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

Dual heat pump systems

funkytut | Posted in General Questions on

I am in the process of upgrading my heating systems and am currently in the “oh crap, what have I done?” stage of the process.

I have a 3200 sf rambler with walkout basement in northern Minnesota.  It was built in 2007 and they called it “energy efficient”, but I think it’s pretty conventional 2×6 construction.
It has a LOT of windows and 3 sets of double doors, which also detract from its efficiency.

The current heating system is a Steffes 5140 hydronic boiler with in-floor heat throughout the house.  In the winter, the Steffes radiates a lot of heat and it’s common for the mechanical room to be very hot. 

The most distant loop, which includes the master bedroom, has trouble maintaining temperature above 62F when outdoor temperatures are -35F.
I haven’t yet determined if this is caused by furniture placement over in-floor heating coils or bad system design.

There is currently no cooling system, but summers are hotter than when I was a kid.

I’m planning on installing a 6-ton vertical loop system (six 200 foot bores spaced 12 feet apart) with a 5-ton heat pump in the basement.
I’m also planning on using that system to provide domestic hot water.

Because of the house design, the contractor is recommending I approach cooling and supplemental heat using a 2-head mini split air-sourced heat pump that supposedly works to -31F (Gree)
Trying to augment the ground-source system with the cooling/heating heads would require running water lines outdoors. (too disruptive to run them indoors). 

It bothers me that I can’t make use of the ground source for the cooling.  I’m spending extra money for a separate system, there will never be a point where I’m pumping heat back into the ground, and the efficiency will vary more than a ground source.

We’re on a lake, so the water table shouldn’t be a problem and I wonder if the “never pumping heat back into the ground” concern might not be warranted.

Some questions:
Can refrigerant, instead of water, be pumped to the heads from the heat pump in the basement?

Can a small hut be built to provide a heat exchange from the ground-source liquid loop to the mini-split system?  I think that would improve cooling efficiency on the hottest days and make it possible to run the heating system even on the coldest days.  It might also reduce maintenance by keeping crud out of the coils.

Would it be possible to do that in the basement, next to the ground source heat pump instead of in a little hut?

Will the heating system be able to keep up on the coldest days since the exiting water temperature will be cooler than the Steffes? 
Will a desuperheater fix that? 
How will that impact domestic hot water?

Does the domestic hot water heating work well in all seasons?   Minnesota can be very hot in summer and very cold in winter and my preference is to open windows when the weather is nice.

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. walta100 | | #1

    Have you done a fuel use study to size your heat pump see the link below.

    6 ton may well be the largest available and the default recommendation.

    It seems unlikely that switching to a heat pump that is likely to make much cooler water will improve the poor performance in the bed room.

    If you are on city gas changing to a heat pump seems likely to be a poor financial decision.

    Consider having blower door directed air sealing done along with thermal photography and increase the water flow to the bedroom.


    1. funkytut | | #3

      That's a good suggestion. Fortunately, my power co-op can provide usage data split into heat, hot water, and other. They also provide temperature information. Unfortunately, before I moved in, I kept the home at 61F while not here, so the usage history won't reflect future usage. (61 because of concerns about sheetrock integrity, I would have kept it colder otherwise)
      Even so, I was able to see that my peak monthly usage for heat was in January of 2022 at 6556 kWh. My peak daily usage was 343kWh during a different time. The Steffes system charges the heat reservoir during the period between midnight and 8am and on that day, it was pulling 44kW for almost the whole period. The hourly chart showed power usage tapering toward the end, so the reservoir did fully charge and did not begin the day at a deficit. (we see -35F just about every year for a short period)

      There are certainly larger heat pumps. The field is to be sized at 6 tons and the pump at 5 tons. I will review the above about peak usage with my contractor before proceeding.

      I didn't mention that I'm rural. I'm 15 miles from the nearest natural gas source, so electric, propane, wood, and sunshine are pretty much my options.

      I was informed that there wasn't much they could do about the bedroom, short of ripping up the floor and laying tubing at a denser pitch. I'm not sure what the drawbacks are of increasing flow rate. The heating guy I consulted did not feel that increasing flow rate would be possible, but I'm uncertain why. (turbulence?) Raising temperature might have an effect on the hardwood floors, but I can't imagine a small change would be a problem.

      I took a look at the article you linked and made some assumptions I'll run by you. My Steffes boiler is electric. There are about 4000 lbs of bricks which get heated to 1500F on the coldest days. I believe that resistive electric heat is 100% efficient, so the calculations that try to estimate in and out of BTU's could be reduced to how much power is drawn by my system.
      The highest daily load I saw was 343kWh, which translates to 19.5 hours operation of a 5 ton system. (17.58kW = 5T) I think your article points to a way to estimate usage if we were to raise the desired temp from 61 to 68, but I haven't worked it out yet. It's also a bit complicated because it's not clear how much charging would be from catching up the previous day's load and how much is from real time usage during charging.

      1. Expert Member
        DCcontrarian | | #4

        If you look at usage and degree-day info over a month the heat capacity of the bricks will even out.

        Heating load should be proportionate to the difference between indoor and outdoor temperature. If we assume that peak day was on a -35F day, 61F is a delta of 96F and 68F is a delta of 103F. That's a difference of 7%, so you'd expect your energy usage to go from 343 to 368kWh. The difference is probably higher than that because I imagine it wasn't -35F all day long, so the average deltas are smaller and the percentage difference is bigger.

        I missed the part about all the heating elements being in-floor. Yeah, there's no easy way to increase the capacity of an installed floor. The heat output is directly related to the temperature difference between the floor and the room. There's a limit to how hot you can make a floor and still have it be comfortable to walk on. Once you get there that's it.

        Does this room have a crawl space or basement or attic above or below it? What you probably want to do is put in another heat emitter, probably a wall radiator. That's easiest if you can just run some PEX pipes back to the boiler room.

  2. Expert Member
    DCcontrarian | | #2

    Agree with Walta on the fuel use study.

    Any system installation should start with a Manual J or at least a fuel use study. Your proposal sounds especially complicated and unusual, this is no place to be sizing by guess or by golly.

    Since it's just one room that can't reach temperature when it's very cold it's likely that the radiator in that room is too small. As Walta notes, the heat pump will probably have colder water than your current boiler so the output of that radiator will probably drop. I don't expect a minisplit to give meaningful heat at -35F so I don't see that helping in that room.

    The usual way to do cooling with a water-to-water heat pump is to use air handlers that can do both heating and cooling, something like this:

    Of course then we're talking about replacing all of your radiators too which makes an already-complicated project even more so.

  3. paul_wiedefeld | | #5

    Will the heating system be able to keep up on the coldest days since the exiting water temperature will be cooler than the Steffes?

    Yes - in-floor heating uses (or at least should) lower temperature anyway. Think ~100 degrees.

    Will a desuperheater fix that?

    Not relevant.

    How will that impact domestic hot water?

    A desuperheater is like a “trickle charger”: it saps a little bit of the capacity for DHW. It’ll have resistance backup, so it’s no big deal if 1 or 2 nights per year you might come close to maxing out the heat pump’s capacity for central heating alone - the tank provides a buffer and you’ll have backup. During the summer, you’ll have to run the geothermal heat pump in heat mode I guess to get it to work if you can’t get it to provide cooling.

  4. gusfhb | | #6

    The room that runs cold:
    Is the room leakier or more overglazed than the rest of the house?
    Is there carpet or something else that differs compared to the rest of the house?
    Consider a blower door test to find if there are defects particular to that room
    Is the rest of the house reaching temperature and the thermostat satisfied at very cold temps?
    If so, the way to balance a radiant system is you turn the heat down via loop valves in the rooms that are warm so that the colder rooms heat up.
    You don't increase velocity, but you can increase water temperature. Most radiant systems run much cooler water than the boiler creates, 100-140 deg F. Increasing the water temps a bit, then throttling back the loops that don't need it might help.

    If the room isn't leaky or carpeted, and tweaking the system is ineffective, they make panel radiators designed to run on cooler water that could be added.

  5. funkytut | | #7

    Thanks to all for your input. I will be doing additional testing this winter, including using my new thermal camera to pinpoint the locations of all the tubing and see if we were inadvertently covering up critical areas. We are aware that carpeting is not great and that furniture can interfere, so beds have to have legs, etc...
    I very much want to minimize the use of the heat from the Steffes, eventually eliminating it if possible.
    I have decided to discard the air-sourced heat pumps and plumb in the water lines to put additional heads on the new system in locations near where the air-sourced heads would have gone. This will allow me to supplement heat in that bedroom no matter what the outside temperature is, it will eliminate some unsightly refrigerant lines on the outside of the house, it will allow me to send heat into the ground source in the summer, it will save the cost of the additional air sourced heat pump system, and it will allow me to use a desuperheater for more months for DHW. My only concerns now are how much comfort this system will provide and how destructive this install will be. (sheetrock repair)
    Should I have more concerns? From my electrical usage analysis, I think I might want to upsize the capacity of the heat pump to 6 tons.
    Also, the ground source was going to be oversized initially at 6 tons. What kind of safety factor is built into the ground source size estimate?

    1. Expert Member
      DCcontrarian | | #8

      Please, get an independent expert to do a room-by-room Manual J.

      This is looking like an expensive undertaking. Spend a little money up front on engineering.

      1. funkytut | | #9

        I'm going to do a heating load study and am exploring doing it myself or hiring someone to do it. Looking online, I'm not having luck finding someone who does this. The fact that the heating contractor I am working with didn't offer to do it suggests they don't. I'm not sure if some of this is available from the builder, who likely has records, or some other entity who might have specified or approved the current heating system.

        Under the DIY category:
        A quick check using a block load on the CoolCalc website reports 69,00 heating BTUH but I don't know if they're reporting a peak rate or some sort of "smoothed" number. The construction details of the house also don't match the choices given in this tool that well, either. If 69,303 btuh ---> 5.8 ton system, then I imagine I should be looking at a 6-ton instead of a 5-ton system. This points in the same direction as the napkin analysis above where my peak draw was 343kWh but my thermostat was set low during that period. Adding DHW to the mix adds another half ton.
        I don't believe I need to worry about short cycling on this system. There's 50 gallon reservoir that's part of it and the estimate suggested another 10 gallons in the lines.

        The YouTube example I saw using CoolCalc showed that a block load overestimated the heating load, but I can't be certain if that's always going to be the case or if it could go either way.

        Taking a tape measure and having some information about surface area and R or U value seems like the best way to estimate heat loss, but I see that overhangs and air leaks matter, too, which complicates things.

        Under the HIDFY category: (have it done for you)
        Is there an organization that does this in Minnesota? When I search, I mostly come up with companies who have something to gain by selling you a heating system or general information about improving your efficiency using weatherstripping.

        Finally, I have additional questions regarding the vertical loop field. When they say that each 200ft borehole provides 1 ton of capacity, how much range does that estimate have and what assumptions are likely to be made? (We are in a sand/clay mix and our drinking water well is about 90 ft. deep. ) Do they do a field assessment when they drill? If the loop field can't provide the necessary source/sink, how easy is it to expand it? Should it be oversized by some amount to allow for the fact that heating load is going to be larger than cooling load? (the CoolCalc estimate showed 69,303 heating and 5,800 cooling btuh)

        1. Expert Member
          DCcontrarian | | #11

          With floor heat you really need to be doing room-by-room calculations.

          I know the tubing is already down, but room-by-room will tell you what BTU level each room will need. From that you can calculate floor temperatures, water temperatures and flow rates. That will tell you before you even begin whether it's going to be possible to hit your heating loads. Don't assume that it is.

          It happens all the time with systems like this that people just assume they're going to work. And then they don't.

  6. user-5946022 | | #10

    If your HVAC sub does not do Manual J calcs, you need a different sub.

    You can hire Energy Vanguard to do a Manual J if you can provide the plans and all the details they would need, such as wall thickness, insulation thickness, % of fenestration and specific location of each, specs of the glass, AND blower door test results.

    So we come back to did you do a blower door test? The cost is chump change compared to what you are getting ready to invest in this system. Get one done by a third party energy rater who hopefully knows what they are doing more than an HVAC sub. Be sure you have two smoke sticks and 2 rolls of blue tape and a helper available for yourself when you have the blower door done; go all over the house and mark where the air leakage is - maybe different colors of tape to signify different levels of leakage? A few kids to take cell phone videos of the smoke sticks would also be helpful.

    Also, you mentioned you are on a lake in MN. If you own part of the lake, have you considered laying your geothermal loop in the lake bed in lieu of the six 200' bores? Seems less expensive and easier to maintain. I knew someone who did this on a lake in MN in the early 2000's and it worked well.

    1. funkytut | | #12

      Thank you to everyone who's providing advice. I believe that following it will help me make sure that I'm not sorry about the purchase later.

      I'm looking into the Manual J and blower door testing.

      The lake I'm on is 112,000 acres, so it's a huge resource for this, but it's between 40 and 45 feet lower than the house and about 250 feet away. In recent memory, we've seen up to 52" of ice, so the loop would be better placed deeper than that, which would require another 100-200 feet of horizontal run to be sure. The lake bottom is lightly sloped.

      Some of the run would best be buried under the lake bottom to prevent damage when the ice starts grinding rocks around. That and the last 250 feet from the lake to the house would have to be buried on a hill. Horizontal drilling around here is problematic because of glacial till with lots of boulders. I understand that vertical drilling can punch through the boulders but that horizontal drillers have to start over and most contractors quit after a couple of tries.

      I do have concerns about the pressure on the glycol lines at 50+ feet of vertical drop and I think the process of burying them adequately would also be expensive. The 200ft boreholes are being quoted at $3500 per hole, which also makes for an expensive proposition.

      Adding the push of the ice sheet in spring that happens infrequently but can be destructive, muskie fishing in the summer over the area in question, and frost heave for the long run from the loop to the house, I felt that the wells would be more of a known quantity.

      I am open to suggestions if these issues have been dealt with before, but I'm operating without any real background in it. Sometimes my concerns are unwarranted.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |