Need sizing advice on a minisplit unit
I installed a Mitsubishi FH12NA in summer of 2014 when they first came out. I live in inland San Diego, and it can get hot, like 10-15 degrees hotter than the coast.
My condo is a studio and is only 420 sq ft. I’m on the upper unit of 2 stories. I face East and South. I have a skylight and dual-pane windows. Building is 1978 construction.
My Mitsubishi is weak. I set the temp and it will barely do anything. I have to set the temp to 61 and manually put the fan on high for it to do anything. Even “powerful” mode is weak.
I’ve had my AC contractor come out and he cleaned the outside unit (I clean the inside filters) and it helped a tiny bit. He recommended I move up to a 18k unit. The 18k head units are enormous. I really don’t want it that overwhelming for such a small space.
I have been looking at the Fujitsu 15RLS3Y, a new model with wifi built in, but has the same specs as the regular 15RLS3. It actually goes up to 18,400 btu cooling. From what I’ve read on other threads they size it as a 15k unit (14,500 nominal) to get the efficiency ratings higher, but it’s actually a 18k unit in cooling mode.
The heat doesn’t matter much, I might use it for 2 months a year and only for an hour each day. I also do some hairstyling on the side from home once or twice a week and have many lights on and a hair dryer.
Do you guys think the 15RLS3Y would cut it? I’m kind of turned off by Mitsubishi now, but maybe it has nothing to do with them. Their equivalent unit the FH15NA goes to 19,000 btu/h cooling (their minimum is higher at 6450 btu/h), but the SEER is a little lower at 22, and the airflow rate is quite a bit lower, and no built in wifi. The Fujitsu is 25.3 SEER and has higher CFM airflow with the wifi.
What do you think? Thanks.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
The first step to these minisplit sizing questions is always the same: you need to do a load calculation. If you haven't done a load calculation yet, we can't provide advice.
Here are some links to get you started:
Saving Energy With Manual J and Manual D
How to Perform a Heat-Loss Calculation — Part 1
How to Perform a Heat-Loss Calculation — Part 2
Calculating Cooling Loads
Unless you have huge west-facing windows or a very leaky thermal envelope, an appliance rated at one ton of cooling capacity (12,000 Btu/h) should serve 1,000 square feet or more.
-- Martin Holladay
A medium to large sized skylight can drive the peak cooling loads stratospheric, especially if it's cheap clear glass double-panes on a south or west facing pitch.
Circa 1978 double-panes are usually clear-glass with a very high solar heat gain coefficient, which could be another major driver, especially if the south facing windows aren't shaded from the exterior.
If you can figure out the major source of gain, it may be more cost effective to keep the FH12 swap out the skylight for a low-gain version or install some heat rejecting low SHGC windows (if that's allowed under the condo rules).
The FH12 should be able to deliver over 13,000 BTU/hr cooling when it's 75F indoors, 90F outdoors. While there is some amount of temperature offset between the room temp and the temperature sensed at the head, the fact that you have to run the setpoint down to 61F for it to do anything makes me think there's something wrong with the unit, or the installation. An improper refrigerant charge level would be one strong suspect.
Or, it could be operation mode: If you're leaving it off all day then expecting it to pull the temperature down from a stagnated 85F or higher down to 75F in a reasonable amount of time, forget about it. The "right" thing to do with modulating equipment is a "set and forget" approach- let it run all day. At it's minimum speed it's efficiency is 2-3x what it is at high speed, and you use less power just letting it modulate with load than trying to play catch-up for hours in a heat-saturated house. The typical 3-4x oversized split AC can usually catch up in a reasonable amount of time when used that way, but a right-sized unit (mini-split or other) won't. Oversizing the AC so that you can play catch-up is also about the worst thing you can do for the grid, since it piles on with 3-4x the amount of grid load of a right-sized unit, driving the grid-load peaks skyward (and generator ramp rate requirments) skyward too.
thanks for the responses. My AC guy checked to make sure it was working properly and didn't have a leak and it was fine. it works, just not very well. I did the load calculation on CoolCalc that was given by Martin. It said I would need about 16,000btu of cooling for my load, which confirmed my suspicions that my unit isn't powerful enough for the job at hand.
The skylight isn't huge, maybe 2x3, but it is clear and dual pane Velux skylight put in around 2007. My other windows face East - one slider, and one smaller taller window, and both are Low E3 put in around 2009. The front door faces South. I'm on the 2nd floor of 2 floors so it is always warm.
I also have a full kitchen, big tv/stereo/cable box etc creating heat, and 2 dogs. I'm thinking that Fujitsu will do the job since it can do between 3000-18,400 btu, and is barely less efficient than my current Mitsubishi (25.3 for Fujitsu 15k vs 26.1 for Mitsubishi 12k).
Which direction does the skylight face? What is the pitch of the roof? What is the SHGC of that unit?
Even 6 square feet of clear glass double pane angled directly at the summertime sun adds a huge solar gain.
What does CoolCalc tell you the cooling load is if you eliminate the skylight from the input?
The skylight added about 1,300 to the load - it's a flat roof and the sun comes in directly during the summer. It is horizontal facing west to east. It's nice but adds heat. The interior lights, all the kitchen appliances, and my stereo/tv/computer also added a lot. And i do some hairstyling once in awhile too with a hot blow dryer with lights on and my 12k unit is no match whatsoever. My revised load calc is around 16,900 sensible, 840 latent.
Do you think the 15RLS3 would be powerful enough? Its nominal cooling is 14,500 but max is 18,400. I was also looking at the mid level Mitsubishi MSZ-GL18NA. That one's nominal is 18,000 with max cooling at 22,000 and the airflow is higher - up to 646 cfm. I also found out the head unit dimensions are about the same as my current unit. Seer is lower at 20.5 though. Is it overkill? Or is it better to be a little oversized to have the unit running at a lower load since these are all variable load compressors? I have no idea.
They are all variable output, but the range and turn down ratios are not equal, and it matters.
The minimum modulated cooling at test condition on the MSZ-GL18NA is 5800 BTU/hr:
http://meus1.mylinkdrive.com/files/MSZ-GL18NA-U1_MUZ-GL18NA-U1_ProductDataSheet.pdf
The minimum modulated cooling on the 15RLS3 is a bit more than half that, at 3100 BTU/hr:
http://www.admorhvac.com/admor-files/submittals/ASU15RLS3.pdf
The "nominal" or "rated" numbers in the submittal sheets are the modulation levels at which efficiency was tested.
The maximum output number is important for judging whether it has sufficient capacity for the design load.
The minimum output is important for gauging how much it cycles rater than modulates which affects both efficiency and comfort. The more modulation range at the low end, the more comfortable you'll be overall.
The 15RLS3 has some margin over your 1% design load, and will deliver more modulated output comfort than the GL18NA, and would be a better fit to the load.
Thank you, Dana! Always so informative.
One final thought I had was, do you think there are any other units I've overlooked that you think I should consider for my scenario? Or is the 15RLS3 the best choice.
I just make it up as I go along! :-)
There are lots of mini-splits out there, and I'm certainly not familiar with them all. It can be important to have reasonable local support (multiple local installers, but also local distributors) of any equipment you intend to install. The top tier vendors in terms of quality equipment and usually decent factory/distributor support seem to be (in no particular order, so let's make it alphabetic) Daikin, Fujitsu, and Mitsubishi, but not all models are as popular (read: "likely to have spare parts in the warehouse if it ever needs repair") as others, and going with less supported vendors or less popular models isn't always a good idea, even if it's a perfect fit.
That said...
LG's Art-Cool Premier series have decent specs and tend to be value-priced. Even the 1.5 tonner modulates down to ~3K in cooling mode (just like the 15RLS3), and the 1.25 tonner has a max cooling of 21,000 BTU/hr, the 1.5 tonner's max out is 29,500 BTU/hr. The 1.5 tonner wins in the "overkill" category, but will still deliver decent efficiency and comfort.
http://cdn-tp2.mozu.com/11590-15709/cms/files/ArtCool_Premier_LA150HYV2_Submittal.pdf
http://cdn-tp2.mozu.com/11590-15709/cms/files/LA180HYV1_Submittal.pdf
modhttp://www.totalhomesupply.com/18000-btu-24-seer-art-cool-premier-single-zone-heat-pump-mini-split/p/LG-LA180HYV1
If you see LG compressors all over the neighborhood (a indication of good support) they're worth at least considering.
Wow, thanks for reminding me of the LGs! My installer likes them too, and Samsung. I looked at those too but they have high minimum cooling around 6k btu. The LG 15k unit looks perfect actually! It's amazing that the range is 3k-21k! I have never seen such a large range. And pretty good efficiency at 24 seer. The only negatives to the LG are that it's significantly bigger inside, about 7" longer than my Mitsubishi, whereas the Fujitsu is about the same. And the LG is slightly louder too. The airflow CFM is crazy high though, like 700 something on max which would cool down my place quick I bet. Not an easy decision to make!! I wish there were some reviews online of this particular unit, but I think they just released it in Dec of 2016 from what I could gather.
The Art Cool Premier been out there for well over a year now (maybe even more than 2 years?), but the amount of online feedback is spotty. Several years ago LG had a reputation among some installers of having a high manufacturing defect rate, but most were still running fine years after installation. I'm not sure how much the factory quality control has improved. I have no direct experience with LG (then, or now) but if your installer is willing to support the product, it's worth considering. Check the warranty info.
I recently saw Mitsubishi's warranty in action at a relative's place in WA. The FE18 that failed is 4(?) years old now and still under warranty. The failure was caused by a MAJOR powerline fault. (A truck knocked over a power pole, dragging multiple transformers on the string down onto the pavement, with one or more of the 2KV phases shorting onto the house-voltage output lines, blowing up 3 transformers and lots of residential equipment in the neighborhood, not just the mini-split.) Mitsubishi accepted it as a warranty claim, and although it took four site visits (and a 3 week wait for one component to arrive from Asia) to find & replace all sub-components that failed, it's back up and running. The labor costs are not covered by Mitsubishi, and the labor DID add up (some of it was eaten by the repair tech for not getting it all right the first couple of times), but she's making a claim to recover those costs with the power company, who is passing it on to the truck driver. But it was surprising to me that Mitsubishi would cover ANY of it under warranty, given the cause of the failure/damage! It's not clear whether other vendors would be as accommodating!
wow, that's crazy Mitsubishi honored the warranty given the circumstances. They must've had a good contractor to advocate for them as well.
As for the Art Cool Premier, yes I think you are right - they have been out for awhile. I was just talking about the HYV2 (2nd generation) of them. The HYV1 are still for sale in certain sizes. They're basically the same but the new ones are more efficient I guess. Not sure I want such a large indoor unit. I'll have to think it over. Thanks again for your help with this! no one knows what I am talking about.
Wanted to give an update: I had the Fujitsu 15RLS3Y installed today. AMAZING!!! so powerful. I actually had to turn it down because it was getting so cold. It goes to the set temp in no time. Once there it cycles on and off. On low it puts out way more air than my old FH12NA. That thing does nothing compared to this. I wonder if I could've gotten away with a 12RLS3Y but this is fine. I will need the extra power on the hot summer days approaching. I think this will be much more efficient than the Mitsubishi...that one had to be on at full blast to do anything and was only putting out 60 degree air on AC full blast. This one is putting out air at 43 degrees with a much higher airflow! I should've listened to my installer 3 years ago when he tried to steer me away from the Mitsubishi's. The Fujitsu is SO MUCH BETTER in every way. I used to be able to hear the old condenser droning and vibrating away under full load. I can BARELY hear this one. The wi-fi access is a nice plus too. I have it all set up on my iphone already and it works great. Thanks again for your help, Dana!
Getting down to the setpoint quickly is usually a symptom of being oversized. I'm not sure what the minimum modulated output of that unit is, but usually a "set & forget" strategy will use less power. The efficiency when running at full blast is very low compared to when it's running at minimum-speed.
If it's cycling on/off at minimum speed it isn't as efficient as modulating at low speed, but it's still WAY more efficient than when running at full speed. See Figure 14 of this document:
http://www.nrel.gov/docs/fy11osti/52175.pdf
When it's ~87F outside at max speed this older Fujitsu 1-ton is delivering a COP of ~4, but at minimum speed it's efficiency more than doubles, to the 8-9 range. Even when taking a hit from cycling it's still NOTHING like the efficiency hit it takes from running full blast on a recovery ramp to a lower set point. Unless you are only running the recovery ramp during a very low priced time of use rate structure would turning it off or bumping it up while you're out for the day be more cost effective than "set and forget".
Minimum capacity is listed as 3,100 btu on the 15RLS3Y and on my old FH12NA it was 2,500. So I thought pretty similar. On the link you provided in figure 13 (section 5.1) it says on the 12RLS they weren't able to achieve their minimum modulation of 3,000 and instead were only able to hit 5,000. I wonder if that is something common with Fujitsu's? That is 2 model generations back so maybe things have changed. At any rate I guess it doesn't matter since my old one wasn't even able to hit a set point ever until long after the sun went down. The 15RLS3 hits set point within about 30-60 mins which for me is nothing compared to the Mitsubishi!! Hope i'm not oversized but i'm just glad it can cool this place down now. It really removes a ton of moisture too. You should have seen the condensate line outside, it was easily more than double the water removal. It's not quite as quiet as the Mitsubishi I had, but the cooling is so much better - like night and day. I'm glad I didn't go with the LG ArtCool Premier. That would've been much bigger evaporator (like 30-40% bigger), louder, with extra capacity up to 21,000btu that I wouldn't have needed. I agree though, the set it and forget it approach is where the highest efficiencies are achieved.
With the same minimum modulation as other Fujitsu units, if it's oversized the down sides are higher minimum cfm at the head with higher noise levels and crummier latent cooling. In San Diego the latent loads are usually negative and the peak latent loads still quite small. If you're happy with the sound level (which is pretty quiet for most Fujitsu units) you should do OK, even if it's oversized.
Thank you, Dana. You have all the answers!
Let me ask you this: if my old 12k unit was noticeably undersized, how did I end up oversized at 15k? Maybe im just a bit oversized? Do you think my 12k unit was defective or improperly installed?
I had the same contractor do both jobs and he said the Mitsubishi was indeed working normally.
Having a low min capacity is actually where you get into trouble in the summer. The min CFM is too high so you end up with 10F drop only so you are at 65%-70% humidity if you run at those min capacity. Typically a min capacity of 6000btu or more is better.
Research high humidity issues and you'll see people having mold issues with these low min capacity brands (fujitsu, lg, etc).
Yes, there was probably something defective with the FH12, but can't really speculate as to what was defective about it.
The max cooling on the FH12 is 13,600 (same as the 12RLS3) but if the 15RLS3 maxes out at 18,400 BTU/hr (not sure about the -Y version), which is a substantial step, a much bigger step than the "nominal" modulation levels would indicate. It's really more like a 1.5 tonner than a 1.25 tonner, but it's test numbers would't be as favorable if tested modulating at 18,000 BTU/hr. The choice of calling it a 15,000 BTU/hr unit is largely a marketing issue, which allows them to test it's efficiency at much higher efficiency point on the curve.
High humidity is not a problem in San Diego- the average latent loads at 55% RH are negative in that location, and the peak latent loads are quite small. In other locations (such as swampy shores of the Gulf of Mexico) oversizing and low modulation can be real issues, but in all mose cases that can be improved by running the unit only in "DRY" mode to guarantee at least some latent cooling, even if it means more cycling.
On the old FH12NA the CFM's range was about 140-400. new one is about 260-550. I thought I would need more CFMs because I couldn't really feel it in my kitchen where I work sometimes. The FH12NA was excellent at dehumidifying (Relative humidity around 45-50%) but struggled with actually reaching and keeping a set temp. The new Fujitsu 15RLS3Y (the Y suffix just adds wi-fi) does a great job with reaching and keeping set temps but the humidity is higher, around 55% RH. I guess that's not really that high of a RH but it's noticeable the added moisture in the air. I go down to see if it's putting out condensate outside and it is dripping like it normally would before.
Now I'm confused on why a low minimum capacity would be detrimental to humidity in the summer time. I thought that would be an advantage that it could modulate lower. Why would you want a higher minimum? 3k vs 6k minimum is basically what the difference between the units were.
Maybe I should just give it some more time. The other day wasn't that hot. only around 85 and not very humid. San Diego's summers have become more muggy than I remember growing up. Now we usually have about 2 solid months of what we would consider muggy weather - dew points between 65-70, and temps in the 90's for inland SD.
This is such an interesting topic. I had no idea HVAC designs could be so complicated with so many variables. At this point it's all installed so not much I can do except run it on quiet mode to keep the compressor running longer trying to achieve set temp, right? Dry mode would do the same thing?
Thanks
It really depends on how low the CFM can get. Eg: if min CFM is say 300, and min load is say 3000btu, that gives a temperature drop of about 10F. Say you are at 75f, that means dew point can be as high as 65f, which gives relative humidity of 63%. If min CFM is 150, then you won't have a problem.
I didn't realize you are in San Diego, so you don't have to worry about it. It's a huge problem if you live in the South though.
If it's going to be humid-to-muggy out (read "outdoor dew points north of 60F") , run it in "DRY" mode and you'll be just fine. In DRY mode it will cycle on/off a bit more if it's muggy and not so warm outside, but it's really not a big deal- efficiency is pretty high even when cycling when it's not very hot outside.
In San Diego outdoor dew points don't tend linger in the 60s or 70s for long (even if it peaks in the 60s & 70s during the day), adding up to less than half the time even during the summer months. This is very different from the right coast (or gulf coast) states where midsummer average dew points are well north of 60F (even north of 70F in some gulf state locations.) If outdoor dew points aren't above 65F except when it's hot out (a real sensible load) it'll probably stay below 60% RH indoors except on the worst days with "set and forget" operation and standard "COOL" mode operation.
For those with dust mite allergies it's important to hold the line at ~50% RH indoors, but for most others 60% is still fine and healthy. Mold doesn't become a serious problem until it's higher than 65% RH indoors for extended periods of time.
thanks. I'm beginning to wonder if my CFM's on the FH12NA just weren't high enough to cool my place off. It did a great job of dehumidifying. it's CFM's were: 137-167-221-304-398. btu's 2,500-13,600. Maybe it was defective in some way though. it really struggled on hot days.
Now i'm looking at the 9k and 12k units again. It looks like most of them offer the same capacities basically. Most go from 3k btu to 13k or so btu.
I'm looking at 2 now:
The Samsung Smart Pearl 9k which offers 3.1k btu - 13k btu. cfms 240 to 460 (doesn't specify sleep mode cfm).
LG Art Cool Premier 9k offers 1k btu -13k btu. cfms 283 to 547 (doesn't specify sleep mode cfm).
What's the difference having the minimum capacity of the LG at 1k vs to Sammy at 3k? The LG throws more air around it looks like. Any suggestions?
I'm going to have a company do a Manual J for me to be sure of my needs.
Greg,
I am also considering the 15RLS3(H)Y. Regarding cycling, I read on other threads on this site that a remote thermostat might help; there were concerns with the sensor being tripped owing to not much clearance where the indoor head might be installed.
Also, FWIW:
https://www.greenbuildingadvisor.com/community/forum/energy-efficiency-and-durability/47671/remote-thermostat-fujitsu-mini-split