Comparing Minisplit Options
I have gotten quotes to add a mini-split system to my Climate Zone 6 (Concord, NH) 3-bedroom, 1.5 bath, 2-story (+unfinished poured concrete basement) Cape built to code in 2002/2003. Generally, all the installers agreed with my Schedule J estimate of a 3-ton unit being needed, but how they distributed the head units was different.
I am currently debating two quotes/systems out of 4 quotes that I received.
One is a 4-head Mitsubishi HyperHeat system using a 48k base MXZ-5C42NAHZ-U1 compressor rated at 20 SEER, 11 HSPF, 13.4 EER). The other is a Samsung MaxHeat using a 36k base compressor (that I believe is JXH36S4T, which is rated at 19 SEER, 9.8 HSPF, 11.75 EER).
My hope is to use the chosen system not only for cooling (rather than the three window AC units) but also heat for much if not all of the winter rather than my 83% efficient propane baseboard heat. However, I’m not entirely sure how to compare the systems. Looking at the ashp.neep.org site, I can see the Mitsu is more efficient (https://ashp.neep.org/#!/product/31908; https://ashp.neep.org/#!/product/31974) but I really am not sure how that computes to actual electric usage for the approx. 7383 heating degree days in Concord, NH. Same for the roughly 300 cooling degree days.
I did check the AHRI directory, but I’m not sure that its estimated costs per-year were much help. (Or, at least, I couldn’t figure out how they calculated those estimates.) Those estimates showed that the Samsung was cheaper to operate, which I’m not sure makes sense given that it’s slightly less efficient.
In addition to comparing the two systems, I’d really like to know if and when I should consider switching back to the baseboard heat in winter in order to minimize costs. I’ve heard ‘don’t bother,’ switch when temps hit about 10-15 degrees, and switch from about Dec. 1 – March 1 (the coldest three months). Can anyone give me some information to assist me in this regard?
Any help would be appreciated as this is not my area of expertise, and I’ve found it incredibly hard to find good information.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Natch,
How many squared feet of conditioned space do you have?
Oops, probably should have added that in my description. Conditioned space should be right around 1,800.
I'm not going to address the specific efficiency and operating expense differences between the two systems, but I would urge you to consider this problem from a 'systems' point of view.
Say you undersize the heat pump system on account of having an existing propane-fired baseboard system. Fine. But what happens when you need to replace the propane boiler? How are you going to feel spending ~$10-15K on a boiler replacement for a system that runs only a little bit each year?
I'm hoping not to undersize the system. By way of additional information, the Mitsu quote has heads of 30k (w/a spare space on the splitter should I need to add another one in the future say, for instance, if I finish the basement) with a max rated capacity of 42k, whereas the Samsung has heads of 36k with a max rated capacity of 36k. If you see anything wrong with the calculations, I'd love to know.
As far as running the systems go, unfortunately, past practice is not a guarantee of future performance. We're seeing increasingly variable weather patterns in my area, so whereas about a half-decade ago we were seeing regular intrusions of the polar vortex, bringing temps down to below -10 for an extended period of time (with routine dips below -25) on a regular basis, the last couple of winters we haven't seen that, with last winter's max low at my house being -12 with negative temperatures occurring, maybe, 4 days, and the weather hanging at or above 10 most days. Trying to look forward over the next 20 years, if we have more polar vortex visits due to instability in the global climate, then I'd have made a not-so-great investment on the heating side simply because I don't expect current day heat pumps to be able to produce heat, much less heat efficiently, much below their rated temps of around -13. So, in that scenario, a well operating boiler is necessary. If, however, we have winters like the last couple, then I've made a great investment on the heating side because while there will be a handful of days that are really low in temperature, it's not surpassing the ability of the systems to produce heat. So even if it's the case that my boiler does die, in that scenario, I can just take it off line.
At present, I'm using right around 1k gallons of propane per year, with around 3/4 of that going to my 83% efficient boiler. Although prices fluxuate, that's around $2-3k per year just to heat. Assuming that I understand the math well enough, I'd basically cut that price in half if I could switch to heat pumps full time. If I make a bad bet, there's a longer ROI and the system may not entirely pay for itself, but, hey, I'll have whole house AC which, with the trend toward more 90+ degree days in summer will be a blessed relief from my room-by-room window AC setup. If I make a good bet, then the system has entirely paid for itself and is actively saving me money.
So, ideally, I'll have a correctly sized system that will provide all the heat each year, but, at this point, I think I'm willing to accept the risk of that not happening.
It seems you should be able to know kWh per BTU at every outside temp. This would allow you to model the running cost of the proposed heat pump system.
Two complicating factors: Boilers are the longest lived heating systems. Compressors are the shortest lived. Also, in the future you should expect electricity to become very expensive on the coldest and hottest days.
It may be worth considering installing a less expensive mini split that doesn't perform well at the few days a year when temps are below zero. Then what would probably happen is that the house would become all heat pump in the future when the first heat pump system aged out.
>"It seems you should be able to know kWh per BTU at every outside temp. "
That's not really possible to do with any precision, since the kwh per BTU is not only dependent on outside & inside temperatures, but also the modulation level of the ductless. Taking a look at the references Samsung:
https://ashp.neep.org/#!/product/31908
Note that at +17F the COP at minimum modulation is about 2.8, and at max modulation it's only 1.8. that's a pretty big difference in $/MMBTU.
Natch,
It would probably be worthwhile to hire an independent engineer to size your system. (See https://www.greenbuildingadvisor.com/article/who-can-perform-my-load-calculations.) To my non-expert eyes, your proposed systems look to be vastly oversized.
FWIW. There have been lots of posts on GBA about the potential downside of using multi-head systems. You have a relatively small story-and-a-half house that might be better served by individual systems. From what I've read (and experienced), you want to lean toward undersizing the HVAC so it runs for long periods of time at low speed. That's the best way to keep your energy bills down while promoting comfort.
The Mitsubishi unit you posted shows a minimum heat output of 24k BTU/hr at 47 F outdoor temperature. This will be more than your load. The unit will cycle a lot and comfort and efficiency will suffer. Do not go with a multi split on an 1800 sqft house. The right one to one units will modulate down to ~2k -3k but/hr. A floor plan would be helpful to make additional suggestions.
I agree with Kyle_R that multi-splits are a generally poor choice for a 2 story house, and something that delivers 48K @ 5F is suboptimally oversized for the whole house load.
Why?
>"1k gallons of propane per year, with around 3/4 of that going to my 83% efficient boiler."
>" 7383 heating degree days in Concord, NH"
1K gallons @ 91600 BTU/gallon x 0.83= 76 million BTU
divided by 7383= 10,294 BTU/HDD
divided by 24 hours in a day is 429 BTU per degree-hour.
The 99% outside design temp for Concord NH is 0F (see: https://higherlogicdownload.s3.amazonaws.com/ACCA/c6b38bda-2e04-4f93-bd51-7a80525ad936/UploadedImages/Outdoor-Design-Conditions-1.pdf ).
Since the degree-days were in base 65F, that's 65F-0F= 65F heating degrees, for an implied design heat load of...
65F x 429BTU/F-hr= 27885 BTU/hr
...which at (27,885/1800'=) 15.5 BTU/hr per square foot, a credible ratio (or even somewhat on the high side) for a new-ish 1800' two story.
Even using ASHRAE's recommended 1.4x oversize factor (reasonable, for non-modulating equipment) you'd be looking for something no bigger than 1.4 x 27885 BTU/hr= 39,039 BTU/hr @ 0F, but for modulating ductless w/backup auxilaray heat it would be more comfortable an more efficient to go with an oversize factor between 0.9x to 1.2x. The 3 ton Samsung is good for about 35K @ 0F, which would be an oversize factor of ~35K/28K= 1.25x which isn't terrible, but the heads/cassettes would still need to be reasonably well matched to their zone loads.
As a general rule it's easier to size systems correctly in a 2 story using two systems, one for each floor. (I suspect your propane burner is also zoned by floor ?) If the room to room temperature balance is reasonable using the propane burner you may be able to estimate the room by room proportional loads by the amount of radiation in each room, but without more precise duty cycle information on the zone calls it's not going to be super-accurate for determining the total zone loads.
When looking at modulating ductless systems it's important to run fairly aggressive Manual-J-ish load calculations to avoid oversizing the heads. A utility consortium in the northwest developed a handy (and free) HVAC design tool (targeted at HVAC contractors) that comes with a decent easy to use load calculator, that unlike most calculators uses appropriately aggressive default assumptions on U-factors, etc.
http://hvac.betterbuiltnw.com/Account/Login.aspx?ReturnUrl=%2f
Take your time and enter the data, and see what you come up with. If it's within 15% of the crude fuel-use based load estimate for the total whole-house load (aka "block load") it would be good enough to use for sizing ductless systems. If you can come up with wintertime-only fuel use and exact filling dates for the propane to correlate with a nearby weather station's HDD data the fuel-use based block load will be pretty accurate, and you can use that to scale the room load numbers that the Manual-J tool comes up with. A more complete explanation on fuel-use load calculations can be found here:
https://www.greenbuildingadvisor.com/article/out-with-the-old-in-with-the-new
(^^^The same bit o' bloggery Akos linked to below.^^^)
No need to guess on sizing, run through the calculations here:
https://www.greenbuildingadvisor.com/article/out-with-the-old-in-with-the-new
Generally for a reasonably well sealed and insulated house, the best option is either a fully ducted system for the whole house. If you have an open main floor you can also go for a single wall/ceiling mount on the main floor and a slim ducted unit for upstairs.
A wall mount in every bedroom on a multi split is a very bad idea from comfort and efficiency. Bedroom loads are so small that even the smallest head is about 3x oversized, no way to make it work well.