I have a Viessmann gas boiler for baseboard heating, there is also a Fijitsu minisplit (7k+12k) system in the house. I use the minisplit when the outside temperature is above 40F. But when it gets colder I turn on the boiler and also I stop using the 12k unit in the living room so it won't affect the boiler thermostat to cause any potential pipe freeze. The other 7k unit is in my bedroom, and for many nights I set the boiler thermostat to 65F and use the 7k unit for additional heat in the bedroom (68-70F). My question is at what outside temperature is the minisplit less efficient than the boiler? So far this year the coldest night was 5F, and I wonder if it would cost more to run the minisplit. Electric rate here is 20 cents per kwh. Viessmann can achieve 90%+ efficiency most of the time. Thx!

The mini split is ALWAYS more efficient than the mod-con. Even at -10F it'll have a COP higher than 1 (=100% efficiency), whereas the boiler's raw combustion efficiency never reaches 100%, let alone accounting for the pumping & control power used to get that heat into the rooms.

What you're really only talking about cost, not efficiency, but you've only quoted the electricity rate. The gas price is...??? (Freeeee...!?!)

The crossover point will differ if it's $0.40/therm delivered compared to $1.40/therm. (And if your gas is free it's always going to be cheaper. :-) )

Then, mini-split efficiency depends both on temperature and oversizing factor, there is no simple correlation that is temperature only. The modulation level/speed it's operating affects efficiency- a LOT.

If the mini-split is running constantly at low modulation at 40F, it's efficiency will be about 400%. If it's going full blast at 40F it'll be between 200-250%. At +5F the difference is more like 200% at minimum speed, 150% at maximum speed (though some will still do a bit better than 150% at +5F.)

If the thing is cycling on/off a lot at 40F rather than running constantly at low speed it's efficiency will be somewhere between 250% and 350%. The more it's cycling, the lower the net efficiency.

The minimum compressor output at 47F for most Fujitsu multi-splits is about 6000 BTU/hr, and if your heat load at 47F is much less than that it'll be less efficient at 47F when it's cycling than at 35F when it's running constantly.

Oops, I forgot to mention that I am paying $2 per a gallon of propane. =) And I meant cost, not efficiency, I had the title edited.

I notice with the outside temperature gets below freezing, the unit goes into defrost mode from time to time, will this add up the cost by much?

The defrost modes are pretty efficient, but it does cut into efficiency a bit. It's usually defrosting the most when it's just above freezing and humid outside. As the air gets much colder than freezing the moisture content of even 90% RH outdoor air drops pretty low.

A gallon of propane contains 91,600 BTU/gallon of source energy, and at 90% efficiency delivers 82,440 BTU /gallon into the heating system. Normalizing to million BTUs-delivered (MMBTU), that's 12.13 gallons/MMBTU.

At $2/gallon that's about $25 / MMBTU, ( a bit more than $2 x 12.13= 24.16 to account for the pumping and control power being used.)

A kwh of electricity is worth 3412 BTU/kwh if used at a COP of 1 (say, an electric baseboard.) Normalizing to MMBTU, that's 293 kwh/MMBTU. At 20 cents that's $58.62/MMBTU, assuming a COP of 1.

At a COP of 2 it's $58.62 / 2 = $29.31 / MMBTU , 15-20% more expensive than condensing propane

At a COP of 2.5 it's $58.62 / 2.5= $23.45 /MMBTU, about the same or slightly less than condensing propane.

At a COP of 3 it's $58.62 / 3 = $19.54 / MMBTU , 20% cheaper than condensing propane

At a COP of 4 it's $58.62 / 4 = $14.66 /MMBTU quite a bit cheaper than condensing propane.

Without knowing how it's modulating or what the heat loads are, but it's highly likely that it's cheaper to run the mini-splits than burn propane until the outdoor temps are at least under +25F, maybe even until it's under +15F, depending on the sizing factor relative to the loads at those temps.

Most multi-splits will deliver a COP of about 2 when operating at their rated modulation when it's +17F (an HSPF test temperature), but if it's still operating in the mid-speed range at that outdoor temp could be doing ~2.5 or slightly better. It it's maxed out at +17F it'll be doing slightly less than 2.

This is your compressor unit?:

http://resource.gemaire.com/is/content/Watscocom/Gemaire/article_1365692189727_en_subs.pdf?fmt=pdf

Note: It's minimum output at 47F is 6800 BTU/hr.

It helps to know what your heat load is at both +17F, and at +47F to come up with reasonable estimates of modulation levels across temperature, and where begins it's on/off cycling. At 40F it'll be no worse than condensing propane even if it's cycling, but it's better if it doesn't begin cycling the compressor on/off until a somewhat warmer temp. The individual room heads may begin to cycle on-off even with the compressor running constantly, but as long as the heat load is 6800 BTU/hr or more a bit of cycling on one head or the other won't hurt efficiency very much.

No, I have a AOU24RLXFZ. It's actually a 3 zone system but the third zone is never used in winter. Heat loss is about 4000-4500 BTU/hr at 17F and 2000 BTU/hr at 47F.

The min-modulated output @ +47F of the 24RLXFZ is also 6800 BTU/hr. At +17F it probably can modulate down to 4500-5K which would roughly match your whole house load. The only question is whether it's actually modulating at that temp which you can probably figure out. Even if the interior heads are cycling on/off, if at least one is going it's modulating- if both are off, it's not.

http://bangorwinsupply.com/wp-content/uploads/2014/02/AOU24RLXFZ-Submittal.pdf

With the minimum modulation being oversized for the average seasonal load your average COP is probably going to be in the mid 2s, but it also means that the efficiency at +17F is also going to be in the mid-2s, or even high 2s if both heads running fairly continuously at that outdoor temp.

So, it's probably going to be comparable cost to $2 propane, or slightly less.

This is a fairly common complication for getting the efficiency out of multi-splits on high-R houses- not enough load. The Mitsubishi FH09NA single head min-split will modulate down to 1700 BTU/hr @ +47F, but can still deliver 10,900 BTU/hr @ +5F, which is a better fit for low-load houses due to the wide modulation range. But putting multiple FH09 heads on a multi-split doesn't help, since the Mitsubishi multi-splits have the same high minimum problem at the compressor end. I takes multiple separate units to gain the advantage.

The min-modulated output @ +47F of the 24RLXFZ is also 6800 BTU/hr. At +17F it probably can modulate down to 4500-5K which would roughly match your whole house load. The only question is whether it's actually modulating at that temp which you can probably figure out. Even if the interior heads are cycling on/off, if at least one is going it's modulating- if both are off, it's not.

http://bangorwinsupply.com/wp-content/uploads/2014/02/AOU24RLXFZ-Submittal.pdf

With the minimum modulation being oversized for the average seasonal load your average COP is probably going to be in the mid 2s, but it also means that the efficiency at +17F is also going to be in the mid-2s, or even high 2s if both heads running fairly continuously at that outdoor temp.

So, it's probably going to be comparable cost to $2 propane, or slightly less.

This is a fairly common complication for getting the efficiency out of multi-splits on high-R houses- not enough load. The Mitsubishi FH09NA single head min-split will modulate down to 1700 BTU/hr @ +47F, but can still deliver 10,900 BTU/hr @ +5F, which is a better fit for low-load houses due to the wide modulation range. But putting multiple FH09 heads on a multi-split doesn't help, since the Mitsubishi multi-splits have the same high minimum problem at the compressor end. I takes multiple separate units to gain the advantage.

Same problem I'm seeing with multi-splits. For a home with low heating/cooling loads the outdoor compressor cant modulate low enough to match zone loads for long run times.

To install smaller multiple mini-split units outside such as the MUZFH can better match low loads, but is overkill to achieve seasonal BTU's required for each zone. Using multiple outdoor heat pumps will kill your ROI because of the additional equipment cost, phantom loads and long term maintenance requirements.

We can only hope that sooner rather than later, multi split manufactures will give designers what they need, "Ultra Low Modulation".

You can accomplish that with a Chiltrix air to water heat pump and a buffer tank. Use as many heads as you want.

Posted By jonr on 10 Jan 2016 10:16 PM
You can accomplish that with a Chiltrix air to water heat pump and a buffer tank. Use as many heads as you want.

Chiltrix claims on performance seem to be a bit unbelievable. I would have to see some third party testing. Assuming they are still in business in 10 years, can you get replacement parts. Warranty, 3 year compressor and 1 year everything else....That Sucks! I know there are some VRF manufactures starting to market and deploy new hydronic systems in multi story building and the reasoning behind hydronic is to eliminate harmful refrigerant leaks and eliminate refrigerant leak detection monitoring systems.

We can only hope that the larger manufacturers will offer small home sized A->W heat pumps in the US. Add a very large storage tank and they can also be used some solve some PV solar issues.

Thanks for all the suggestions. I do have one more question on maintenance, how many years can the system be running without recharging refrigerant under normal conditions?

Under normal conditions there are no refrigerant leaks, and the heat pumps should go forever without re-charging. They develop other problems though.

Cleaning the finned coils of air-source heat pumps is often necessary on a periodic basis. How often they need cleaning is a function of how much airborne crud you have in the local environment, and how much air you're moving through the coils.

Environmental degradation of the coils takes place over time. In seaside salt air environments significant degradation can happen in just a handful of years. As a general rule of thumb, in more benign environments you should be able to get 15-20 years out of it.

Some variable refrigerant valves on variable speed air source heat pumps will fail, though the track record is pretty good and getting better year by year. That's usually a reasonable repair charge when it fails, not a total unit replacement.

Mini-splits where the interior heads are NEVER cleaned may eventually be sending liquid refrigerant back to the compressor and destroy it, but again, this is more likely in an underinsulated house owned by a feline enthusiast with a couple dozen shedding cats that need to stay cool than in a tight high-R house occupied by a few humans & fish.