Minisplit efficiency seeming to drop
Hello,
I had a 5 head 42kBtu Mitsubishi split installed and started in 2017. There are 4 half ton heads and one ton and a half. This is a hyper heat unit. This last year it seems as the efficiency has been dropping as my kw usage is climbing, but my habits haven’t changed. It also doesn’t appear average temperatures have changed too much in the past few years (based on the electric companies graphs).
I also have noticed the largest head seems to drift 4 degrees or more away from set point (as indicated by two green lights on the head) regularly, but rarely, if ever, runs on the highest speed.
I did have a few issues with the initial install ( refrigeration leak) and they programmed the unit improperly. I believe the programming was fixed, and obviously had them pump down the system and recharge it after noticing the leak.
Anyone have any thoughts or things I can check? I wash the filters at least monthly, many times they aren’t that dirty.
Thanks in advance!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
It's probably low on charge from a small leak. You might find it useful to buy a leak detector and find the leak yourself.
This is the one I used, and it worked great to find a leak that caused my minisplit to perform poorly after ~1 year of operation. It turned out to be scratch on the flare surface.
https://www.elitechustore.com/collections/refrigerant-leak-detector/products/elitech-ld-100-refrigerant-leak-detector-cfcs-hcfcs-hfcs-freon-gas-leak-sniffer-semi-conductor-detector-with-long-probe
+1 for checking for leaks, especially since you’ve had leaks in the past.
The “wands” that sniff out leaks are pretty sensitive. Check around ALL fittings, and any shape bends in the lines. I’ve seen some screwy brazed butt joints in refrigerant lines before, so be on the lookout for any oddball trouble spots while doing your checks.
The wands don’t always catch very small leaks. For really small leaks, I like the dye that gets injected into the refrigerant lines. The dye can accumulate in the area of very small leaks over a period of days or weeks, then you go looking for the glowing spots using a blacklight. If you suspect a leak, but you don’t find any leaks with the wand, have your HVAC guys put some dye in the system, wait a few days to a week, then go looking for leaks with a black light.
Bill
Bill,
These newest detectors are much better than the old ones. I came from automotive leak detectors to newer heated diode ones. The difference in performance is astounding. There are, however, some tricks. You need to be working in still air, if there's a breeze you'll never find the leak, it gets carried away too fast. Also, you've got to scan the probe very slowly along the underside of the line/fitting, holding it quite close.
Yeah, I've been impressed too. The only times I've had trouble with them is leaks that are in air paths (like around the TXV), and with slooow leaks that would be a few bubbles a day or so if you bubble checked them. The dye is good for those really slow leaks that gradually bleed off refrigerant over a period of months. The wands are good for faster leaks.
My usual methodology would be to check with wands, and if nothing was found but a mysterious leak was still suspected, we'd put in a cartridge of dye. After putting in dye, we'd sometimes check the following day, and if we still saw nothing, we'd check a week later. It was VERY rare to not find the leak with this method. I think the only time we missed it was on a carlisle compressor around something screwy like the electrical connection.
Bill
Thank you for your replies. Hopefully this is not the case on some of the joints that are within interior walls ( installed during an addition) and will be real hard to detect. Would it be best to test while heating and cooling or will the pressures be high enough in both lines during heating alone?
Do the test while in heating mode. The pressures in heating mode are 300-500 psi compared to 100-200 in cooling mode.