Mini split experts: system seems to be losing efficiency

[Flargen]

Hey guys,

I installed a Mitsubishi SLZ/SUZ-KA15NA mini-split setup in my garage a few weeks back, and over the last few days it seems to have started having trouble holding ~75°F. I'm not sure if this is due to the fact that it has started to get even hotter, but as a reference, the current outdoor temp is 88°. The garage is fully insulated (max R13), but there are some small air gaps (haven't installed the molding yet around the exterior door or the attic access). There are minimal air gaps on the garage door itself (about as tight as I've ever seen).

I've gone through all the maintenance menus and found no error faults, etc.

I put the system into "Test Run" mode and after a few minutes, the coolant pipe temp hit a low point of 10°F. After a few minutes, it stabilized at 21°F. Still no error codes. During said test run, the air coming out of the cassette was WAY cooler than it is in normal operation.

Do these things adapt to their environment in order to maximize compressor efficiency? If so, is it possible to reset these adaptations or modify them in any way?

I'm about to check the low-side pressure but would love any input you can give.

Indoor: SLZ-KA15NA
Outdoor: SUZ-KA15NA
Controller: PAR-33MAA

Thanks in advance.

 

[Flargen]

Went out and checked the condenser and found that the liquid line (small) was completely iced over, which immediately pointed to a low charge... I went to loosen the low-side service port cap and realized it was only on there finger-tight.

Low side pressure (in "Test Run" mode, which should be full-tilt): 60 PSI. I think I've found the problem.

Gonna empty the system, tighten all flares, pull a vacuum, and wait for a buddy to arrive with some new refrigerant.

Given that the system held a full vacuum for 20+ minutes when I set it up, I get the feeling the service port valve may have been leaking, causing the refrigerant loss. Woo!

 

[PoorOwner]

As far as I know the test run is at 75 degree set point with indoor fan on high.
I have seen test run with the condenser fan running slow so it’s not a full speed run.

But 60 psi is low. No matter what speed it is at.

The service port cap can be not there and it wouldn’t have leaked out that fast unless the valve is truly defective.

 

[Flargen]

Hmmm... interesting. Sounds like I may need to get a nitrogen cylinder and pressure test the system since it held vacuum without issue.

 

[LS6 Tommy]

Went out and checked the condenser and found that the liquid line (small) was completely iced over, which immediately pointed to a low charge... I went to loosen the low-side service port cap and realized it was only on there finger-tight.

Low side pressure (in "Test Run" mode, which should be full-tilt): 60 PSI. I think I've found the problem.

Gonna empty the system, tighten all flares, pull a vacuum, and wait for a buddy to arrive with some new refrigerant.

Given that the system held a full vacuum for 20+ minutes when I set it up, I get the feeling the service port valve may have been leaking, causing the refrigerant loss. Woo!

Vacuum is not for leak checking.
I must post that ten times a year...

Tommy

 

[Flargen]

Vacuum is not for leak checking.
I must post that ten times a year...

Tommy

It'll reveal some leaks, but only catastrophic ones. But yes, you are correct... I've been putting off buying a nitro rig for a long time but can't put it off any longer.

 

[PoorOwner]

Did it hold vacuum -29.9 inHg or did you hold a micron reading?

-29.9 mercury is like -14 PSI.

Micron gauge reading can detect amount of air infiltration. Unfortunately alot of gauges they let air infiltrate as well..

 

[Flargen]

-29.9 inHg, or at least as precise as my gauge set is (cheap Dayton setup from Grainger).

Just grabbed a nitrogen cylinder and will be picking up a regulator on my way home.

I'm guessing I can close the liquid/discharge valve and run the unit to capture what refrigerant remains? And then close the suction valve, shut it off, and pressure test?

 

[fnbadaz06]

When I had my mini-split tested, the vacuum being applied was read as a micron reading.
He pulled vacuum for 15 minutes and said anything under 400 was good.....mine was under 360.

 

[Milton Shaw]

I would pressure test with the refrigerant and a sniffer, nitrogen only tests with bubbles leak unless you add refrigerant . A good sniffer should find any leak. Pressurize the system with R410 to over 200 psi with vapor from the tank not liquid and you should be able to find the leaks. I have even used an ultra sonic leak detector to find leaks with.

 

[PoorOwner]

Follow the manual's pump down procedure, via test mode, but with a known leak you should shut it off at just before 0 psig. Get a helper to shut it off.. it probably will take 2 presses on the button.

I would not go too much higher than 300 psi because the nitro may make it into the condenser. Remember to isolate so only the gauge is holding pressure, because the regulator will replenish any loss pressure if you keep it on there.

You may want to consider to "top" off, it may only take 1 pound or less to make the frost go away and blow cold like new again. Observe for the rest of the summer making sure it does not go low again. Then you should pull all charge and weighting in, which is the ONLY, PROPER way to charge a mini split.

here is a video of a tech just topping off a frosted up unit. Not the right way, but not a big deal for them to come back...

 

[justinjoyal]

Measuring vacuum with refrigerant gauges is NOT the way to go.

Also I always use Nylog on caps, in case the valves would leak.

 

[Flargen]

We ended up starting by topping off the system to get the low-side pressure to ~125 psi. My buddy came by with his refrigerant, scale, etc... And the system took 1 lb 11 oz (!!) to return to normal operation.

We busted out our sniffers and couldn't find any leaks outside, but got one small blip at the cassette end of the lineset (but only with the sniffers in their highest sensitivity modes). Given the lack of any other false-positives, I'll be taking the insulation off and inspecting it further sometime in the next couple days.

Thanks for the tips, guys.

 

[bzinsky]

It'll reveal some leaks, but only catastrophic ones. But yes, you are correct... I've been putting off buying a nitro rig for a long time but can't put it off any longer.

I’ve only done 20 minisplits, but hired some ***** to help with install and he screwed up half the flare connections.

Every single leak I could tell by using a vacuum and micron gauge. Talking extremely small leaks that held nitrogen for 24hours. All showed up.

I suppose there are some leaks that would seal themselves under vacuum and open under pressure, but from what I’ve seen, when it’s specifically a flare connection leak, it’s all the same with negative or positive pressure.

 

[bzinsky]

We ended up starting by topping off the system to get the low-side pressure to ~125 psi. My buddy came by with his refrigerant, scale, etc... And the system took 1 lb 11 oz (!!) to return to normal operation.

We busted out our sniffers and couldn't find any leaks outside, but got one small blip at the cassette end of the lineset (but only with the sniffers in their highest sensitivity modes). Given the lack of any other false-positives, I'll be taking the insulation off and inspecting it further sometime in the next couple days.

Thanks for the tips, guys.

Soap and water will find it if its at the flares