GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Factors affecting CoP of Heat Pump (Air-to-Water)

advice_yesplease | Posted in Mechanicals on

Hello there

So I’ve had my air-to-water heat pump system doing its thing for nearly a year now, seems to work great. This winter I’d just like to tweak things a bit such that I’m overall getting more efficiency/higher CoP out of it.

Brief description of the system – main floor (~900sqft) is radiant in-floor embedded in 1.5″ concrete pour, and upper floor (about 1100 sqft with) has 3 zones (with about 30 lin/ft of hydronic baseboard per zone). Radiant in-floor has taco mixing circulator (to bring down water temp previous provided by boiler, still using it though for the lower temp output of heat pump), and the remaining zones use a simple single-speed circulator, controlled by zone valves. There are basement/garage zones that I leave unheated, but they also has radiant loops.

There is electric backup that I don’t have enabled (and whole house wood stove as another backup), but overall heat entirely with my 2 ton monoblock air-to-water heat pump and it seems more than capable (has kept up nights at -25C last year). System has a 40 gal buffer tank.

I keep reading the common wisdom that heat pumps run better when running continuously and at a low percentage of their total capacity, but I am wondering under what conditions this is true and to what degree. I read the NREL report ( which tests some variable speed air-to-air, which seems to suggest improvements to CoP with reduced outdoor unit power (Table 6) albeit they seem to not be that significant at outdoor temps 32F and below (bigger factor seems to be indoor unit fan speed).

Anyhow I am wondering how these impacts on CoP translate with my monoblock air-to-water heat pump and how I can improve overall HSPF. I have some good flexibility over how the unit operates (can tune the PID controller in the unit, set start/stop thresholds, limit capacity during certain times).

Here are the things I am considering so far – I plan to measure their impact once I get my energy meter/data logger but for now hoping to get your thoughts: 

1) Do what I can to make Make it run more continuously:

I figure that I can improve overall HSPF by avoiding startup/shutdown cycles as much as possible, that involves doing various things to extend run-time – most of the following points (2-4) result in extended run times which I figure is good. 

Once I get an idea on the actual effect on CoP of the heat pump itself I’ll try and factor in the increased operating time of the circulation pumps as I understand that may be significant (I have two pumps for the heating loop, one for radiant in floor – the taco mixer-, and one more for the upstairs zones).

2) Reduce output water temp:

I understand that overall CoP (and BTU capacity) is improved when output water temperature is lower, so I decided to move the output water temp from 45C to 40C. 

As an aside: should I also increase the circulating pump speed (the one that’s pumping water through the air-to-water unit) to reduce the DeltaT? Or slow the pump to increase DeltaT – it is a three speed pump

As the radiant slab is mixing to 35C, this doesn’t really affect the radiant floor dynamics other than taking longer to reach that supply temperature – so I think that’s fine.

This has more of an impact on the hydronic baseboards upstairs – they have a low output due to low water temps, but seem to be able to keep up just fine even at 40 C (though we keep the rooms quite cool, 17 to 19 deg setpoint) – have an electric boiler to boost the temp in baseboard loops if needed but so far unused.

2) Operate at low capacity at night:
The air to water unit has a “night mode” which limits overall capacity to 50% (for both outdoor fan and compressor) during night time hours, my observations so far is that this causes it to run all night especially on very cold nights to keep the main floor zone at its setpoint (where as at full output it was more intermittent, though cycle times were still a hefty 1-2h), 

3) Reduce cycle rate of themostats in upper floor zone 
I set all thermostats in the upper zone to have a reduce cycle rate (I think they are set for 2 to 3 cycles per hour max, or whatever the lowest cheapie honeywell thermostats allow). The cycle times are already long due to low output but I figure this will help extend them further.

I was kind of uncertain about this one as, when these loops do cycle on, due to infrequent use the “cold” water in the loop tends to shock the buffer tank temperature and increase output of the air-to-water heat pump if it was previously operating at low output. 

4) Heat up the slab in advance
I have a scheduled “slab heat-up” time that runs for approx 3 hours (from about 3pm to 6pm) – here I set the thermostat on the main floor such that, regardless of what the temperature is, the thermostat will turn on and try to bring slab temperature up.

Anyhow, this gives a continuous run for 3 hours, and it tends not to call for heat again until night time due to residual heat in the slab.

My theory here is that these hours 3pm-6pm are, on average, warmer and will give better CoP over the season. During the time the slab is heating up the room is typically either calling for heat already (on a grey, cold day) or already warmer than the slab (due to solar gain) – in the latter case I figure I’m not adding much heat to the room since air temps warmer than floor surface – so the heat is only released once nightfall hits and it starts to cool down.

I may add some weather prediction to this “slab heating” thing to avoid this when these hours are forecast to have colder air temps but have not tried that yet.

5) Avoid temperature setback
I tried this for a bit but since there is always someone home, the only times that make sense for a setback are at night 

It takes a long time to bring the temperature up even by just 1 degree (both in the main area due to slab mass and in upper floor zones due to low BTU output),

Any thoughts? Until I can actually perform some measurements, I am wondering if the above tactics seem like a good approach or if there’s any flawed logic? Anything else to try?

Also, from that NREL report the indoor unit fan speed seemed to have a big impact on CoP all other things equal (I understood that this is because it reduces condensor fluid temperature more, bringing it closer to evap temp which makes cycle more efficient?) – at a given compressor/outdoor fan output what would be the equivalent here in this monoblock air-to-water system where both the condensor and evaporator are outside? Have I done this already by reducing water temps? Should I increase speed of indoor circulating pumps to reduce the input/output deltaT?

Thanks for reading!!!!

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member
    DCcontrarian | | #1

    I would talk to the manufacturer. My experience is they have spent a lot of time tweaking for performance and they actually appreciate it when someone cares about it and want to talk about it.

  2. walta100 | | #2

    "1 Do what I can to make it run more continuously:"
    I think this point is over blown on this site. I do not think the start and stop penalties are as large as we are assuming.

    "2) Reduce output water temp:"
    I think you could make a big difference on this point.
    Everything you can do to reduce the delta between the outdoor air temp and the hottest water you are making will increase the system's capacity and efficiency.
    If the upstairs baseboards can’t produce the required number of BTUs at the new lower temp consider adding more baseboard.

    "2) Operate at low capacity at night:"
    Chousing to use this mode seems silly to me unless the system is so noisy that it keeps you awake.
    The more Kwh you burn at night the more efficiently your provider can operate its equipment and lowers it's costs.

    "3) Reduce cycle rate of thermostats in upper floor zone"
    You do understand when you increase cycles per hour you also increase the number of degrees between the cut in and cut out. So it will get colder before the heat comes on and warmer before the heat turns off. Are you willing to give up that amount of comfort?

    "4) Heat up the slab in advance:"
    Seems like another silly option best turned off. I think Heat pumps work best when you set a temp and let the system run unchanged for months at a time.

    "5) Avoid temperature setback"
    I used a big setback when I had a 3X oversized gas forced air system. It worked well enough but I had spots that would cool first and fall below the dew point and try to grow mold. I gave up setback in the new house with a heat pump and do not miss it. Seems silly to expect tons of concrete to chance temp much in a few hours.


    1. Expert Member
      DCcontrarian | | #3

      "3) Reduce cycle rate of thermostats in upper floor zone"

      He's got a 40 gallon buffer tank so this probably isn't going to affect heat pump operation at all.

  3. Expert Member
    DCcontrarian | | #4

    You don't say how your buffer tank is plumbed.

    Heat pumps are all about temperature deltas. You want the coldest water you can get going into the heat pump return, and you want the hottest water you can get going out into your radiators. Have you looked into 2-pipe and 3-pipe buffer tank plumbing?

    You'd get the highest efficiency if the water coming out of the heat pump goes directly to the radiators, and the water coming off of the radiators goes directly into the heat pump -- the way the system would work if there was no buffer tank. The problem is when demand for heat is low you need the buffer tank to keep the system from short-cycling. What 2-pipe and 3-pipe plumbing allow you to do is have your system behave as if there is no buffer tank when demand is high yet have buffering when it is low.

    Traditional buffer tanks have four pipes: the output from the heat pump goes into the top of the tank, and the input from the heat pump comes out of the bottom of the tank. The input from the radiators comes out of the top of the tank and the return from the radiators goes to the bottom of the tank. The water temperature only varies to the extent the tank stratifies.

    In a three pipe system, there is a tee at the top of the tank connected to the output of the heat pump and the input to the radiators. When demand is low more water will come from the heat pump than goes to the radiator, so hot water flows into the tank and the tank warms until the setpoint is reached. When demand is high, the same amount of water goes to the radiators as comes from the heat pump, so nothing flows through the tee into the tank. Under really high demand you can have water flowing out of the tank and the tank cooling. At the bottom of the tank the return from the radiators goes in and the return to the heat pump goes out.

    A 2-pipe system is similar, except that the cold side is teed the same way as the hot side. I wish I had good diagrams.

    I have a Chiltrix system, and it will modulate as low as about 25% of capacity. If the demand for heat is lower than the minimum modulation, it will heat up the buffer and cycle on and off. If the demand for heat is high enough, it will try to set the compressor speed and the circulator pump speed to match the demand for heat. I have a three-pipe system, and when it gets into this mode it will run for hours at a very steady pace. I believe that this is the most efficient way to run.

    As an example, the week before Christmas it was very cold in DC (for here), single digits. I have my water temperature set for 95F. I was monitoring the heat pump and it was very stable with the water coming out at 98.4F, return water at 89.6F, circulator at 4.2 GPM. (That's 19,360 BTU for the nerds). When it's doing that the buffer tank is doing nothing.

  4. Expert Member
    Akos | | #5

    Air to water is some ways are similar to a modcon.

    First step is set up an outdoor reset. You want the water to be as cold as possible and your zone circulators to run pretty much all the time. With a good outdoor reset curve and well set up zones, internal thermostats are pretty much redundant.

    The next thing to watch is flow rates. Most hydronic setup I've looked at are WAY over-pumped. You want nice big delta T across your emitters, 10F to 20F are typical but you can go for even more, for example on my modcon I run almost a 40F delta on some old cast iron rads.

    When you have different temperatures for zones, you can do some creative plumbing to squeeze extra efficiency out. For example, I have plumbed low temp and high temp zones in series before, the water from the high temp zone is still warm enough to feed the low temp zone. The return from the low temp zone is much colder than if I had them in parallel. Another thing you can do is use the return from the baseboards and T it into the cold water mix input for the floor heat to scavenge the left over heat.

    1. Expert Member
      DCcontrarian | | #6

      Chiltrix has an automatic outdoor reset feature:

      1. Expert Member
        Akos | | #7

        Just because it is there, it doesn't mean it is set up. My experience has been that outdoor reset is rarely configured properly by HVAC techs and sometimes not even set up even though everything is wired.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |