why weather compensation (with or without an element of load compensation as a secondary 'tweak') will often (not always) result in a more stable outcome than load compensation for most (not all) of the time.  Solar gain, cooking, and wind in particular can disturb wc, which is where some form of secondary control comes in.

@andrewj, you’re absolutely right that these systems work best when the house is well balanced and properly designed from the outset. Even with compensation enabled, issues like oversized emitters in one zone or unexpected solar gain in another can throw off comfort levels across the home. Your example of the thermostat in the living room responding to a combination of occupancy, appliances and sunlight, while other rooms get cooler, is a perfect illustration of the challenge.

Where you place the load sensor absolutely matters, and you’ve highlighted one of the most practical options: a neutral hallway unaffected by direct gains or unusual occupancy heat. That said, even then, it all falls apart without a solid design and correctly sized emitters.

We totally take your feedback on board about future content… there’s a strong case for dedicating a full episode to real-world compensation strategies: ideal sensor placement, zoning logic, temperature consistency and the limits of what compensation can actually solve. I’ll schedule one in the next few weeks so it’s out well before we enter the next heating season.

This stuff is often presented as a silver bullet, but as you’ve pointed out, it’s more nuanced than that.

Really appreciate your engagement and suggestions… please do keep them coming.