Mitsubishi Mini-split Overshoot — time to cut jumpers?
I have a 1.5 ton Mitsubishi mini-split serving a living room of 500-600 sq ft. Since installation, it has never been able to regulate the room temp close to the setpoint. The problem is that the temp at the air intake close to the ceiling is not representative of the room at living level and is adversely affected by outdoor temps during the day particularly in the cooling cycle.
An MHK2 thermostat was installed two years ago. It changed matters so that now, in cooling mode, the unit drops the room temp about 2 deg F below the setpoint. Even though the room is below the setpoint, the unit cycles on from time to time. The offset between the setpoint and room temp varies day to night and with the outdoor temp. This mini-split is easily the worse heat pump I have ever lived with and the only one that can’t regulate the room temp.
I have found Mitsu Application Note 3048 — How to turn off the indoor unit fan when set point is met. On my unit, this is done by cutting two jumper wires. A local HVAC company spent one hour on the phone with Trane tech support to confirm that cutting the jumpers (and which two) will shut off the indoor fan when the setpoint is reached. That is a recommended solution to poor temperature regulation that I’ve found on this forum and others.
Is cutting the jumpers the right solution to my problem? I ask this because I can see that the onboard control is still active and trying to control room temp based on the unrepresentative intake air temp. Will stopping the fan take the onboard control out of the loop?
I know that the onboard control is still active because I placed temp sensors at the air intake, the cold air outlet and next to the MHK2 to monitor what is going on. As the room warms up, the air intake temp (top of the unit) reaches the thermostat setpoint first (due to stratification and the outside influence), which triggers the unit to come on even though the room temp is still 2F below the setpoint.
Thanks for any help you can offer.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I have found that mini splits in general do not do well regulating to an accurate setpoint when left to their own devices.
i manually adjust the fan do get it to do what I want.
You know the pattern you are seeing.
I turn the fan down to low at night, and then depending one what the weather is going to be, turn it up in the morning.
If it is actually cool at night I may shut it off completely
Let me add the resolution to my mini-split control problem. I now have this running satisfactorily. I run a ceiling fan continuously in the center of the room to mix the air vertically so that the air temp at the intake to the unit is representative of the room. The setpoint has to be +4 deg F above the temperature I want! Some of this is a temp delta from 5' above the floor to the intake point (9' off the floor) plus it appears that the internal temp sensor is not in the intake air stream but located off to the side where it is heated by the equipment running inside the indoor unit. The ceiling fan keeps that differential constant. Then, I covered up the IR sensor with aluminum foil except for one small corner so the remote works. This sensor looks at an exterior wall (3 of 4 room walls are exterior) that is >50% windows and out across the porch into the courtyard. It apparently sensed the outdoor temp as if it was the temp on the interior surface of the wall causing the setpoint to drop by 2 to 3 F during the hot afternoon hours. Now, I get about 1/2 deg of setpoint drop in the afternoons, but that is about OK. Overall, I managed to re-engineer the room and negate the unit’s IR sensor to the point that it holds the room a comfortable temp. The thermostat is used only to set the setpoint. What a piece of HVAC crap.
The Mitsubishi mini-split may be a fine piece of equipment when installed in the theoretical room that it was designed for. But that's not my (pretty typical) home. Maybe these work fine in European or Asian markets where building methods and standards are a lot different, but not in the U.S. My installer (RTFT here in Tucson) did me no favor recommending this, installed it too close to the ceiling, failed to understand (or tell me) that the old system's return air duct had to be insulated and roofed/drywalled off to eliminate the outdoor influence, and has spent the last year avoiding my calls. Trane did nothing to help me either. They did recommend 5 "top" installers in Tucson who had received their highest level of training. Number 1 had already been out and said they had never seen a mini-split on a thermostat. Three more weren't interested in the job. Number 5 sent a tech out, who spent an hour on the phone with Trane tech support. The tech told me that they were "traned" to handle the mechanicals including the compressor and refrigerant, but had no ability to diagnose control system problems. BTW, Trane confirmed that you can convert a Mitsubishi mini-split to work on thermostat control by shutting down the indoor unit fan (see Mitsubishi Electric Application Note: 3048). It is done by programming on some units, by cutting jumpers on other units including mine. After Trane identified the jumpers to cut, the tech's mgmt refused to do it since they didn't want the responsibility. So much for "traning".
Sorry for the rant. I hope this information helps someone else.
Is the problematic head the only head attached to the compressor?
IE
mini split = one head and one compressor.
multi split = several heads connected to the same compressor.
Walta
Yes. Mini-split = one indoor unit + one compressor.
Deleted
18,000 BTUs of equipment seems like a lot for only 600 square feet.
Is a hard wired remote sensor like this one available to fit your model?
https://www.ecomfort.com/Mitsubishi-PAC-SE41TS-E/p30827.html
I would be surprised if Mitsubishi was using the remote controls IR receiver to sense the rooms temperature. If you have and documentation relating to the IR sensor, I would like to read it.
Walta
It is a 1-1/2 ton unit for 500+ sq ft with 85 percent of the wall area being exterior. The same installer did a Manual J for the predecessor (roof top) unit 15 years ago and came up with 1.5 tons cooling. Maybe it should have been 1 ton for the mini-split. Right or wrong, that is what I have and why.
I do not know about the wired remote sensor. I believe the unit senses intake air temp using a “room air sensor” on a circuit board off to the right side of the unit inside the case, which is why it runs several deg F above the intake air temp. Would the wired sensor replace the onboard sensor for that purpose?
I have the wireless remote temperature sensor for the MHK2 thermostat and have used it in the recent past (but not now). When under thermostat control, and sensing the temp at either the thermostat or the wireless sensor, the on-board control system continues to run in the effort to fine-tune the temp of the intake air -- UNLESS you shut off the fan when the setpoint is reached by programming or cutting jumpers (per App Note 3048). As best I can tell, it still uses the intake air temp in that effort.
I decided not to cut jumpers once I determined that I could run it with on-board control by stabilizing the sensed intake air temp and adjusting the setpoint +4F above the desired room temp. It now runs with the thermostat providing the set point but the air temp sensed at the indoor unit.
If I cover the IR remote receiver with aluminum foil, the unit will regulate to a very constant 79F room = 83F setpoint, with the ceiling ran running to stabilize the room-to-intake differential. Pull the foil off and it drops the room temp by 2 to 3 F every afternoon as it looks out into the courtyard under the Tucson sun. I have assumed this is the onboard control system using the IR radiant energy that it senses to make a setpoint adjustment. With just a corner of the sensor area left uncovered, I see a regular drop every afternoon of around 0.5 F, which is not bad and does improve the comfort level (even though I am not in the room to enjoy it). This unit does NOT have the I-see sensor. (In the main part of the house, the two remote sensors do this for the Honeywell VisionPro thermostat that controls a rooftop Trane HP but it is a small adjustment. None of them look outside through a window.)
If it is actually something else going on with the IR receiver, I’d very much like to know. Thanks for your comments, Walta!
Mitsubishi and Fujitsu will cool to -2 of set temp and heat to +1… or thereabouts. This is the nature of the beast. They are not precise. Set up the target temp to your liking and let the machine do its thing
I did not do your manual J but at 400 square feet per ton is 30% above the local lazy rule of thumb of 600 SQF per ton that almost inevitably oversizes a system by 30% or so. Seems like you could almost leave the window open and still have the unit cycle off.
I suggested the remote sensor because I personally do not believe the onboard sensor can accurately sample the room temp and the remote sensor would allow one to find good sample.
Does it seem your unit cycles on and off several times per hour? With your variable speed compressor on hottest days, it should speed up or slow down and never turn off.
Consider if your unit is oversized it would almost always be running at its slowest possible speed. Note at slow speed getting oil to flow back to the compressor is an issue after so many minutes of slow run time the computer must run for a few minutes at max speed flushing the oil back to the compressor. An oil return cycle running could explain your overshooting the set point from time to time.
Walta
Thanks for your reply. I am finishing the drywall closure of the old return air duct above the indoor unit and will repeat my tests after that before moving forward. It's already filled with fiberglass batt, so closing it up should not make a lot of difference. I like the idea of a wired room sensor at living level if one is available for my unit. That will rationalize things.
IIRC, the original Manual J said 1 ton cooling load then 1/2 ton was added for a rooftop mounted unit. The mini-split outdoor unit is on a pad and never exposed to direct afternoon sun, so could probably have been a 1 ton unit. I detect no short cycling.
In fact, the unit is clearly moving refrigerant through all the time and modulating it. On a recent afternoon, the outlet air delivered into the room between 2 and 3 pm ranged from high of 68F down to low 60F when outdoors was reaching 105F and the indoor room temp was steady near 79F. There are 2 or 3 modulation cycles within a typical hour in the afternoon. Sound right?
At this point I am glad I did not cut the jumpers and decided instead to try to re-engineer the room first so that the unit could work under on-board control. I am outraged at my former installer for walking away from this, at Trane for giving no help, and even the company that sent a tech to confirm which jumpers to cut and then refused to do it. Maybe I should be glad for their caution, because it led to the current situation which I think is better than a conventional on/off thermostat controlled heat pump.