GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter 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

Cost and effectiveness of GSHP vs. minisplits in Zone 5

Ben Rush | Posted in Mechanicals on

I understand that the installed cost is much higher for ground-source heat pumps (GSHPs) than for mini-splits.

I had thought this was due to the cost of excavation, but I recently read (here, in one of Martin’s posts from 2013) that the GSHP equipment cost is much higher too.

To address the first issue, could the ground loop go under a basement floor? (New construction, presumably- or in my case, a replacement basement. But that’s another story.) Would any additional excavation be required? How much? Would 2 or 3″ of sub-slab foam insulation be enough to separate the conditioned basement from the year-round +/-55F (zone 5) soil? Bottom line: would this be an inexpensive- yet effective- way to install the ground loop?

On to the equipment cost issue: What makes GSHP equipment expensive? Is it so different from CAC, window AC, mini-splits, or refrigerators?

I really like the idea of GSHPs, for two theoretical reasons, and one practical one: A) In zone 5, the soil is cooler than the air in summer- and warmer than the air in winter. Why would I want to put heat into 90F air or take heat out of 10F air? B) the volumetric heat capacity of soil is about 1000 times that of air, and C) in Chicago, occasionally it might be too cold to heat with a mini-split (Not sure if that’s three times a year- or once every three years- but it could happen.), but it will never be too cold to heat with a GSHP.

I’d appreciate any insight you can provide.
Thank you,

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. GBA Editor
    Martin Holladay | | #1

    There are three problems with your suggestion.

    1. Loops need to be in deep trenches, with lots of soil on all sides of the tubing. Digging deep trenches under a slab undermines the footings, and it's expensive.

    2. There isn't enough area under a basement slab to accommodate the length of the required loop.

    3. During the winter, this type of ground loop lowers the temperature of the soil, and an installation like the one you propose risks freezing the soil under your house -- an undesirable outcome.

    Some theorists think that the high cost of ground-source heat pump equipment is due to the 30% federal tax credit -- in other words, manufacturers keep prices high because they know they can. The equipment cost is subsidized (and, arguably, artificially increased) by the federal government.

  2. GBA Editor
    Martin Holladay | | #2

    You wrote, "In Chicago, occasionally it might be too cold to heat with a mini-split. (Not sure if that's three times a year- or once every three years - but it could happen.)"

    Homeowners in Vermont who have been heating their homes with a Mitsubishi or Fujitsu minisplit are finding that these units work well, even when the temperature drops to -20°F. So I think that your concern is misplaced.

  3. D Dorsett | | #3

    It's sometimes possible to drill WELLs under the slab suffficient to heat & cool a high-R house, but a trenched slinky approach typically has a geographical footprint 5-10x the footprint of the house. In The Netherlands (where the water table is high everywhere drilling through layers of peat, sand, and clay is cheap & easy, and the outside design temps are modest) single-well systems taking only a few square meters of real estate have been used to heat & cool high R row houses. It may or may not be cheaper than mini-splits, but the European preference for hydronic heating systems with low temp panel radiators or radiant floors makes GSHP a reasonable choice when it can be done cheaply.

    It's a more challenging problem for a code-min house in Chicago, which (unlike NL) has a significant cooling load, much lower 99% outside design temps, and different geology.

    There are many houses heated with mini-splits with outside design temps much lower than Chicago area ~0F (give or take a few) design temps. The primary difference is in net efficiency, and upfront cost, and heat distribution. A typical GSHP will deliver an annual COP of about 3.5, best in cla$$ systems will be a bit north of 4.5. Five years ago a right-sized ductless system would tip just north of an average COP of 3, but typical systems would run between 2.5-3. The cool climate mini-split technology has improved in the past 5 years, and if done right it's possible to hit north of 3.5 with ductless systems, north of 3 with best in class mini-ducted mini-splits. From a net cost point of view, it's often cheaper to go with mini-split and a slightly larger rooftop PV array to cover the additional power use of the lower efficiency than it is to go with better-class GSHP, but pricing varies (a lot) from location to location.

    A retrofit project I was involved with a handful of years ago is a 3-story house heated by a single Mitsubishi -FE18 mini-split per floor. The 99% outside design temp at that location is +5F, but it regularly hits below 0F, and goes to negative double digits at least once every five years. This past February the local temps hit -16F, and didn't break out of positive single digits for a few days, but the mini-splits had no problems keeping up, despite having no specified output capacity below -13F. The total cost for the three mini-splits was about USD$13K, about 1/3 what it would have cost to install a single GSHP system big enough to handle the load in my area. Even if you discount the pricing of the GSHP system due to the subsidy, say it can actually be installed for half the numbers being quoted, it's still more expensive. (That may not be true in every region, but it is in my area.)

    The cost adder of GSHP has to do with the ground heat exchanger, the lower production volumes of the heat pump hardware (compared to mini-splits, which are a mass-produced commodity in a very competitive market), and every system is a semi custom design requiring a lot more engineering/designer time. It's more labor to install a GSHP due to the ground loop, and there is greater design risk. The actual (as opposed to theoretical) efficiency of a GSHP system is in the hands of the designer, whereas mini-splits are a "system in a can", with far fewer ways to screw it up. That doesn't mean it's impossible- the more idiot-proof you make things, the more creative the idiots become.

  4. Charlie Sullivan | | #4

    Dana summarizes the situation and the reasons for it really well. There are a few recently emerging things that could make GSHPs a little more competitive with mini-splits, but I think they only narrow the gap a little--mini-splits are still a better deal.

    1) For a while, mini-splits all had modern variable-speed compressors while GSHP has single or two-speed compressors. Now there are some variable speed compressors available. This allows higher efficiency in part-load conditions, i.e., most of the time. With variable-speed compressors in both, the efficiency advantage of the GSHP becomes stronger.

    2) There's a new approach to drilling wells for GSHPs starting to be used in some places: Using directional drilling systems to make a borehole that goes down diagonally, instead of going straight down deep. It sounds like that should be more expensive because it's higher tech, but the raw power needed is less, because you don't need a crane capable of lifting 450 feet of drill shaft straight up, and you are less likely to need to bore into bedrock. Also, you can drill multiple short wells going off in different directions from a single point, reducing the amount of trenching needed to reach the heads of a set of conventional wells. Whether you'll really be able to get a cost reduction from this approach depends on the capabilities and pricing structure of local drilling companies, but what I've read indicates that you can achieve a substantial price reduction

    Insofar as that approach allows you to use a larger number of short wells, if you plumb then in parallel you can also achieve reduced pumping energy.

    Of course, both of these advantages are only realized if you find a team of professionals who know how to design and implement a system that takes full advantage of them, and who charge a reasonable price for the service. With new approaches like directional drilling that means either finding someone who is already ramped up in applying the new approach, or someone who sees the potential in it and is willing to invest the extra time and effort to get familiar with it without charging you for that time. Otherwise, you might pay the same as you would for a vertical well system, or even more.

  5. Ben Rush | | #5

    Great info.
    I didn't realize the slinky approach needed so much area. It's hard for me to let go of ideas that I like; usually I'll have a few "but what if...?" questions- but in this case the required slinky area and the available house footprint aren't even close: maybe even an order of magnitude apart.
    I also appreciate the points about the efficiency of high volume manufacturing of mini-splits, the possible unintended effects of subsidies, the reduced risk of screwing up a mini-split installation, and the challenges in finding a team of competent professionals... to apply a new approach... economically.
    Thanks guys!

  6. Expert Member
    Dana Dorsett | | #6

    I can't tell you tne number of truly insane mini-split proposals I've reviewed in the past (I usually stop counting stuff like this at three, after that it's "many" :-) ), but it's important to understand the magnitude of the loads correctly and size the equipment optimally to really get the efficiency out of the. The hacks all seem to think that the fact that they modulate speed up and down means you can just randomly oversize them and it'll be OK, but the modulation range is not infinite. It doesn't take hard math or real engineering to get it right.

    The proposals I really despise are those with a ductless head in every room with the minimum output of the combined compressors exceeding the 99% design heat load, right up there with proposals for equipment that has no specified capacity at the 99% outside design temperature for the location. There are mini-splits with specified capacity as low as -22F (wet bulb) or -15F (dry bulb) out there now- there's really no excuse, but they're out there! If you get it wrong by either gross oversizing or inappropriate equipment designed for more temperate climates it's pretty easy to hit a seasonal COP of only ~2 in a Chicago type climate. Happens all the time...

  7. Reid Baldwin | | #7

    The source/sink of the refrigeration cycle in a typical GSHP is actually the water/glycol mixture. There is an additional heat exchanger between this and the 55F ground. Heat transfer to a solid is really inefficient compared to heat transfer between fluids. So even though this heat exchanger (the ground loop) is orders of magnitude larger than a typical heat exchanger, there is still a substantial temperature difference. Consequently, the sink is warmer than 55F in the summer and the source is colder than 55F in the winter.

  8. Ben Rush | | #8

    Remember the dumbfounded look on Nigel Tufnel's face in This is Spinal Tap when he pointed to his amplifier and said, "But this one goes to 11"? He just couldn't comprehend that a different label on the volume dial wouldn't actually change the volume.
    Now picture him (or me) in my un-conditioned, un-insulated basement, in the summer, saying "But it's cooler down here". It drives me nuts- thinking there must be some free cooling down there- and I imagine contraptions to take advantage of it. I'm (almost) certain that if I had a better grasp of math and engineering, I'd know to disregard these ideas sooner. Are these as dumb as the misguided concept of directing a dryer vent inside?
    Dumb idea #1: A return duct near the ceiling on an upper floor- that takes warm summer air and sends it to a duct manifold under the basement slab. Not a replacement for other cooling. Just a free addition. Requires a fan, and airtight ducts (glued PVC?) under the slab to avoid water and radon infiltration). Just turn it off in the winter.
    Dumb idea #2: Same concept, but skip the sub-slab manifold. Just blow basement air to upper floors- or vice versa. Requires the absence of sub-slab insulation, and the use of moveable insulation on top of the slab- that can be removed for summer- but installed for winter. Wrestling mats, maybe?
    Dumb idea #3 (bonus!): I wish I could heat my house with junk mail. Free fuel, delivered! Obviously, if this could work, people would already be doing it. I suspect there's not enough BTUs, even in all the catalogs we receive. Plus, it would need to burn really clean, or it would negate the "green" cost. Actually, I probably do have the math skills to calculate the BTUs in my junk mail...
    Dumb idea #4: Can I get some free cooling from my 55F tap water before it goes to the water heater? It seems strange to pay to heat water and cool air- separately- at the same time! This exact issue may have already been put to bed with the invention of the HPWH.
    As always, I deeply admire your knowledge. Thanks for putting up with my uneducated ramblings.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |