ASHP having to work hard to deliver lots of hot water

I thought I’d share this example of a 14kW air source heat pump having to get on with providing four showers after a few weeks of being switched off. We have a 300L tank. It was 22 °C outside.

I was surprised it took an hour to get hot at the start. It was also pretty rubbish in that half-way through the first wave it switched to running the radiators. I certainly didn’t ask it to switch and neither did my scripts.

You can see my algorithm asking for a target temperature of 60 °C because there hadn’t been a recent legionella run. I had everything completely switched off whilst we were away. The algorithm decided it didn’t consider it had been hot enough for long enough so the second wave of heating got a 60 °C target too. Then it judged the second wave had been sat in the tank long enough so the third wave of heating only had a target of 38 °C. I’m wondering if I’d have been better off just flushing 300L of the stale water out to the garden and just heating up to 38 °C and not worrying about legionella in the tank. Of course we’d have still needed to flush the pipework leading to the shower.

The second wave looks more like my normal heating run of 15 minutes because the tank was already pretty warm. The sensor is low down in the tank, it’s the default Mitsubishi layout.

The other thing that caught me by surprise was the lack of solar diverter. That would have normally covered most of our needs on a sunny day like today, but it was also turned off whilst we were away. We usually would have spread the showers out throughout the day, but because we all got home at the same time and wanted showers it created an unusual demand. I can see why someone might prefer their combi-boiler in a situation like this, but I’m still happy with the heat pump.

1 Like

Do you actually use the stored water? I thought most tanks now used a coil to heat the water used as it is used (direct), rather than a coil to heat the water in the tank (indirect)?

Yes, we use a 300L store of domestic hot water.

There’s an external heat exchanger on the tank that transfers the heat from the heat pump to the water from the tank. There’s a pump to move the cold water from the tank through the heat exchanger. We draw the warm water down from the tank.

It’s quite useful to allow us to (normally) choose when to heat the water.

I’ve also had a mysterious switching-to-heating in the middle of the hot water cycle in the spring, but it hasn’t happened since.

With a modern unvented domestic hot water system with regular usage, I feel that the risk of legionella is so low that I don’t worry about trying to prevent it. This article (h/t @TrystanLea) seems to support that idea: REHVA Journal 06/2020 - Energy savings in hot water supply by legionella modelling. There’s a much greater risk in old non-domestic systems (e.g hospitals), or air conditioning systems.

I feel that having the temperature sensor at the bottom of the cylinder makes one think the tank needs heating when it doesn’t really. You can see that the middle hot water cycles rises very quickly, suggesting that it was mostly hot still. I think you could have maybe run 4 showers from a single tankful, depending on how long people take. (Maybe 5 litres/min from the tank? So 300L would last an hour?)

I find that I can get away with heating my 250L tank just once a day, and we (almost) never run out of hot water. I’d really like to get a thermometer into the top portion of my tank; I’m not sure there’s a pocket for one. I wonder if the return temperature might be indicative of how hot the top of the tank is?

1 Like

It depends more on whether it is direct or indirect tank. If the water in the tank simply heats fresh water flowing through a coil on it’s way to the tap, the risk of legionella is extremely small even in rarely used showers. It is the cycling of the water temperature that, AIUI, is the major risk factor. If you actually use the water that is in the tank at the taps, there is a higher risk (but still small).

I have a whole string of temperature pockets in my tank and the variation down the tank is striking.

1 Like

Here’s how I understand it, and I’d be glad if someone could correct me:

I would call what you are describing a thermal store: where the potable water isn’t stored in the tank itself. It only passes through a heat-exchanger on its way to the tap.

A ‘direct’ hot water cylinder refers to having a heating element in contact with the potable water in the tank.

An ‘indirect’ hot water cylinder is when a coil passes through the tank of potable water. The coil is fed by water from a heating source / boiler; the water that comes out of the tap is still sitting in the tank.

When I looked it up (Google), the definition seemed to be that if the water in the tank passed through the boiler, it was direct and if the bolier heated the water in the tank via water in a coil, it was indirect.

If you heat the water in the tank directly, then use a coil to heat the DHW, there is minimal risk of legionella forming so no heat cycle is required.

You can also have an electric element in either case (direct or indirect).

Well Mitsubishi decided to confound you both @Gwil and @borpin :slight_smile:

There is a 300L tank of potable water that is used in the taps (and showers).

In the old days that would have been heated by a coil snaking around inside the tank. However, Mitsubishi decided to put the “coil” outside the tank and pump the water from the tank out to the “coil”. Except that the “coil” is a plate to plate heat exchanger.

Here’s the huge amount of pre-fitted hardware that comes with the tank.

And here it is in real life:

You can see the pipes going into it on the left are the flow and return to the air source heat pump. The pipes going into it on the right are the ones going into the 300L of potable water. You can see the silver Grundfos pump that moves the water from the tank out to the heat exchanger.

So, I believe there is a risk because we have 300L of water just sat around. Normally we’d consume that in a day or so, but this time it was sat around for weeks.

Thanks for your comments.


Yes, definitely an increased risk.

Do Mitsubishi have any advice?

The bolt on plate heat exchange seems to be the preferred method for many heat engineers on twitter. Citing better performance than a coil I think?

You also don’t lose the volume inside the cylinder that the coil would have taken up.

The bolt on plate is also how Mixergy class their cylinders as “heat pump ready”. I will be having one (and a pump) bolted on to my Mixergy when the heat pump goes in.

Heat Pump Tank - Mixergy.


Thanks for spelling this out - I hadn’t quite twigged that this is how our EcoDans were plumbed, but it makes sense now that I think about. Here’s a diagram from Mitsubishi’s website:

Slightly off topic, but I notice that the temperature of the return on my system behaves very differently to yours, despite being quite similar. On mine, the return temp drops as soon as the DHW cycle ends, but quickly returns to what it would have been. It then gradually falls slightly faster than the flow temp for about 25 minutes, before jumping up to be close to the flow temp. I see this exact same profile every time, whereas David’s chart in the first post shows the return dropping to 30 degrees and moslty staying constant.

Anyone got any idea what’s going on here? I wondered if I was seeing a “wave” of heat travelling around the pipework from the hot flow pipe to the cooler return pipe.

1 Like

For comparison, and back on topic, here’s my 11.5 kW EcoDan heating up 250L from nearly cold (water out of the tap was barely warm). It took a little over 90 minutes to heat up to 53°C, delivering 10 kWh of heat with an average COP of 3.5. Outside was 19 - 20°C.

Note that I gradually increase the target temperature to keep the HP power draw to around 35%, so that it stays within the generated solar. I find that if let EcoDan do it’s own thing, it ramps up the compressor quite hard as it races to get the flow temperature up. I prefer it to be slow and steady, though I’ve not noticed that it makes any difference to the overall COP.

1 Like

Thanks for sharing @Timbones

Aren’t you tempted to just heat to 40 °C and get a better CoP? You’d have stopped at around 10:30 and you wouldn’t have gotten into the long tail of lower efficiency. I’m assuming you’d end up doing another heating run later once that wave of water was drawn down.

The way you are walking up the desired temp is really nice. I tried that with my 14kW but it just decides to shut off so I abandoned hope. I can do that walking with the space heating if it’s less than 9 °C outside (otherwise it races up too quickly and shuts down).

BTW you are doing way better than I am - my DHW CoP is usually 2 to 2.5.

1 Like

Crikey, that’s really weird. The only thing I’ve seen people talking about that could even vaguely cause something like that is a low loss header, but even with that I can’t see how you’d get the exciting situation you’ve got. I’m even more perplexed by the temps coming back together at 12:15 if you don’t have any pumping going on.

It’ll be interesting to find out what’s causing it when you tell us in a few years :slight_smile:

1 Like

On sunny days when I’m generating more solar power than I can use, I’ll heat the tank as high as the HP can get to (53-55°C), with a little bit of immersion to get those last couple degrees. Consumption on those days can be up to 3 kWh. I figure that when the electrical input is free, the achieved COP doesn’t really matter.

On cloudy days, I’ll only heat the tank up to 42°C, or maybe not at all if it was heated to a high temperature on the previous day. This reduces daily consumption to 0.5 - 1.0 kWh (excluding idle power), allowing me to put more juice into the battery while still having enought hot water.


1 Like

The simple algorim that I find works well is to target tank temperature + 10 or 12 degrees, updated every 4 minutes. This seems to be a large enough delta for the EcoDan to keep heating, while minimising power consumption. YMMV. I have a more sophisticated formula reduces the delta as it approaches the maximum, though I don’t think this makes any difference to overall performance.

For regular space heating, I only control the target temperature based on outside & inside temperatures, and let the EcoDan decide how best to meet that. Basically weather compensation + programable thermostat implemented in software. Less chance for system failure, and I can easily override in Melcloud. Quite happy with how that worked back in March, but will have to see how it fairs come winter.

1 Like

Is that including standby consumption? My average daily CoP during the last few months (i.e. no heating) has been closer to 2.3. The coefficient looks waaaay better when I disregard the idle power (~24 W)…

Really interesting thread here. I always find it hard to know when/how to chip-in, partly because I have far more questions than answers… i tend to go into thought-overload… hard to know where to start!. General thought,… its quite a lot of heat (14kW) to ‘squeeze’ into any heat-exchanger (coil or plate)., so flow/tank dt will be quite large. The flow temperature spends quite a bit of time at 60c, so COP unlikely to be very good. Shame you cannot ‘cap’ the output to say 7kW when on DHW.

1 Like

Yep, I always include that, I feel it’s the only way to keep me honest.

1 Like

Oh so very true @johncantor

At the moment I’m still not willing to hack into my heatpump because I want the RHI money and the warranty to be respected. I’ll tolerate some inefficiency until that expires.

It will certainly be nice when I can explicitly tell my Ecodan to modulate down and behave a bit more gently.