Minisplit decisions
Sorry, it’s going to be another one of those mini split questions.
Background info:
2 story house in zone 6, SW Ontario, heating demand 12kBTU at 0F.
Currently have in floor hydronic, fed by electric DHW tank. Was led to believe by the crazy builder that the heating demand is so low that it doesn’t make sense to have a heat pump. This was a bit of a dubious claim to begin with, but coupled with the fact that actual heating energy has been at least double what was predicted, now it’s time to look at heating alternatives.
Part of me wants to salvage the in floor heating, adding either a Chiltrix or SanCo2. The Chiltrix just looks too expensive (conservatively C$13,000-16,000 for the outdoor unit, control, tank, two fan coils for cooling). For way less than half of that I can get a mini split. The SanCo2 is a little more reasonable, but still significantly more than a mini split and it provides no cooling. There’s also a rebate for air-air heat pumps, not for air-water.
Based on energy modelling I did at the tail end of the winter, it seems like 90% of the heat going into the house is going in on the main floor. So for heating purposes, it would probably make the most sense to just have a single 12kBTU zone on the main floor. The main floor is much more suited to single point heating as well, being an open concept. However, if I’m going to have a heat pump, I would want to have some cooling in the summer as well. Would cooling just the main floor help the upstairs much, or would it result in a huge temperature difference between the floors? Not really needing a lot of cooling, just dropping the temp 4 degrees F during the warmest 3 weeks of the year would be fine.
If I want a 2 zone system, is it acceptable to have the capacity of the 2 interior heads to exceed the outdoor unit, with the intention being that one be shut off in the heating season, and the demand will be lower in the cooling season? Like if I got a 12kBTU outdoor unit, a 12kBTU indoor unit for the main floor and a 6kBTU unit for the second floor?
If I went with just a single zone and found out it was a mistake, is it possible to add a second zone later on?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Depending on the model, you can have up to about 130% sizing diversity between inside and outdoor units on multi head systems. But your total capacity at any time will be limited to what the OD unit can deliver. If the outdoor unit has left over capacity from the diversity of not running all indoor units, the remaining units will not suffer from reduced capacity at low ambient conditions. (meaning full capacity (12kbtu) even at -25C downstairs).
I'm in Southern Ontario, and found it more cost effective to get multiple single heads. It's also more efficient and you get a big advantage in wide modulation ranges if you choose your single head models carefully. Multi head systems don't usually modulate down, so a 12kbtu head will always try and deliver 12kbtu and will cycle more often than a 1:1 modulating 12kbtu head that can go down to 2500btu.
I can see going with multi heads once you have 4 or more inside heads, since it could be an eye sore to have many outdoor units on the same house.
A pair of separate single zone Mitsubishi MXZ/MUZ- FH09NA mini splits is a far superior solution to a 1.5 ton multi-split with two zones and (trying to) turn one zone off. The minimum modulation at +47F of any 2- zone multi is on the order of 6800-7200 BTU- more than half your design load, and will spend most of it's time cycling rather than modulating. A single FH09 delivers about 10,000BTU/hr @ 0F, and is good for 12,000 BTU/hr of cooling, yet it can throttle back to 1,600 BTU/hr @ +47F in heating mode.
https://nonul.mylinkdrive.com/files/MSZ-FH09NA_MUZ-FH09NA_Submittal.pdf
The 3/4 ton Fujitsu cold climate mini-splits have more capacity, but also a much higher minimum modulation of 3100 BTU/hr- a pair of them would have a minimum modulation nearly as high as a 1.5 ton 2 zone multi-split.
During the summer you can probably turn the downstairs unit completely off (throw the breaker!), and during the heating shoulder seasons you may be able to turn the upper floor unit completely off. When a zone of a multi-split is nominally "off", there is still refrigerant flowing through the coils whenever any other zone is running, it is still heating/cooling the "off" zone. In cases of extreme oversizing that can even lead to overheating or overcooling comfort issues.
The installed cost of a pair of single zone mini-splits is usually cheaper than a multi-split of similar capacity.
Regarding the Sanden: It's not a trivial engineering task to get a CO2 compressor to run a radiant floor efficiently. Unlike R410A and other common refrigerants, CO2 refrigeration systems do not undergo a phase change from gas-to-liquid the way other heat pumps do- it needs a higher temperature difference than you would get in a radiant floor system to run in it's more efficient operating zone. It's one thing to be taking 5-10C water up to 55-60C, quite another to take 30-35C return water on a low temp hydronic heating system up to 40-45C. Sanden is currently selling combi-systems, but without a large domestic hot water load their efficiency isn't going to be as good as an R410A reversible chiller.
Thanks, this makes it pretty clear that my two options are:
1) install one unit on the main floor, try it for a year and see how the upstairs behaves
2)install one unit upstairs, one downstairs at the same time
The only reason for option 2 would be to save the extra cost of the second site visit, as the total amount of work would be the same.
Dana,
Based on the heating load, is this the model you'd say was the most appropriate?
MUZ-FH12NAH
I'm assuming I want the hyper heat option.
In my area a pair of FH09s installed at the same time would come in under USD$7000 (~CDN$9.5K) in competitive bidding, maybe even a hair under USD$6K if done at a time of year when the installers are more hungry for work.
A single Fujitsu -9RLS3H has about 14K of capacity at 0F (more than your estimated load, and more than the FH12NAH, and with a lower minimum modulation than the FH12NAH) and would usually be USD$3-3.5K in competitive bidding in my area, but it's minimum modulation is 3100 BTU/hr @ +47F. If trying to do it all with a SINGLE ductless this would be a better choice than an FH09NA.
If you later decide to add another mini-split upstairs, an FH06 or FH09 would be a better choice than a second -9RLS3H.
All of the "FH" series single zone units are hyper-heating, no matter what the suffix letters happen to be.
So you lean toward the Fujitsu? I've never seen one of those in my area, and a couple of places I have been in contact with handle only Mitsubishi.
I just did a contractor search for Fujitsu, and the closest one is a hundred miles away.
*edit: I see the first models you mention are Mitsubishi, just with the MUZ left off. I saw model numbers starting with F and thought Fujitsu.
*2nd edit: I found a Fujitsu contractor just by randomly calling hvac places. Don't rely on Fujitsu's contractor locator!
For delivered low-temp capacity you get a lot more out of a 3/4 ton Fujitsu than a 3/4 ton Mitsubishi (and slightly more than a 1-ton Mitsubishi.) The minimum modulated output of the 1-ton Mitsubishi is 3700 BTU/hr @ 47F, which is slightly higher than the higher low-temp capacity Fujitsu.
HVAC installation companies usually handle only Fujitsu or only Mitsubishi- I've never found one that did both.
I don't particularly favor one VENDOR over the other. Among the major vendors it's MODEL that fits the application best is what matters. But local support matters too. If the nearest Fujitsu installer is 100 miles away, 1 ton Mitsubishi on the first floor (to cover the design load) isn't a disaster.
Dana, where would I find the extended performance specs for either the Fujitsu or Mitsubishi?
The Mitsubishi FH series AHRI submittal sheets specify capacity at 70F (21C) indoors and +5F/-15C, -4F/-20C, and -13F/-25C, which for a home heating application should be "good enough".
See the capacities at different temps listed on the top half of the second page(s):
http://meus1.mylinkdrive.com/files/MSZ-FH06NA_MUZ-FH06NA_ProductDataSheet.pdf
https://nonul.mylinkdrive.com/files/MSZ-FH09NA_MUZ-FH09NA_Submittal.pdf
http://meus1.mylinkdrive.com/files/MSZ-FH12NA~MUZ-FH12NA_Submittal.pdf
http://meus1.mylinkdrive.com/files/MSZ-FH15NA_MUZ-FH15NA_Submittal.pdf
For the Fujitsu xxRLS3H series I find it easier to refer to the NEEP spreadsheet, which lists the capacity at +5F, and -15F, but also for the xxRLS3 (no-H) at -5F. The only difference between the -H version is the pan heater- the capacities are identical.
NEEP is currently re-compiling the data, and the spreadsheet aren't currently downloadable. But according to the last one released the 9RLS3H is good for :
15K @ +5F
3.5K @ -5F
11K @ -15F
The 12RLS3H is good for
16.5K @ +5F
15K @ -5F
11.5K @ -15F.
The 15RLS3H is good for
20.5K @ +5K
18.5K @ -5F
16K @ -15F
At +47F all xxRLS3H models can modulate down to 3100 BTU/hr
I agreed to a proposal for the 9k Fujitsu for $5000, less $1500 rebate (which expires April 1, so a bit under the gun to make a decision). A couple of days later, he called to inform me that the rebate would only be $1000, because despite the Fujitsu operating down to -26C, Fujitsu failed to have it registered with NEEP as a cold climate device. Any way I can confirm or dispute this? Seems pretty weird to me. I'm already giving up the 2-year labour warranty because I'm 10 minutes outside the dealer's service area, the reduced rebate is just another kick in the pants. Wishing I'd got a quote on a Mitsubishi from someone else at the same time, but certainly too late to try to get that done before end of Monday.
*edit: Using the wayback machine, I was able to read the exact requirements of the program:
"B. meets the requirements in NEEP’s Cold Climate Air-Source Heat Pump Specification (NEEP, 2017 or more recent; ask your Participating HVAC Contractor for a copy of this specification); and also meet the following two additional requirements: (a) COP at -15° C (5° F) greater than 2.0 at maximum capacity operation; and (b) maintain a maximum capacity at -15° C (5° F) that is greater than or equal to 55% of maximum capacity at 8.3° C (47° F)"
So while the guy wasn't quite accurate in what he told me (it is on the NEEP list), it does fail the first additional requirement, the minimum COP of 2 at 5F. It's 1.95. This is for the model AOU9RLS3H. If I went with the AOU9RLS3, without the H, the COP is 2.09 and would qualify. Would that unit still be appropriate for zone 6? We see below -20C, but not very often and we have plenty of supplemental heat to cover.
Another option would be to go up to the 12RLS3H. Has the same minimum output as the 9, but for whatever reason the COP at 5F is higher than the 9. This option is $450 more than dropping down to the 9RLS3 (no H), but $300 less than the 9RLS3H.
Also noticed that the Mitsubishi has better COP at 5F than either of the Fujitsu models (2.33), so now I'm really sad I didn't get a quote on one of those.
I was kind of freaking on the $5K quote until I realized it was Loonies, not greenbacks. That's about USD$3.75K, a bit higher than I would expect for a 9RLS3H in competitive bidding, but not crazy-high the way USD$5K would be.
Unless 5F is your 75th precentile temperature bin or something the COP at +5F isn't very relevant (but it's capacity at +5F is important if it gets that cold in your neighborhood, which it does.) The COP at your average mid-winter temperature is a far more important data point.
Eyeballing the temperature graph for London Ontario for January/February it looks like the binned hourly mean temp for the coldest 6 weeks of the year is in the low to mid 20s F. That's the temperature at which you start caring about the COP performance. Since the COP at +17F is listed, that's a good starting point.
It's too bad whoever decided on the rebate criteria didn't talk to you first. I figured out that the RSL3 line only differs in lacking the pan heater. So it gets that COP boost by not having any defrost action. From what you've said before I gather that the pan heater is pretty important for my climate?
Pan heaters aren't essential if you're the type to occasionally check that the drain isn't plugged with ice during periods of extended cold/very cold temperature operation. Damaging levels of ice build up are rare in zone 6, but can be expensive when it happens.
The RLS3 definitely has periodic "defrost action" on the coil, and it's the defrost water dripping on to the pan than needs to drain away before it re-freezes. The pan is designed to drain, but when it's VERY cold out it can re-freeze before it drains. A pan heater that comes on only during defrost cycles during very cold weather prevents the refreeze from happening.
What would constitute extended cold / extreme cold period? If it's once or twice a winter, that's not a big deal. What do you do when you see the drain plugged, go at it with a heat gun?
When it's near freezing but still below the drip water is warm enough to flow. I'm not sure there's a single temperature or duration that works as a general rule of thumb, but if it hasn't been above +10F for a week it would be worth checking.
A heat gun/hair drier approach would be enough if caught early. Heat tape use for freeze protection on plumbing wrapped around the pan could also work unless it's -10F or something. Of course buying a mini-split with the controls and heater already built in works. I don't know for sure (haven't checked into it), but I'd be surprised if an RLS3 didn't have the control features on board, and just needed the pan heater itself to be added.
I suspect you're right about the controls, though they might be firmware locked out. The question is whether Fujitsu offers a pan heater upgrade kit.
It appears that the heater is integrated on the Fujitsu models, so it's not something you can buy as an add-on. That would have been ideal.
Between the 9RLS3 and the 12RLS3H, are they going to be pretty close in terms of overall efficiency? The HSPF is almost identical. The EER is quite a bit different, but the cooling is going to be a much smaller slice of the usage pie than heating.
I would like the peace of mind of the pan heater, and would probably cough up the extra up front to get it, but this blurb from their marketing spiel makes me hesitate:
"When temperatures reach 36ºF, a built-in heater warms the base of the outdoor unit so that condensation doesn't freeze. Without a heater, freezing condensate can cause noise, damage to the fan blade, condenser, and system performance."
They make it sound like the heater is running 100% of the time the ambient temperature is below 36F. This seems to be supported by the specs in the NEEP product listing. The difference in input power for every capacity and every temperature below freezing for the RLS3H is always exactly 150W higher than the RLS3. That doesn't seem like a very "smart" system. You can see the effects on the COP at the min. capacity at 17F, on the unheated model it holds almost steady compared to the COP at rated capacity, whereas the heated model drops by 36%.
While I'm willing to pay more for convenience of not having to check the pan occasionally, I don't like the idea of a 150W heater running 24/7 outside when it's below 36F.
I thought that in Fujitsu's original cold climate RLS2H series the heater ran continuously when it was sufficiently cold outside ( which was a real hit in HSPF efficiency compared to the non -H RLS2), but that the RLS3H runs the pan heater during defrost cycles and perhaps for awhile after the defrost cycle ends. That said, I haven't had a chance to observe one up close in action during a defrost cycle.
If the effect of the heater running constantly showed up in the HSPF rating, it's a good indication that is not the case with the RLS3H series, as the HSPF only drops by 0.2 from the RSL3. Maybe the input power and COP they have to report to NEEP is worst case, i.e. in defrost mode?
>"Maybe the input power and COP they have to report to NEEP is worst case, i.e. in defrost mode?"
That would be my guess too.
The contractor dropped the ball, and failed to get the process started before the rebate program was terminated. They thought they were going to be able to just back date it, or that there would be a grace period, but that was not the case. Very frustrating, as I met with the guy on Tuesday, and he knew the deadline and had four days to get it done. That mistake cost me at least $1000, perhaps $1500. So I'm not particularly inclined to give them any business, which means Fujitsu has fallen out of my favour a bit (unless I can find another contractor to deal with).
Are there any other brands to look at for quotes, or are Mitsubishi and Fujitsu in a class on their own? I now have all the time in the world to make a decision.
>"Are there any other brands to look at for quotes, or are Mitsubishi and Fujitsu in a class on their own?"
LG's LAN090HYV1 /LAU090HYV1 delivers 11,220 BTU/hr @ +5F/-15C, 7920 BTU/hr @ -13F/-25C, so one per floor would handle your loads.
https://s3.amazonaws.com/s3.supplyhouse.com/product_files/LG-LA090HYV1-Submittal-Sheet.pdf
https://files.lghvac.com/resources//EM_SZ_ArtCoolPremier_ALL_01_17.pdf.pdf
I think that model is usually shipped with the pan heater option pre-installed, but verify that with the contractors.
This series modulates down to a very low 1023 BTU/hr @ +47F, but it's efficiency at that very-low output is pretty lousy (not that it's going to matter much). It's HSPF tests at 12.0, and it's efficiency at your mid-winter average temp is a COP of about 3 at roughly half-speed, which is where it would be running.
Support in my area for LG isn't nearly as well developed as Mitsubishi (with a regional training center less than a half-hour drive from my house) or Fujitsu. A handful of years ago field reports locally indicated possible quality control issues with them, but I haven't heard of any "lemon" units that failed early or never worked right more recently than that. It's worth investigating where the nearest training center is and how many LG trained installers are working in your area. Most problems with mini-splits can be avoided with competent installers who are familiar with the product.
I seem to see about five Mitsubishis for every Fujitsu, and five Fujitsus for every LG. The newer colder temp versions are fairly recent though, so I may start seeing more of them.