Hey folks, I've been experimenting with cooling using my heat pump and thought I'd share my experiences here. If you have an ideas for making this work better I'm all ears! I have a 8kW Mitsubishi Ecodan ATW ASHP that was installed 4 years ago and has provided me with wonderfully cheap heating without burning stuff. Heat Pumps can, in theory, be reversed, so that instead of moving heat into the house by making your radiators hotter, it can move heat out of your house by making your radiators colder. I've seen the Heat Geek videos where they play around with reversing the heat pump to provide cold water to the radiators and conclude it's probably not worth it, but I gave it a go anyway.

How I did it

In short, I did this:

I turned off my heat pump at the Consumer Unit (or “fuse box” in old money), opened the cover of the indoor control unit, flipped a dip switch (SW2-4), turned it all back on, went to the settings on the LCD control panel, and lo and behold, it now has the option for cooling mode!

You need to do this because most heat pump installers will leave cooling mode disabled. My heat pump was pretty easy to enable cooling mode on, but yours might be more difficult. It might involve buying and installing a chip, which might seem more daunting. This probably won’t void your warranty, but I did wait until mine was more or less expired before doing it, just in case.

Why you shouldn’t do this

Installers leave this setting disabled because it’s very much outside the boundaries of what an ATW ASHP should be doing. Running cold water through radiators may cause problems. If the water in the pipes and radiators is below the dew point in your house, water will condense on those pipes. Over time, that water will start to cause potentially serious problems, such as rotting the wood that holds your floors and ceilings up. Installers don’t really want to be on the hook for that stuff, so they leave it off.

What to expect

Installers will also tell you that the cooling effect is minimal and not worth the risk of damange to your property or to the heat pump itself. They might be right. This isn’t air conditioning: your house won’t suddenly be 21 degrees when it’s 30 outside. It will lower your room temperatures by 1 degree, maybe 2 if you’re lucky.

But that 1 degree is hugely significant to how you will feel in your house when it’s ridiculously hot outside. It won’t do much on its own, but if you couple it with other cooling techniques such as using fans, having windows open at night and closed in the morning, curtains/blinds closed to reduce solar gain, etc., it will make your house feel much more pleasant on those really hot days.

The difference between 25 and 26 indoor temperature is, in my experience, very noticeable. And at night when it gets cooler and you open your windows again, you will be cooling your house down from 25 instead of 26, which makes the nights more bearable too as you reach lower temperatures sooner.

Some graphs and data

Here’s the comparison of temperature with and without the cooling. Both days were the same outdoor temperature, and nothing else was different. It’s about as fair a comparison as you’re going to get I think:

The important thing here is where it flattens off after I turn it on. For a good chunk of the day, between 11am and 3pm, the cooling power of the radiators was enough to match the heating power of the sun. Eventually, nature won and the temperature started to rise as it got hotter in the mid-afternoon. But for a good chunk of the day, the house simply didn't get hotter.

The heat pump did about 1.5 to 2kW of cooling on average over the 3 days I monitored it. It had a CoP of around 3, which is pretty rubbish for air conditioning, but I have solar panels so when it's this sunny I don't care!

Things I’ve learnt

The temperature of the radiators doesn’t matter that much

You can control the temperature of the water in your radiators; in fact, for my heat pump, that is the only possible setting in cooling mode. But it actually doesn’t matter that much when your house is already at 25 degrees. What really matters is how much heat from the air in your rooms can be transferred to the cold radiators. The colder your radiators, in theory, the more heat it will absorb from the air. But air is really crap at transferring heat. The normal way radiators work in heating mode is by circulating air via convection currents. The constant airflow created means that cold air is constantly being drawn up through the radiator, warming along the way, and being circulated back into the room. But this doesn’t work in reverse: the cold air from the radiator mostly just stays on the floor, meaning the air immediately surrounding the cold radiator gets cold, but the rest of the room stays hot.

Reducing the temperature of the radiator won’t improve this much. The amount of heat absorbed by the radiators when at 12 degrees is more or less the same as at 17 degrees.

So, to reduce the risk of condensation, I monitor the dew point and set the flow temperature to be 2 degrees above that. Typically, this means 15-17 degree flow temperatures.

What really matters is the Delta T between Flow and Return (at a given flow rate)

The basic equation for how much heat is being absorbed by the radiators is Q = m * C * DT, where Q is cooling power, m is the flow rate, C is the specific heat capacity of water, and DT is the difference between flow and return temperature to your radiators. I know from previous experiments that my heat pump measures m and DT fairly accurately, and therefore calculates Q fairly accurately, so I can more or less trust the measurements from the heat pump itself.

However, even if I don’t trust it, the basic physics is the same. All you have to worry about is whether the return temperature is higher than the flow temperature. If it is, then the cold water in your radiators are, in fact, absorbing heat and getting hotter, meaning your house is getting cooler! So just monitor the flow and return temperatures, either manually with a thermometer or automatically if you have the means and/or trust your heat pump.

You’ll want to do this anyway, in order to avoid condensation… On my unit, the flow temp is slightly below the temperature I set on the control unit, meaning the pipes near the unit are cooler than the 17 degrees I might set for the flow target. This means that I need to set the flow temp a couple degrees higher than the dew point, to ensure that the outgoing pipes aren’t colder than the dew point to avoid condensation.

Delta T is pretty much 2 degrees no matter what the flow temperature is

Anywhere between 12 degrees and 17 degrees, for me at least, DT is 2 degrees. Probably there’s some rounding in there, so 12 is closer to 2.4 and 17 is closer to 1.6, but in practice, the unit itself reports very little change in cooling output when I run it cold vs hot. I suspect this is because the temp of the rads simply isn’t the limiting factor here: you need to get push warm air to the radiators (or draw cold air away from the radiators), and this is what limits the cooling power of the radiators more than the temperature itself.

So don’t worry too much and set the flow temp to be a good amount above the dew point. I know I’ve said this 3 times but it’s important!

My unit reports that it does between 1.5 kW and 2 kW of cooling on average. This is pretty good! It’s enough to reduce the air temp by 1 to 2 degrees.

The Dew Point goes up slower than the indoor air temperature

Generally speaking, in the morning, the dew point is below 15 degrees, and by the evening, the dew point is more like 16. So the dew point rises by around 1.5 degrees throughout the day. Therefore, my flow temperatures rise by 1.5 degrees over the course of the day. But my indoor temperature goes up from 24 to 27 over that same period. This means that the difference in temperature between the air and the radiator actually increases as the temperature in the house increases, meaning that a 17 degree flow temp at 27 air temp will do more cooling than 15 flow at 24 air temp. So again, don’t worry about raising the flow temperature as the air in your house gets hotter - if it does make a difference at all, it will be in your favour as the day goes on.

Radiator fans don’t increase cooling output, but do make you feel a bit cooler

I spent nearly £400 for 5 rad fans, and they do nothing for cooling output. I was going to buy them anyway so wasn’t a total waste, but they do nothing to improve the cooling effect. They’re simply too small, and can’t draw air up through the radiators. The best use for them is to put them on the bottom of the radiator, so that the cold air goes onto your feet. This only works if your radiator is out on the open, and you’re able to sit relatively close to it. Probably it adds more to the cooling effect than the 6 watts or so that it draws (and therefore produces in heat), so I keep them on. But they are not worth getting if you don’t already have them, as the same thing can be accomplished cheaper and easier with a £10 desk fan.

An actual fan is much better

Point a big fan at a big radiator and the air will certainly move around. The air from the fan will disperse laterally mostly, at body height, so the cool air will be more likely to hit you rather than settle on the floor by your feet.

In any case, the basic problem with this set up is getting heat TO the radiators. The radiators don’t “radiate cooling” in the way that a hot object would normally radiate heat. Radiative heating works because the metal radiator heats up and emits photons. When those photons hit other objects, like your body or your sofa or whatever, the photon releases some of its energy as heat and warms up that object. Typically it gets reflected and bounces around a few times, releasing heat every time it hits something.

This doesn’t work in reverse. A cold metal radiator won’t “attract” photons to it purely by being cold. Photons need to be colliding with it for “radiative cooling” to happen. I don’t personally believe that this is a significant mechanism or effect. I think what people think of as “radiative cooling” is merely the cold air that surrounds the radiator being and feeling cold. But that’s conduction! The air is (slowly, poorly) conducting heat to the cold radiator, and getting cold itself. Then your body is (slowly, poorly) conducting heat to the cold air, reducing your body temperature slightly.

In my view, conduction is the thing we need to focus on. How can we conduct the heat away from the air to the cold object?

We have probably all heard of putting ice below a fan so that the fan blows icy cold air around. I think this is the best bet for ATW cooling. Set up fans so that they blow hot air over the cold radiator, and blow cold air from the radiator around the room. We can’t rely on convection, but we can use fans to replicate some of its effect.

(Another idea I had, which I don’t think is a good one, is using sand or ice packs on the bottom of the radiators. My theory is they slowly cool down, then I move them to a hot part of the house, e.g. the kitchen, where they absorb some of the heat from the kitchen. Then I move them back to the cold radiator. Basically, I am a manual heat pump, and sand/ice packs are my medium. …Yeah, I don’t think this is a good idea, but who knows? Worth a try.)

The house dries up pretty quickly

I turn the cooling on in the morning and off at night. This allows any condensation that I can’t see to dry up. To be clear, there is no condensation on any of the pipes or radiators that I can see. At most, I get some in the bathroom after a shower, but I leave the window open in there anyway (and shut the door) to reduce humidity. It is possible that the wood is rotting underneath the floorboards, or some other scary thing is happening, but I doubt it… Buildings are designed to get a bit wet and then dry off, and in any case I’m not running this thing 24/7/365. I’ll be running it for 12 hours a day, a week at a time, during the hottest, driest part of the year. I think the risk I am running is pretty low, personally, though other opinions are available!

Conclusion

I’m sat here in my living room when it’s 30 degrees outside and I’m a bit cold. I’m pretty impressed, quite frankly. It cost me nothing (other than sort of wasting money on those radiator fans), and provided light relief this past week. The theoretical performance is about 1.5 kW to 2kW, which in practice reduced the temperature in my house by 1 to 2 degrees or so. That’s impressive for something that cost me nothing!

Don’t get me wrong, this is by no means a replacement for any of the other things that people recommend doing to cool the house down. Those things are FAR more effective. But this actually does do something, it is pretty low risk the way I’m running it, and again, it literally cost me nothing. I have the kit in my house already, I just flipped a switch and now my house is a degree cooler than it was before.

Has anyone else experimented with this? What did you find? What worked well for you? Any tips?

Cheers!