Are setbacks always less efficient with mini split heat pumps?
I recently had a four zone Mitsubishi system installed which consists of three wall/floor units and a ducted air handler. Over the weekend I also installed a Sense energy monitor.
I’ve read many articles and posts about how setbacks increase energy usage with a MSHP as it will cause the system to jump to a less efficient operating range when the setback ends.
Based on some experimentation today this doesn’t seem to universally be the case for mild setbacks on my system, at least under current conditions.
Basically, what I’m seeing is the energy usage for the house doesn’t seem to change at all if I bump the setpoint for a given zone down by 1-2° as long as that zone is already at/near the previous setpoint. I’ve not yet tried seeing what happens when I drop the temperature for multiple zones at the same time.
Anyone have any thoughts or experience to share? Could it be I’m just seeing an ideal alignment of circumstances (mild outdoor temperature relative to setpoint) that are leading to this behavior? I really wish there was a way to gain some better I sight into the behavior of these types of systems.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I was wondering about this myself. It seems to me that if you had a well insulated house with a good thermal mass, it could conceivably make sense to have nightly setbacks. This would allow the heat pump to spend more time operating when the ambient temperature puts it in a more efficient range. This would only be effective if the nighttime temperatures are quite cold, otherwise the change in efficiency would be negligible. It would probably also require accepting a wider indoor temperature swing.
I do exactly this and it's been backed up with my electrical monitoring data. In the cold winter months I have it scheduled to turn up the setpoint in the early afternoon when outiside temps are much warmer and it operates way more efficiently. The house acts like a giant heatsink and it is extremely effective.
Does the sense let you monitor individual circuits? If not I would connect the sense directly to the mini split circuit and monitor it for awhile. I have had a efergy hooked up to mine since Jan. Seeing the live electrical modulation has been amazing. You can reall fine tune on how to run it the most efficiently. Its amazing how little power these units use.
Here is a random snapshot of my monitoring.
> I’ve read many articles and posts about ....
Most "information" on the Internet is wildly over-generalized from limited data and then repeated over and over.
Ideally one would collect lots of data using both methods and normalize all of it to kWh/HDD.
Setbacks during Spring and Fall are a lot different than trying to recover a couple degrees when it's 0 degrees out, so I'd be curious to know how it reacts in the dead of winter.
Sometimes it's tough this time of year with solar gain in the afternoon and cool overnights. The 2 degree increments with my Fujitsu's sometimes are just a little too warm or cool so I find myself scheduling a boost in the morning and let it drop rest of the day/overnight.
>"Setbacks during Spring and Fall are a lot different than trying to recover a couple degrees when it's 0 degrees out, so I'd be curious to know how it reacts in the dead of winter."
When it's in single digits F around 0F the COP at max speed isn't very different from the COP at minimum speed, so setbacks won't have a huge negative impact on the extreme temperature days.
During the shoulder seasons some mini-split compressors running at max speed can have HALF the efficiency (or less) of when running at low to mid range. See Figure 5 for third party bench tested efficiency at different modulation levels of an older 1-ton Fujitsu across temperature:
https://www.nrel.gov/docs/fy11osti/52175.pdf
With it's 35F outside, at min-speed it tests just shy of COP 5, and at max speed it's about COP 2.3. When it's 17F outside the efficiency clusters in narrower (and lower) range. When it's 5F or lower it's going to be 1.5-2, no matter what speed.
While newer generation Fujistus are somewhat higher efficiency overall, the relationship between modulation level and efficiency are similar, but not as dramatic. For instance, at +5F the AOU/ ASU12RLS3Y has a COP of 2.77 at minimum speed, 2.1 at maximum speed- about a 24% lower efficiency. At +47F at min-speed it's running 4.54 , at max it's 3.34, about a 25% hit in efficiency.
https://ashp.neep.org/#!/product/25334
That's still an efficiency hit, just not as dramatic as with the older 12RLS2.
There are yet other products from other vendors with abyssmal low-speed efficiency at +47F, but still pretty reasonable mid -range efficiency, with another dramatic drop at high speed. Sizing those correctly and letting them modulate will almost always result in lower energy use.
Still thinking a lot about this topic. Is there a way to determine (or lookup) the COP curve for the particular combination of equipment I have to better understand what efficiency can be expected at various loads/conditions?
I have a four zone system consisting of the following:
MXZ-8C48NAHZ
2x MSZ-FH09NA
MFZ-KJ09NA
SVZ-KP24NA
I know we're not talking about huge saving here with the small setbacks I described in the original post, but I'd still like to figure out what the best approach is definitively.
Honestly, I think part of the problem is that it feels wasteful to leave my office set to cool at 70ºF from 5PM until 8AM the next morning rather than having it set back to something like 76º while it's unoccupied.
It seems like much of the research that has been done so far has focused on single zone systems used for heating in a cold climate. I'm really curious to know just how the numbers work out for a zoned system like mine being used for both heating and cooling in Central PA.
Start recording daily energy usage data with and without setback (alternate days). Also record average temperature for the day (which will normalize out some of the errors). Post enough data and I'll give you a definitive answer.
I already have the power data being logged in Sense, as well as per-room temperatures, outdoor temperature, solar radiation, etc being logged via Ambient Weather.
I'm not logging thermostat setpoints yet, so let me see if I can quickly setup some automation for alternating the setback through the Honeywell TCC API every other day on a consistent schedule. For the purposes of this test I think I will limit the setback to a single room (the office I described previously) to make the data easier to analyze.
If the setback is only for a single room (eg, 3/4 ton on a 4 ton compressor), catch-up will rarely cause the compressor to reach high modulation levels. I expect there will be a savings (although it may be lost in the noise).
Some data showing that a fixed thermostat setting isn't the optimal strategy for energy use. Not clear how widely applicable it is - nothing beats data collection on your own system.
https://aceee.org/files/proceedings/2014/data/papers/1-768.pdf
The Sacramento case has nothing to do with operating efficiency of modulating mini-splits, and is only applicable in locations with time of use rates AND a climate amenable to nighttime ventilation strategies for cooling. While the energy use savings was into low double digits in the benchmark, it's not particularly relevant to this discussion. The energy COST savings of pre-cooling was much higher, due to the rate structures. It's not at all clear if a right-sized modulating system would have underperformed the SEER 13 and SEER 15 systems monitored on either energy use or energy cost, but it's hard to believe that in that location with mostly negative latent loads a right sized SEER 20+ ductless system wouldn't have beaten it (with margin) on both energy use & cost, whether a pre-cooling strategy was used or not.
The Phoenix case only showed 3% energy savings, and again this was not modulating equipment, which would usually have more than 3% savings by operating in a modulating load rather than pre-cooling.
The Tucson high performance high thermal mass house case with the air source hydronic chiller used for both fan coils and slab cooling is even less relevant. The monitored results of 30% savings by pre-cooling on extreme peak days was interesting, but the average indoor temp was higher on those days too- it's not apples-to-apples, or even apples to pears. Pre-cooling a slab would be more akin to using the Ice Cub thermal storage cooling system than a standard split system or a modulating mini-split, with or without set backs.
While it's an interesting study, none of those cases is really applicable to whether a modulating ductless is better off with a fixed temp vs. setback strategy.