How to reduce high energy costs for ductless minisplits
Hello Martin,
I have the Mitsubishi H2i heat pumps as my primary heat source at my south east MA home. Seven indoor units [9 and 12K’s] and two outdoor units [MXZ-3C30NAHZ, MXZ-4C36NAHZ]. The 2,400sqft 1980 contemporary style house has lots of [1st gen. Anderson double glazing and Pella single with interior “storm” panels] windows. The Pella’s are actually [relative] better thermal performers. The wall insulation is typical of that generation, poor to fair. The attic has about 20″ blown in insulation.
The electrical heat pump usage cost is more [maybe up to 25%] than for most expensive oil at very cold winter [about 3 winters ago]. Nothing has changed with the house since the oil heat, except some insulation improvements.
Been scratching my head, since all independent research and Mitsubishi specifications say that I should be seeing maybe 25% savings over oil.
I came across your 6/19/2015 article How To Buy a Ductless Minisplit.
You clearly define what I have been experiencing with my heat pumps at my house. That is, their efficiency is much lower [maybe 50-65% at real cold temperatures [below 20F].
Thank you so much for the great information.
Maybe you can be of further help. My contractor is basically not competent in this arena and Mitsubishi does not seem to want to be helpful.
Where should I turn to start to resolve my heat pump problems? I have contacted Mitsubishi Customer Care, but I’m not sure that will lead to substantial solutions.
In advance, thank you so much.
Kazirode
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I'm sure the experts will weigh in shortly- but its possible that your system is dramatically oversized. They can modulate way down but seven is probably too much and is reducing efficiency. . Perhaps try disabling 4-5 of the indoor unites and see what happens?
Rick,
Thanks for the response. My apology, I forgot to include that I have two of the indoor units shut off [two bedrooms]. When it's below 20F, it won't be comfortable shutting down additional units. The four units downstairs struggle maintaining 68F/65F. And, the one in my bedroom has to be running for 68F/65F with doors closed.
Kazirode, did you have a Manual J done to evaluate the room-by-room loads? At this point I'd suggest two paths: you've gotta make sure the units are working properly (sure sounds like they're not), and you gotta figure out exactly how leaky your house is. A competent technician has to found to do the first; as for the second, a few hundred bucks would hire an energy auditor. Their blower door measures airtightness and infrared camera pinpoints exactly where the wind is blowing in and where your insulation is weak. It's been a cold winter, but even with airtightness problems those units shouldn't have any trouble keeping you toasty.
Kazirode, I looked up the stats on your outdoor units using NEEP's data: http://www.neep.org/initiatives/high-efficiency-products/emerging-technologies/ashp/cold-climate-air-source-heat-pump
At 5 degrees, your units have a minimum output of 12,500 btus/hr and 22,500 btu/hr, respectively. These are probably on the big side but don't seem egregiously over- sized. The COP at 5 degrees for each outdoor unit at the minimum output is 2.29 and 3.71. These are pretty good numbers from an efficiency standpoint when compared to your old oil furnace which was maybe 0.85 COP at best.
Oil might be cheaper that electricity in MA but, like you, I would still expect your mini-splits to be cheaper than oil heat. There has been talk on GBA about remote temperature sensors for the mini-splits as well as proximity of the units from the ceiling. I'll see if I can find some links and post them here when I have more time.
This sounds like the "ductless head in every room" oversizing problem.
The 3C30NAHZ is good for 28,000 BTU/hr or more at the +10F (give or take 3) SE-MA type outside design temperatures, the 4C36NAHZ is good for over 40K.
http://meus1.mylinkdrive.com/files/MXZ-3C30NAHZ_Submittal.pdf
http://meus1.mylinkdrive.com/files/MXZ-4C36NAHZ_Submittal.pdf
So combined that's about 68-70,000 BTU/hr of capacity at design temperature, for a load that is dramatically less than that. Without enough zone load the heads will cycle on/off a lot, taking an efficiency hit, and if the total load isn't big enough to keep the compressors from cycling on/off it's an even bigger hit.
A typical reasonably tight 2400' sub-code-R 2x4/R13 house with wood or vinyl sash clear glass pane windows + storms and R19 in the attic and no foundation insulation would come in at ~45,000 BTU/hr @ 10F if it's a single story + full 2400' basement with NO foundation insulation, ~35,000 BTU/hr @ +10F if it's a 2 story or single story with insulated basement, a bit less if its a 2-story with insulated basement. There are lots of exceptions (in both directions), but odds are pretty good the load is at or under 40K, making the ductless well over 1.5x oversized for the whole house load.
A 40K load @ 10F outdoors, 70F indoors is a 60F temperature difference and a load of about 667 BTU/hr per degree F below 70F. So at 47F, it would be a temperature difference of 23F, and a load of about 23 x 667F= 15,341 BTU/hr. The minimum modulated output of 3C & 4C compressors @ 47F is 7200 BTU/hr each, 14,400 BTU/hr total, a bit less than the approximate load, so you're pretty reasonably well off from a compressor cycling point of view, as long as the combined zone loads are reasonably proportioned between the two.
"... four units downstairs struggle maintaining 68F/65F..." probably means they're reasonably well sized to their loads. They should be running CONSTANTLY, almost never cycling off (except during defrost) whenever it's below 47F. There is usually a temperature offset between the sensed air temperature at the head and the average room temperature. That offset can be resolved by wall thermostats, if you like, which would adjust the output of the heads automatically to make it seem less like they're "struggling", and actually maintaining the setpoint without having to bump the setpoints up with the remote
If the other 3 zones on the other system aren't "struggling" with an offset between the programmed setpoint and the room temp a low outdoor temps it means they're probably oversized for the loads, and cycling the compressor on/off even during colder days, which takes a toll on efficiency.
If you've been setting back the temperatures overnight or away at work, don't. That usually increases energy use, since it has to run at a lower efficiency during the recovery ramps back to temperature. Even though the efficiency is slightly lower at the higher indoor temperature, the efficiency during the recovery ramp back to temp is MUCH lower, more than erasing the savings of being at a lower temp.
Oil prices are volatile. Oil pricing in MA is currently about the same as it was during the 2014-2015 winter, but 15-20% higher than last year, and 35-40% lower than during the winters of 2010-2011 through 2013-2014. Even if it turns out your ductless system when operated optimally is still 25% more expensive than oil heat of 3 winters ago, there's no reason to believe that would be true over the lifecycle of the system. For historical MA retail heating oil pricing, use the widget on this page: https://www.massenergy.org/heating-oil/historical
With the build out of offshore wind (MA has a mandate for that), grid storage (MA has a mandate for that too) and transmission lines from Canada (a work in progress, but a policy goal of the Baker administration) the medium term effect on retail electricity pricing will be downward, since wholesale electricity pricing much less vulnerable to natural gas price volatility. There is currently extreme spot price volatility due to limited gas pipeline constraints, and an over reliance upon combined cycle gas power generation in this region (about half the annual grid mix, but a much larger portion of the peak power.) The Polar Vortex event of 2014 caused retail electricity prices by quite a bit in one jump due to the crazy-high spot market pricing on the wholesale marktes, but they have begun to moderate, and should continue to decline with each rate case adjustment (if slowly) as more zero-marginal-cost renewables come online.
If the basement isn't air sealed an insulated, if you intend to live there for awhile it's worth air sealing and insulating at least the band joist & foundation sill (usually subsidized heavily by MassSave rebates), and if possible insulating the walls (even though MassSave doesn't subsidize that). Insulating an unfinished basement is a reasonably inexpensive DIY project if you're handy, but much more expensive if done by a contractor. If the basement is built out and finished it's a more daunting financial proposition, but still worth considering.
Rick: "At 5 degrees, your units have a minimum output of 12,500 btus/hr and 22,500 btu/hr, respectively."
The min-output @ 47F specs in the NEEP spreadsheet are substantially higher than in the submittal sheets from the manufacturer.
The would yield a substantial efficiency hit at this level of oversizing if the minimum output accurate if the minimum output at design temperature is higher than the actual load at design temperature.
Do you know if there is an explanation for that difference? Does the "minimum capacity" in NEEP terms mean something different than what it means on the submittal sheets?
Dana- we all learn a great deal when you respond to these posts. Thank you!
If you go the NEEP website and click on the excel spreadsheet in the upper left corner- they have data for every ductless mini split available. If you scroll to the right, you will see data for 47 deg, 17 deg, and 5 degrees (including min.max output and COP for each temperature) This is where I found the data.
NEEP has written about how ductless heat pump manufacturers can choose which output range they base their COPs on. Its hard enough to find specific data from each company so they took the time to build this incredible database.
NEEP has a great lecture about cold weather heat pumps- I've posted the link below.
https://youtu.be/lcWxwq_-Z_U
Dana- check out the video link that I posted. Most of it- if not all of it will cover topics that you learned 15 years ago. But, fast forward to 16:00 in the video and the author begins to talk about how manufacturer's data can be different from the "AHRI" figures. The significant point here, according to the author, is that manufacturers can choose the air-flow rate coming from the compressor to rate the output of the unit. They make the case that because of this, the stated manufacturer's rating is, on its own, an unreliable source for sizing the heat pump. Hence, the built this wonderful database.
Perhaps all of this is documented within the submittal sheets- I'll have to track them down to take a look.
Thanks Rick!
The submittal sheets are here:
http://meus1.mylinkdrive.com/files/MXZ-3C30NAHZ_Submittal.pdf
http://meus1.mylinkdrive.com/files/MXZ-4C36NAHZ_Submittal.pdf
Marc's video discussion about AHRI ratings doesn't discuss the minimum modulated output at different temperatures, only the difference between maximum capacity and AHRI rated output.
There is also an error in that part of Marc's discussion. The AHRI rating is for ARI heating zone IV , based on the number of HEATING HOURS per year, and NOT the DOE climate zone 4, which is based on annual heating DEGREE DAYS. ARI zone IV temperate climates can have many more heating hours than a DOE zone 4 location with many more degree-days, and the maps are quite different:
http://www.fsec.ucf.edu/en/publications/html/FSEC-PF-413-04/images/Figure5_lg.gif
https://www.greenbuildingadvisor.com/sites/default/files/images/DOE%20climate%20zone%20map.preview.jpg
For example, coastal Washington state is DOE zone 4C, but ARI zone VI (Roman numeral six), with many more heating hours per year due to the cool but not cold ocean breezes. Coastal WA has many more heating HOURS than most DOE zone 6 locations, despite many fewer HDD.
Thank you Rick, Andy and Dana for the insightful comments and suggestions. Can you suggest good sources for very good professionals in the insulation and heat pump fields? I’m having a tough time finding both. I've been working with MA Save, but that service, although very good, doesn't reach the insulation issues discussed here. I tried to post my comments to all of your comments and suggestions, but got blocked by the spam filter. I also tried attached multiple Flir images, but looks like only one image could be attached, as the next image to be attached over rose the previous.
Heyoka Solutions in Falmouth did a very credible Manual-J on a house I got involved with on Martha's Vineyard a few years ago when I used to specify a mini-split solution. I don't have the files on this computer, but I believe the guy who ran the numbers was Paul Raymer(?).
Whether it was him or not, they have the certifications and talent for specifying insulation upgrades & heat pump equipment.
http://www.heyokasolutions.com/aboutus.asp
Hello Dana,
It has been a while, Heyoka Solutions isn't taking on new customers, but they recommended Chris Larkum. He confirmed that the heat loss/load was calculated correctly. and I'm still struggling with Misubishi and finding an excellent contractor for my high operating cost mini splits. A contractor that I brought in to do a system "commissioning" didn't find anything obvious, except one of the exterior units has a small refrigerant leak which we will repair. That's probably not causing about 2 to 2 1/2 times the cost of oil at the coldest winter and highest price [($3.80/gal). With all in the background, it is our guess that the main reasons for extremely high operating costs are the minimally adequate insulation and the early 80's windows as well as that the heat pumps are not efficient [operating cost] when temperatures drop below 25F. Since the indoor units are located near the ceiling, that probably contributes quite a bit into this equation. I'm in Norfolk county, MA area. Can you recommend excellent heat pump contractors and insulation experts to help a disappointed Mitsubishi mini split heat pump system owner?
Hello Dana,
It has been a while, Heyoka Solutions isn’t taking on new customers, but they recommended Chris Larkum. He confirmed that the heat loss/load was calculated correctly. and I’m still struggling with Misubishi as well as trying to find an excellent contractor for my problems with my high operating cost mini splits. A contractor that I brought in to do a system “commissioning” didn’t find anything obvious, except one of the exterior units has a small refrigerant leak which we will repair. That’s probably not causing about 2 to 2 1/2 times the cost of oil at the coldest winter and highest price [($3.80/gal). With all in the background, it is our guess that the main reasons for extremely high operating costs are the minimally adequate insulation and the early 80’s windows as well as that the heat pumps are not efficient [operating cost] when temperatures drop below 25F. Since the indoor units are located near the ceiling, that probably contributes quite a bit into this equation. I’m in Norfolk county, MA area. Can anyone recommend excellent heat pump contractors and insulation experts to help a disappointed Mitsubishi mini split heat pump system owner? I'll follow-up with more details as we progress.
Multi-splits are not as efficient as modulating mini-splits even when the heads are properly sized to the loads, but even at 22 cents/kwh it shouldn't be costing 2 - 2 1/2 x the cost oil @ $3.80/gallon even at ridiculous oversizing factors. But in your original post you claimed:
"The electrical heat pump usage cost is more [maybe up to 25%] than for most expensive oil at very cold winter [about 3 winters ago]."
So, which is it, something like 2.5x the cost of the oil system, or is it more like your original 1.25 x cost estimate? (That's a pretty big difference in estimates!)
A "small refrigerant leak", often translates into a "large efficiency hit" if it's been going on awhile. I know of a mini-split in western Washington that was running at a COP of about 1 (but still heating the place) for a full season due to low refrigerant charge before it was properly diagnosed and repaired. (It was not diagnosed by the installer, but by a researcher working for the NEEA's Northwest Ductless project, who was measuring and monitoring in-situ efficiency.) That unit happened to be an LG, but any mini-split that is very low on refrigerant will have appallingly low efficiency, using 2-3 times the amount of power that it should, and running out of capacity if it was sized properly to the load.
If one unit has been low on refrigerant for a full season the leak could easily explain the higher than anticipated operating cost at a 1.25 x multiplier, but both units would have to be low to explain a 2.5x cost multiplier.
I have no direct experience with contractors in Norfolk county to recommend. I would expect a "Diamond" contractor in Mitsubishi's "installer finder" listings should be able to diagnose and correct any defects in the installation (incorrect refrigerant charge, etc), but not the insulation/weatherization picture. It's likely that the recommisioning contractor is already on that list.
Click on the "Diamond Contractor" tab and give them your ZIP code:
https://www.mitsubishicomfort.com/get-started/contractors
Hello Dana,
Thank you for the updated information. Very helpful. Maybe someone else will have suggestions with contractors. I tried a few from the diamond contractor list, but haven't been too successful. We even had the Mitsubishi regional service manager do an inspection.
Oops about my 25% vs. 225-250% mistake. Initially, I forgot to include the [12month] solar generated electricity in the usage equation. It is about 225% [more over oil]. But, I haven't calculated in the added A/C usage [which I didn't have before the heat pumps].
I think I will undertake the insulation upgrades myself. I'm leaning toward closed cell 4x8 panels on all outside walls and foundation. I'll use the GreenBuildingAdvisor resources to make sure I'm using best practices and materials for my particular case. I'll probably undertake the window sash upgrades myself as well. I guess an excellent insulation consultant is what I'm also looking for.
Hello GBA contributors,
Thank you so much for all the help. All of your information has been very helpful.
It seems “we have what we have”. Once the refrigerant leak is repaired, there is very little more that can be done with my Mitsubishi mini split heat pump system.
In the hopes of helping others out, I wanted to share some things I've
learned from using heat pumps for heating and cooling a late 70's early 80's era contemporary home in Massachusetts (with solar panels).
Here are some important things to consider when using [Mitsubishi]
hyper-heat mini split heat pumps in colder climates like the Northeast:
1. Electricity cost in MA is double of what it was about three years ago. MA
electricity cost is one of the highest in the nation.
https://www.chooseenergy.com/electricity-rates-by-state/
Make certain to use electricity costs for your area and not average
electricity/unit cost for comparing different energy sources. When I did my
upfront research, I don’t remember seeing this caution in the researched
information, which included many independent efficiency case studies. I’m
sure the caution was there, but it couldn’t have been too prominent as I don’t remember seeing it. As a result, system operating costs are very high, possibly 200+% more than most expensive oil [three winters ago].
2. Indoor mini split heating units are usually mounted at the ceiling.
Since heat rises, and since we live closer to the floor, comfort is a challenge.
Extra energy must be used to compensate for the ceiling level mounting of
the indoor units. This problem is compounded with the location of the
temperature sensor [see 8 below]. This is not an issue for the A/C [since heat rises].
3. If the house has a lot of windows (e.g. contemporary-style homes), especially large ones, problem identified in number 2 is compounded.
4. If the basement is not conditioned and the basement ceiling, foundation, and walls are not sealed/well insulated, heating on the upper floor will be a big challenge.
5. Similarly, if the home's walls and attic are not very well insulated, the units will need to work much harder to achieve the set indoor temperatures.
Numbers 2, 3, 4 and 5 were verified with thermal imaging.
Although the above issues are also valid for oil or gas heat, since the
baseboards are located at the floor and under windows, there is no
pronounced heating problem when comparing apples to apples.
6. The mini split heating is efficient [energy usage] down to about 25F.
Below 25F, the efficiency drops as the outside temperatures drop. This is important for those in colder climates, particularly those with average to below average home insulation.
7. The Mitsubishi handheld remotes are old technology. The display is very small, not back lit [it’s especially a problem at night], and the icons are confusing [such as the hot and cold icons].
8. The temperature sensor is not in the remote, which is located at occupant living level, instead they are embedded in the indoor heating unit at the celling level [see above 2, 3, 4, and 5]. Very bad design.
9. The Mitsubishi indoor unit plastic enclosure is not UV protected. Sun
exposure over time can discolor (yellow) and damage the plastic. The units are ETL safety certified. As a professional in this field, the ETL mark is suspect.
10. If the outside units are mounted to the exterior walls, get ready for train like [vibration] noises during the heating season.
11. "Set and forget" does not seem to be as claimed. Last year, when I
used A/C on demand, I had significant solar generated energy in the "bank"
for the winter months. After Mitsubishi convinced me to "set and forget", I
have about zero solar energy in the bank for the same period this year.
12. [Diamond] contractors are not fluent with this technology. My
experience with a few contractors shows a minimal [at best] installation
capability. Get ready to have minimum service. Also, Mitsubishi seems to
lack fluency with this technology. The heat pumps seem to be a niche
technology here in US. Maybe not “prime time” ready for the US mass market.
I've spoken to folks that use the mini splits as supplementary heating and/or as A/C, and these folks have no "complaints", except for the remote controller.
And the good news:
Great A/C with zone control per indoor unit,
Better for the environment than oil or gas, especially if combined with solar.
If the dwelling is very well sealed and insulated [see 2, 3, 4, and 5], heating should be okay as well.
Probably a very good choice for supplementary heat.
Kazirode,
Thanks for your tips. Several GBA articles, inlcuding "Bruce Harley’s Minisplit Tips," advice readers that floor-mounted units do a better job of heating than units mounted high on the wall.
In that article, I wrote: "Harley also cited a 2015 study of ductless minisplits conducted by Robb Aldrich of Steven Winter Associates: 'The researchers found that if you direct the vanes [of an indoor unit] to a certain direction, you reduce the air flow and hurt the efficiency. They also found that in rooms with stratification, the warmest air in the room is sucked back into the unit. The efficiency would be higher if you could return cool air to the unit.' That’s why it makes sense, if a ductless minisplit is used primarily for heating, to choose a floor-mounted indoor unit."
Concerning vibration associated with wall-mounted outdoor units: Different homeowners have different reports on this issue, but in general, mounting the outdoor unit on a concrete slab -- ideally, on a steel stand that raises the unit above snow height -- is best, as you point out. This issue has also been covered in several GBA articles, including "Net-Zero Homes Show Signs of Convergent Evolution."
In that article, I quoted Paul Biebel, who said, “I’ve also learned that there is always some vibration from the outdoor units, so it’s important to choose the location carefully. ... When things are quiet, you can hear a vibration or hum.”
Hello Martin,
Thank you for your comments. The stratification issue seems to fit quite well with the thermal images I have. That can explain quite a bit about the high operating costs. I shared your observations with Mitsubishi.
Quite a while ago, the Mitsubishi Regional Service Manager set the vanes on all seven units. I forgot to include a comment about correctly setting the vanes. I had difficulty figuring out the optimum directionality adjustments.
Thank you.
A couple of misconceptinos:
"1. Electricity cost in MA is double of what it was about three years ago."
Not true in the majority of cases, possibly true for a very few municipal light utilities. Electricity price spikes at the end of 2014 related to the Polar Vortex event have abated somewhat in the past 2 years, but even those price peaks were only a ~30% rise over the prior years. Almost nobody was paying less than 15 cent/kwh for retail residential electricity 3 years ago (fully delivered, not just the energy portion) and nobody is paying 30 cents now, even on the grid-connected islands.
The energy portion of the power bill in MA is a competitive market- you're not stuck with the utility's "standard rate" in most cases. The state has a website for comparison shopping for power: http://www.energyswitchma.gov/#/
(On a 100% wind power contract I've been paying 1-2 cents less than this years' winter rates, about 0.3 cents more than the current summertime rates. )
"11. "Set and forget" does not seem to be as claimed. Last year, when I
used A/C on demand, I had significant solar generated energy in the "bank"
for the winter months. After Mitsubishi convinced me to "set and forget", I
have about zero solar energy in the bank for the same period this year."
While "set and forget" works better with modulating minisplit than with multi-splits, comparing this year's much higher cooling degree-days compare to last year isn't valid until it's normalized to kwh/CDD. Daily cooling degree-day weather history data from a local weatherstation on degreedays.net can help sort that out.
The other variables that would affect how much electricity you're "banking" are the cloud cover & temperature this year vs. last (output drops with higher temperature, even as AC power use is increasing). Another potentially huge factor is the lower efficiency due to being lower on refrigerant this year than last due to the slow leak.
Suffice to say, blaming it on the set-and-forget strategy is premature at best.
Hello Dana,
The electricity three years ago was 12c/kWh. Then it spiked to 24c/kWh. Now it's down around 22c/kWh.
As you point out, it is difficult to compare apples to apples about on demand usage vs. set and forget. But on average, not sure how much hotter this summer is compared to last year. I may run degree day comparison. However, I know that the system has consumed a lot of energy. When I install smart energy usage monitoring, I'll run a few experiments for comparison. At this time, I'm suspicious of the manufacturer claim. I wish they would provide clear substantiated evidence to support their claim.
Thank you.
Your experience varies quite a bit from the average ratepayer in MA. The state average 3 years ago was about 17 cents, but there were a few local municipal power companies (apparently yours) still in the sub 15 cent range:
https://ycharts.com/indicators/massachusetts_electric_utility_retail_price_yearly
Who is your electric utility? I can only think of two munis that were ever in the 12 cent range over the past half-dozen years, both of which were in Worcester County, and their rates were above that 3 years ago. The NStar/Eversource and Nat'l Grid customers (more than half the state) haven't seen all-in retail electricity pricing that low in more than a decade.
If the system is way oversized for the load a set & forget approach may be less efficient, but even at 1.5x oversizing it's usually going to work out better. The reason set-and-forget is usually more efficient is that the compressor then runs nearly continuously but at a lower, more efficient speed. Running the compressor at minimum speed in cooling mode the steady-state efficiency is about twice that of running at high speed. But the average efficiency won't be that high if it's cycling on/off, due to spin-up losses. The strategy presumes that it's sized small enough to have a very high duty cycle even at part-load.
Since you have two separate systems you might see if one system can be turned completely off to improve the duty cycle of the other one. Repeated spin-ups of the compressor(s) on zone calls is a real efficiency killer. Ideally you'd have at least one zone calling for refrigerant at any one time so that the compressor only steps up/down with zone calls rather than spinning up from zero every time. It's possible that it'll be cycling on zone calls even with just one of the systems running, but if one turns off a few of the heads and let the heaviest load zones cover most of the rest of the load the cycles will be longer, with fewer cold-starts from zero rpm.
Still, single head mini-splits with larger turn-down ratios and a modulating output do considerably better during lighter-load periods than multi-splits, since they can run nearly continuously for hours at a time rather than spinning up to serve one zone for 3-5 minutes, then repeat that 5-10 minutes later for a different zone, etc.