Tricky heating control problem

[John Aston]

Any brainy people out there who can help me with this space heating control problem? None of the manufacturers (Honeywell, Danfoss etc.) seem to be able to...

I am planning to install underfloor heating (UFH) and radiators in each room in my house. I want the underfloor heating to keep the rooms warm when the outside temperature falls below 21°C (the house temperature set point). However, if the temperature continues to fall, the output of the UFH will be insufficient to maintain the room temperature at 21°C. The thermal characteristic of each room is different of course but, generally speaking, when the outside temperature has fallen to about 12°C, the radiators should switch on to supplement the underfloor heating.

Is there a way to automatically switch on the room radiators *only when* the underfloor heating has reached its maximum output capability?

I don't want the radiators to be on when the outside temperature is at or above 12°C. At 11°C the radiators should come on "a bit", at 10°C they should output "a bit more" - you get the picture.

(Note: Underfloor heating is slower to respond than radiators. If the outside temperature falls rapidly from 21°C to 12°C, say, it would be useful for the radiators to turn on immediately until the UFH cranks up to its maximum output. Then the radiators could switch off and handover heating to the UFH.)

Hope that I've explained this OK.

[Tony Bryer]

The inside temperature of an occupied but otherwise unheated building will always be a few degrees above the outside temperature because of the gains from lighting, cooking, appliances and body heat. So your assumption above probably needs to be adjusted.

The test as to whether you need the additional radiator heat is surely whether the inside temp is acceptable or otherwise - though this of course will depend on the outside temp. If you control your radiator circuit though a Honeywell CM67 programmable stat with optimiser the optimiser will estimate whether the programmed temperature will be reached anyway, and if so will turn off the rads. One of the other regulars here can probably give a better explanation of the optimisation algorithm.

[John Aston]

Agreed.

The rads could be on and the UFH could be off and the internal temperature of the house could be acceptable, but I'd rather it was the UFH that was (mostly) supplying the warmth.

Thanks, I had a quick read about the CM67 at

CM67 seems to add some outside temperature-dependence to control of the switch-on time (does it do anything else?) I'm not sure how I can use it to switch off the rads once the UFH is up to speed.

[IMM]

Design and install the UFH to operate at -3C outside and have appropriate controls to modulate the heating system. Abandon the rads

Why?

Have a thermostat in a weatherproof box on the north wall of the house. That is the simplest way.

Modulate.

What you are doing is half cocked. Either have full UFH or full rads. A good modulating condensing boiler with load compensation control will modulate the rads to suit. So why have expensive UFH when the rads, which you have anyhow, can do it all.

It is possible to do what you want. A two stage controller will work and modulate both using modulating 3-way valves. These are commercial products and are very expensive. A simple way is have an outside stat to hold off the rads and an inside stat to cut them off when inside setpoint is reached. The rads will only be used as a boost in an off-on control. This may mean two stats next to each other on the wall. One for UFH and one for rads. Or modulating UFH controls and on-off for the assist rads. But why bother!

[Set Square]

You'll presumably have 2 zones - on for the UFH and one for the rads. Have a separate room stat for each zone (2 x CM67's would be fine) but set the stat controlling the UFH slightly higher than the one controlling the rads. Then, if the UFH is able to maintain the temperature on its own, the rads won't come on. If you want to take outside temperature into account, have an outside stat and wire it in series with room stat controlling the rads. Then, the rads won't come on regardless of the inside temperature if the outside is above (say) 12 degrees. [You could use a CM67 for this too - mounted *inside*, but using a remote temperature sensor which is outside. If you muck about with its setup parameters, you can achieve a degree of proportionality between (say) 10 and 12 degrees.]

[Tony Bryer]

I stand to be corrected but thought that the only thing the CM67's external sensor did was to give you the facility to check the outside temperature from inside. I wasn't aware that it also acted as a weather compensator.

[John Aston]

Wish I could avoid installing the rads! But my house is old and draughty, and underfloor heating on its own will not provide sufficient output to keep it warm on very cold days.

I could install just radiators and forget the UFH, but I prefer the heat given off by UFH. Just my personal taste, really.

(I am, of course, also trying to draught-proof and insulate the house as best as I can.)

See above. On its own, the UFH can only keep the house about 9°C warmer than the outside (this is a conservative calculation based on the U-values of the house). I need the rads to boost the heating when the weather gets cold.

See above.

[IMM]

No. Two heating stages.

[IMM]

appropriate

You have answered your problem. The money your would have spent on a UFH system spend on heavy insulation, glazing and draft proofing. Have a modulating condensing boiler on rads. Sorted.

Does the house have wood or solid floors?

[Andy Hall]

That is what it does - outside display, not control from the sensor.

[IMM]

I don't, to my knowledge.

[Set Square]

The CM67 allows you to connect an outside sensor (not sure what this does!) and/or a *remote* sensor - which allows you to "control" the temperature at a different location from that of the stat body itself. What I am suggesting is to use a remote sensor, but to mount it *outside*. The stat then thinks that's it controlling the outside temperature (which of course it isn't!) but it's connected in series with another inside room stat, and serves to stop the rads coming on when it's warm outside. Once they *do* come on - because it's cold outside, they're actually controlled by the inside room stat. [This solution requires three stats in all - 2 for the rads and one for the UFH!]

[IMM]

What an expensive way to do something so simple.

[Andy Hall]

Sorry, you're right. In the older version of the product, you couldn't do this (I used to have one), but the new one has a choice for how the connected sensor is used. However, it does appear that the choice is one function or the other from the menu settings - not that it matters here......

[Set Square]

Yes, you're right - it is one or the other but not both. I couldn't quite remember when I wrote it, but have since looked it up. As you say, it doesn't matter in this case - a remote sensor mounted outside is all you need. A somewhat unorthodox use for a CM67 though!

[Aidan]

It could be done, quite easily (but expensively), with a BMS (building management system) outstation. These are like small computer devices, without a screen or keyboard; they are used solely to monitor and control building services. Have a look at the Trend website, for example. However, they're only marketed for commercial buildings and the entry level costs (for an outstation, programming, a PC supervisor and the necessary software) is likely to be in 4 figures. If that hasn't put you off, any of the Trend suppliers, TTCs, would be happy to quote you. There are many other BMS manufacturers.

Varying the heat output according to the outside air temperature (weather compensated/ outdoor reset) is a common control strategy. If this is a existing building in which the heat output from UFH would be inadequate to cope with the heat losses, it may be better to have the perimeter radiators running continuously, to counteract cold convection currents produced by the glazing. You will have two separate heating systems in one room, which isn't ideal, but it shouldn't be a major problem.

I believe the domestic systems, e.g., Honeywell Hometronic, are less flexible, and primitive in comparison with BMS; I've had no dealings with them, so can't comment. The necessary programming is kept to a minimum because many domestic installers won't need to be be familiar with the BMS type control strategies.

[IMM]

The simplest and best is to have a two stage modulating controller. Each stage modulates a 3-way mixing valve, one for UFH and one for the rads. One internal sensor would do it to maintain a setpoint. No need for anything outside. A commercial setup so expensive, but better than a BMS system, and cheaper.

The UFH should do this as the whole floor is a heater. He could have rad [perimeter rads on a weather compensator to give background heat and the UFH topping up. The response would be sluggish as the UFH would be slow in raising the temperature. Best have the UFH as background and quick to respond rads to top up.

So, that would be a UFH setup on a weather compensator modulating a valve. When the valve is fully open indicating it is on full, it activates a microswitch to turn on the rads which operate on their own stat. The rads then take the room up to say 22C when the setpoint is reached the rads off, but the UFH is on which you don't want. A weather compensator which has room influence, set to 22C, would then have modulated down, maybe to off, the UFH, until the room is below setpoint then it comes on. This way the UFH may take control and only bring in the rads for boost.

That would "probably" be the cheapest option.

[Tom]

This reads like SPAM to me. Tom

[Aidan]

No spam, Tom. I'm just telling the OP what I've learnt through years of dealing with commercial Trend controlled plant. BMS is great stuff & can control almost anything. I lament the fact that it's price makes it, or anything comparable, inaccessible for all domestic users except the very affluent. And BMS technicians.

The only other solution AFAIK, other than the domestic heating control systems suggested by others, is through commercial PLC controllers. I know zilch about PLCs.

PS. The plumber would have supported your cyl "The simplest and best is to have a two stage modulating controller."

Who makes such a controller and which model is it?

If you're suggesting discreet controllers (i.e., 1 analogue/thermistor input & 1 analogue/0-10V output per controller) these are, in my experience, a waste of space compared to a BMS outstation (multiple analogue or digital inputs and outputs configured by the software).

"When the valve is fully open indicating it is on full, it activates a microswitch to turn on the rads which operate on their own stat."

The OP has stated elsewhere, UFH in a solid floor has a huge thermal inertia. If the room temperature falls, causing the demand to increase to 100%, it could be an hour or more before the UFH is actually emitting 100% of it's output. In the time it takes to turn on the radiators, the room temperature could have increased (e.g., dawn). Once the setpoint is reached and the UFH output reduces, the UFH will be emitting excess heat for another hour, despite the space temperature having exceeded the setpoint. The heating output will lag behind the room temperature. The OP seems to have a good grasp of the problem.

In comparison, weather compensation will react to outdoor air temperature and increase the UFH output. It will take some time for the low outside air temperature to affect the indoor temperature and it will also take some time for the UFH output to respond to the increase demand. Weather compensation ain't perfect (e.g. it takes no account of solar gains or wind cooling) but it's pretty good. Ii'sprobably good for this application.

In the first few lines you had said; "One internal sensor would do it to maintain a setpoint. No need for anything outside."

Later on you say: "So, that would be a UFH setup on a weather compensator (i.e., outdoor air temperature sensor) modulating a valve."

You've got me mystified, International Man of Mystery.

[IMM]

They are unsuitable for this application.