Repairing Mitsubishi Mini-Split Inverter Board

RedOak

We had a very rare lightning strike nearby that caused a power surge. After resetting several GFCI outlets and a couple of breakers - including the mini-split's breaker - we had no AC.

The indoor evaporator unit (MSZ-FE18NA) can be turned on normally, via the remote control, but the outdoor condenser unit (MUZ-FE18NA) is showing very little signs of life.

As I have an electronics background, I felt comfortable with testing the line voltage, pulling the outer disconnect and opening up the condenser unit itself. A careful visual inspection of the inverter board inside of the condenser unit revealed that F62 (2 Amp Slo-Blo Glass Buss Fuse) was black inside and, obviously, open. Needless to say, I was very surprised that no other components on the inverter board appeared to have overheated, so I inserted a jumper across the fuse contacts, crossed my fingers and reinstalled the disconnect. Unfortunately, other than a faint buzz, there are still no other signs of life coming from the condenser unit.

Employing the PDF of the Service Manual for the Outdoor Condenser Unit, I found a schematic drawing for the Inverter Board's Power Supply on pg 8. For the most part, Mitsubishi didn't use standard electronic component symbols on this drawing, which makes troubleshooting difficult, but I found that the open fuse (F62) was in series with a couple of components that are wired in parallel: PTC64 & PTC65. In the bottom-left corner of the same drawing, there is a small chart that names most of the non-standard symbols on the drawing. Both PTC64 & PTC65 are described in the chart as being "Circuit Protection," which leads me to believe that these components - which are both encased in a black plastic case that very closely resemble a relay cover - are either resistors or metal oxide varistors (MOVs) installed for surge protection.

The cheapest Inverter Board assembly I could find for the MUZ-FE18NA outdoor condenser unit costs well over $600.00, which is more than we can afford, so the question I'd like to ask the HVAC forum board is out of necessity: Does anyone know where components for this type of Inverter PC Board can be purchased?

By the way, even if we could afford to buy one, the Inverter Board in question looks way too pristine to simply replace and throw away - without at least having given component-level repair a shot first.

Many thanks for the help ~
Red

Jamie Hall

If you are well versed in electronics, and willing to be rather brave, I think I agree with you --there is a very good chance that those are, indeed, probably MOVs to protect the rest of the board. In which case there's a pretty decent chance that they died protecting the board. At least you can hope.

Obviously the first check is to determine if they really are open.

Then the next trick will be to analyse the circuit. If they are protection MOVs, they will be first in line on the input power, before pretty much anything else except maybe a fuse or two.

Then the next trick -- if they really are -- is to make an educated guess as to a replacement. Digikey has a mind bending list of them... and that I can't help you with, unless there is some kind of markings on them you're just going to have to make as good a guess as you can.

JUGHNE

Before you spin your wheels, did you ohm out the compressor? All 3 phases to each other and then to ground. If any winding opened or shorted to ground then the unit is toast, IMO.

How old is this unit?

We have all probably thrown away good looking boards.

As far as the cost of a new PCB, that is just the way it is in the Minisplit world.

RedOak

@Jamie Hall - Thanks for the reply, JH. After reading the OP again, I can see where some confusion might arise about the components on the Inverter Board in question.

The fact is that we've had one local HVAC contractor come out (so far) to take a look at the mini-split system. We're expecting a couple more HVAC people to come out, so I haven't actually taken the Inverter Board out - in case any of the HVAC pro's want to take a close look at our system.

The first guy that came out never turned the system on, typed the outer unit's Mitsubishi part number into his "tablet PC" and, eventually, gave me a completely outrageous estimate to replace both the inner and the outer unit (we have a single zone system) ... but the point here is that I didn't want to take the Inverter Board out of the outdoor condenser unit after we get a few HVAC guys out here. Who knows, maybe one of them will actually attempt to turn the system on, right?

RedOak

@JUGHNE - Thank you for the post as well. If you read my last post, you'll read why I didn't get too far into troubleshooting the outdoor compressor unit, but, after we've had a few HVAC pro's come out and give us their thoughts on this, I will take the inverter board out for further testing and test the compressor itself.

I've been away from the electronics world for a while now, but I did spend a career calibrating, aligning and component-level repairing precision measurement test equipment (PME) for the USAF and, later, as a civilian DoD contractor ... so it's pretty tough to simply trash a relatively simple PCB without at least checking it out first - especially on a fixed income.

The sad part, in my mind, is that Mitsubishi - the apparent king of mini-split HVAC systems - doesn't seem to be training their people at the component level. The lack of proper electronic symbols on their schematic drawings makes that pretty clear.

EDIT: I forgot to mention that this system is less than 10 years old.

Harvey Ramer

@RedOak If the unit is under 10 yrs, it should be under warranty if it was registered. Warranty replaces the whole board.

Pcb's in residential hvac do not get component level repairs in the US. That's just the way it is. There is no training or documentation available for contractors to work on them. Some other parts of the world are different. 

I've made a few minor repairs to them when the units were out of warranty. Saves the homeowners a lot of money in those cases. But I don't have the knowledge to go deep on boards. 

Frankly, if the board is malfunctioning, you are likely going to have to replace it unless you fix it yourself. 

RedOak

@Harvey Ramer - We've been here nearly five years. I was told that the mini-split system was installed 3 years prior to the sale, but I wasn't given any further information about the system or, more importantly, who installed it. This is something worth thinking about, however ...

We just had a second HVAC pro out, earlier today, and he didn't ask about the warranty period, either .... hmmm.

ratio

I've seen component specs exactly once in the last decade of this kind of work. (Aaon specs the rs485 transceiver chip on some of their boards—& they're socketed!)

Your contractors will not do component level repairs to the board, they will offer at least board replacement, & as you found out yourself they may not even do that. The necessary skills for component-level diagnostics aren't really found in the HVAC service industry. Board-level diagnostic skills aren't even all that common.

I would do it for my own system. IDK if I would do it for Grandma, but I certainly wouldn't do it for anyone else.

JUGHNE

Usually homeowner's insurance covers damaged done by lighting strikes.
If so your best case would be that the compressor is shorted or open and also the PCB is toast.

Any HVAC person can check the compressor. Most of these people would condemn the entire system if compressor is burned.

Warranty at best would give you parts only. PCB is easy fix. Compressor replacement not advised, IMO.

RedOak

@JUGHNE - You used the word "burned" to describe what may have happened to our outer condenser/compressor unit. Do "burned" outdoor units normally show visual signs of being overheated by a power surge?

109A_5

Hello, @RedOak,

Repair Parts... Digikey.com, Mouser.com, Newark.com, aliexpress.com, ebay.com, there are others.

Should be enough information that a good Electronics Technician could repair it unless the micro-controller is damaged or the needed parts can not be found. More information there than many other systems.

Did you check the other fuses ? F601, F880, F901... May be others in there too.

Lightning can do odd and amazing damage.

STEVEusaPA

@JUGHNE gave the best advice. A lightening hit could've taken out the compressor and the fan motor too. Sounds like you didn't have a surge protector.
You buy a board only to discover other components are toast, or the other damaged components wreck your new board.
Make sure you check everything else before throwing any money at it.

dko

PTC64/65 are PTC-type inrush current limiting thermistors 33Ω

But like above said, many other things to check as well.

RedOak

@109A_5 - Thank you for posting the one and only schematic drawing I could find for the Mitsubishi MUZ-FE18NA outdoor condenser/compressor unit.

As mentioned earlier, we're still getting estimates for replacing our single-zone ductless (mini-split) system that was damaged by a nearby lightning-induced power surge on Wednesday, Sept 13, 2023. This being the case, I have not, as yet, removed the inverter PCB. Instead, I tested the incoming power - and both legs measured a normal value of about 120 VAC. After pulling the disconnect, I did a very thorough visual inspection and DC resistance check of the obvious suspects, including every fuse shown in the drawing. As it was one of several glass-tube type buss fuses on the board, F62 (2 Amp Slo-Blo) was blackened by soot, but, otherwise, it was intact. All of the other fuses in the drawing are still functioning as intended.

Once again, the blackened buss fuse was the only visual evidence of overheating - one could practically eat off of the rest of the inverter board.

Thus far, the estimates to replace our single-zone mini-split (both inner and outer) have ranged from well over $10K down to to $4K. The deductible for a claim will be $1K, so, needless to say, I'm rather anxious to get the last quotes and open the condenser unit up again. At that point, I'll remove the inverter PCB and do more testing. I will also test the compressor itself at that time.

So far, the HVAC contractors who have come out are only interested in replacing the entire system, so most of the posts in this thread seem to be right on the money. In short, no one is interested in attempting any form of repair work.

RedOak

dko said:

PTC64/65 are PTC-type inrush current limiting thermistors 33Ω

But like above said, many other things to check as well.

@dko - Are the PTC-type inrush limiting thermistors seen on the mini-split inverter boards typically insulated by a black plastic case?

Once again, I haven't been able to remove the inverter PCB, yet, but the black plastic cases lead me to believe that PTC64 & PTC65 were both MOVs.

That said, I'm clearly not an HVAC expert, so I'm going by what I worked on back in my days on the bench. In short, PTC64 & PTC65 look more like small relays than anything else.

JUGHNE

Compressor motor winding are inside the steel compressor can.
It is actually a 3 phase motor with tightly wound windings of different voltages separated only by the varnish on the copper. Similar to power transformer windings.

With lighting, which can do anything it wants, sometimes a surge can blow thru the varnish insulation and short the windings to each other or even the grounded can.

Whenever lighting or a bad power surge is mentioned, the first thing I look at is the compressor windings. If failed then that is the end of the search.

With the power off for at least 30 minutes (the power supply caps can pack a punch), you can follow the 3 largest wired down to the compressor. The compressor may be wrapped in an insulation blanket that comes off easily.

The connection point for these wires is under a plastic cap held on perhaps by a single nut.

Remove the wires and mark their locations.
Check between the terminals with a good Ohm meter. The readings should all match in Ohm readings.
That is L1-L2, L1-L3, L2-L3 should have matching readings....not very high.

Then from each terminal to ground should be infinity.
It is best to check the windings with a megameter if you have one. This puts a high test voltage on the windings to ground.

Some mini's suggest to change the outdoor fan motor whenever changing the main board FWIW.
Supposedly a bad fan motor can toast your new board.

ratio

It's not a three phase motor, although it looks a lot like one. It's basically a big brushless DC motor—the inverter board puts out positive & negative pulses, not a sine wave. Of particular note is that it won't run correctly if the 'phasing' is backwards.

RedOak

@JUGHNE - I don't have a megger - which is a great suggestion - but I do have a Fluke 87 DMM.

Once I pull the disconnect and let the caps in the outdoor unit drain completely, I'll isolate the compressor from the inverter board and test out the motor windings. As you mentioned, if any of the windings are shorted - or don't [reasonably] match the other resistance readings - it's "Sayonara" to the old outdoor unit and "Konnichiwa" to a new Mitsu condenser unit and indoor head combination.

On the other hand, if the compressor windings do check out, I'll remove the inverter board and proceed with further circuit testing. Given the "like new" [visual] condition of the board, I've got my fingers crossed that warming up the Weller soldering station and replacing a few components might do the trick ... but I won't be holding my breath.

RedOak

@ratio - Your thoughts about "phasing" are very well taken. Considering that the wiring has never been touched, I'll be sure to keep everything labeled properly before taking anything apart. A few "before" photos wouldn't be a bad idea, either. Thank you kindly for the post.

109A_5

Hello, @RedOak,

The PTC64, 65 most likely provides a path to power the logic and communications electronics and maybe a soft start path to charge CB1, CB2 and CB3, since they and fuse F62 makes a path around X64. X64 most likely is only closed if a motor needs to run or after a power up delay.

It would not surprise me if there are MOV on the board too.

ratio

I mention it because tech support & I condemned an inverter board because the compressor wouldn't run right—didn't hold RPMs steady & after a while the armature just spun to a halt. A second, more through examination after the problem didn't go away found the B & C legs swapped. From the factory. If it were a three-phase motor, it would have run fine(ish), just not compress; and I would've found the actual fault immediately. The good news is that everything is labeled, you don't even need the service manual to find & fix it. BTW, all Mitsubishi manuals are online at mylinkdrive.com. (note: NOT my dot linkdrive dot com!)

JUGHNE

Ratio, yes, but one would test it as if it was a 3 phase motor.

Do you know of a good test procedure for the outdoor fan motor?

MaxMercy

Odds are the IPM was damaged. That schematic is actually just a block diagram and is showing very little detail - for instance, the power factor correction circuit is missing.

ratio

@JUGHNE, IIRC the test procedure is just ohm out the windings against the expected value (& check for grounds) & if that's good replace the whatever board. That should be pretty accurate, I don't believe there's much more than the windings inside, if they're good & there's no physical damage to e.g. bearings it's gotta be the board. The service manual may give expected voltages for the board, too. It's been a while & I was just skimming that section.

RedOak

ratio said:

I mention it because tech support & I condemned an inverter board because the compressor wouldn't run right—didn't hold RPMs steady & after a while the armature just spun to a halt. A second, more through examination after the problem didn't go away found the B & C legs swapped. From the factory. If it were a three-phase motor, it would have run fine(ish), just not compress; and I would've found the actual fault immediately. The good news is that everything is labeled, you don't even need the service manual to find & fix it. BTW, all Mitsubishi manuals are online at mylinkdrive.com. (note: NOT my dot linkdrive dot com!)

@ratio - Wow, so you had an experience in which you found an outdoor unit that came from the Mitsu factory wired improperly!?! That's a pretty revealing story ... Hmmmm, if we end up buying a new single-zone system, maybe we should take the last HVAC guy's advice and move to Lennox? Any thoughts?

ethicalpaul

Basing a brand decision on a single case of mis-wiring seems pretty rash.

RedOak

109A_5 said:

Hello, @RedOak,

The PTC64, 65 most likely provides a path to power the logic and communications electronics and maybe a soft start path to charge CB1, CB2 and CB3, since they and fuse F62 makes a path around X64. X64 most likely is only closed if a motor needs to run or after a power up delay.

It would not surprise me if there are MOV on the board too.

@109A_5 - As @MaxMercy wisely pointed out, the power supply schematic pasted into your last post is not much more than a "block diagram" (sans theory of operation), which forces anyone who's worked at the component level to make certain assumptions. I've actually seen MOVs that look a lot like PTC64 & PTC65, so that's why I keep asking questions about those two little "black boxes." They're also described on the drawing as "Circuit Protection," which adds to the guessing game. It's really too bad that they didn't simply use proper electrical symbols in the diagram, but there may be a business "philosophy" (aka, agenda) at work there, aye?

mattmia2

PTC is probably positive temperature coefficient themistor. It is probably potted in a molded case and the number on it probably tell you what it is. It probably isn't your problem although I could see a voltage spike shorting them so it tries to run the compressor through that leg of the circuit when the relay x64 is open. It isn't clear if this is an actual schematic and there are point to point connections not shown or a block diagram.

Jumpering a blown fuse is an extremely bad idea. A blown fuse is almost always a symptom not the problem and a jumper give that excess current an opportunity to destroy some other things beyond the initial failure.

It isn't clear what controls x64, it may be the type of relay with both solder and quick connect terminals and a logic board controls it through wires that plug directly in to it.

RedOak

ethicalpaul said:

Basing a brand decision on a single case of mis-wiring seems pretty rash.

There's also a pretty big price disparity in the mix here, as well. This entire thread is based on the fact that we're on a fixed income ... but that's all been covered in the thread, so I digress.

RedOak

mattmia2 said:

PTC is probably positive temperature coefficient themistor. It is probably potted in a molded case and the number on it probably tell you what it is. It probably isn't your problem although I could see a voltage spike shorting them so it tries to run the compressor through that leg of the circuit when the relay x64 is open. It isn't clear if this is an actual schematic and there are point to point connections not shown or a block diagram.

Jumpering a blown fuse is an extremely bad idea. A blown fuse is almost always a symptom not the problem and a jumper give that excess current an opportunity to destroy some other things beyond the initial failure.

It isn't clear what controls x64, it may be the type of relay with both solder and quick connect terminals and a logic board controls it through wires that plug directly in to it.

I agree with you about the risky jumper bypass, but it was a very warm evening and I didn't have a proper fuse available at the time. The Mrs had already made an appointment for the next morning, which hastened things as well. Regardless, I assure you that nothing - on the remarkably pristine inverter board - was fried by the temporary jumper, which adds to the frustration, quite frankly.

As I mentioned before, I will soon take the outdoor unit apart again and apply the advice I've received herein in an attempt to either isolate the actual issue or bite the proverbial "deductible bullet."

mattmia2

As I am sure you are aware if you have experience troubleshooting electronics, semiconductors rarely have any visible signs of failure when they fail.

RedOak

MaxMercy said:

Odds are the IPM was damaged. That schematic is actually just a block diagram and is showing very little detail - for instance, the power factor correction circuit is missing.

@MaxMercy - I gave the schematic posted in this thread another look and, not being an HVAC expert, I have to ask about the actual function of the IPM? I don't see an IGBT IC in the drawing, so what's it there for? The symbol placed on the IPM "box" in the drawing is an NPN transistor, so is it performing some sort of amplification? As you can tell, the lack of information is pretty frustrating.

By the way, do the acronyms "MC" and "MF" in the drawing represent Motor Compressor and Motor Fan? Scratch that, I see that I made a good guess. "MC" is the compressor and "MF" is the fan motor.

mattmia2

The IPM is some sort of a motor controller, the motor is variable speed and it is the variable speed drive. You will have to look at the board to see what component it actually is, if it is some sort of power IC or if it is a module on a separate PCB attached to this board.

Clearly there is another logic board that controls this board.

RedOak

mattmia2 said:

As I am sure you are aware if you have experience troubleshooting electronics, semiconductors rarely have any visible signs of failure when they fail.

Not in my career. I was once stationed in the "Lightning Capitol of North America," the Tampa Bay area, and I saw all manner of semiconductors that either showed brown lettering [excessive heat] or were simply blown apart ... but what does this have to do with an inverter board that hasn't even been properly taken out and looked at yet? Is it possible that I'm being goaded into some sort of an irrelevant argument? I'm simply too old for that sort of thing, so I'll bid you "good luck" finding someone else to quarrel with.

109A_5

Hello @RedOak,
I agree the drawing is basically a block diagram type drawing, it is often better than nothing. The days of good schematics and manuals with the theory of operation of every significant circuit are long gone for consumer electronics. If it dies just replace it. Manufactures and installers win, the consumer and the Earth looses.

I suspect the IPM is Inverter Power Module. That module may only be the high power switching devices, to make DC rail into 3 phase pulses 120 degrees apart and at a frequency that is needed dependent on the system load. Its is like a built in VFD (Variable Frequency Drive) like with bigger 3 phase motors. The actual control of the IPM may be a big (many pins) chip on the back side of the board or elsewhere on another board. Or it could do it all and the micro-controller just sends a message to IPM for the frequency (motor RPM) it desires.

BTW MOVs for over-Voltage protection would not be wired into the circuit that way. The circuit protection provided by the PTC devices is that it limits the current. However in the case of Lightning it is often amazingly fast with overwhelming dynamics.

ratio

@RedOak, we're still getting COVID units. Miswired, misprogrammed, missing pieces, extra pieces. Loose setscrews on blower wheels. Analog inputs plugged into digital outputs—backwards. Everyone is suffering from a lack of decently trained labor.

RedOak

ratio said:

@RedOak, we're still getting COVID units. Miswired, misprogrammed, missing pieces, extra pieces. Loose setscrews on blower wheels. Analog inputs plugged into digital outputs—backwards. Everyone is suffering from a lack of decently trained labor.

@ratio - That's pretty logical. In fact, it's inspiring me to continue looking into a repair of the outdoor condenser unit. At this point, I'm really hoping that the compressor tests out positively.

ratio

Check the winding resistance (again, IIRC the expected value is in the service manual on mylinkdrive.com; but if they agree to within ±0.2 chances are the windings are good), check for grounded windings. You may see high resistance to ground due to the refrigerant and oil allowing some leakage if the compressor hasn't run in a while, maybe slowly climbing, if you're using a megger/insulation tester.

The only other thing I've heard of, but not experienced myself, is that if the compressor overheats you run the risk of the rare-earth magnets in the rotor loosing their magnetism. Kinda hosed at that point, and I'm unaware of any way to test that without a working inverter board. I think it needs to be really hot for that to happen, though.

MaxMercy

RedOak said:

@MaxMercy - I gave the schematic posted in this thread another look and, not being an HVAC expert, I have to ask about the actual function of the IPM?

IPM is Intelligent Power Module, a general description of a self contained (and generally encapsulated) circuit that includes a microprocessor, driver circuitry, output devices (like IGBTs or MosFets), and a built in heatsink.

Sometimes they aren't fully potted, or they're potted in an easily removable clear jelly snot, or potted in a hard epoxy shell.

If the main board the IPM is mounted to has some printed function designations, you can generally measure the resistance of the output pin(s) to the ground pin(s). Any shorts means the IPM is bad. If there is no short on the output, the module's processing circuitry could still be bad and would be very difficult to check without a schematic of the IPM (virtually never published).

HVAC guys don't get into this level of servicing as the module is field replaceable. I'm an electronic designer and technician so I've been dealing with IPMs for many years now.

EDIT: oops, didn't answer question. In this case, the IPM drives the compressor. Modern AC units don't just show AC across the compressor, they're dynamically driven at various speeds for reasons of efficiency (energy star) and supposed longevity.

RedOak

MaxMercy said:

EDIT: oops, didn't answer question. In this case, the IPM drives the compressor. Modern AC units don't just show AC across the compressor, they're dynamically driven at various speeds for reasons of efficiency (energy star) and supposed longevity.

Given it's placement in the "block diagram" - pasted into this thread (above) - I guessed that it was feeding signals to the compressor (MC in the diagram). Needless to say, this changes things quite a bit. Considering that the system was running at the time, the IPM "compressor driver" could very well be damaged. Hmmm...

Do you think that a replacement for the IPM could be found anywhere?

ratio

It appears to be part of the main board? At least, it's inside the dashed line. It's some kind of transistor H-bridge module to drive the compressor with positive & negative pulses if the notations on the drawling can be believed. I can't guess if it's an off-the-shelf part or a Mitsubishi custom job. Looking into IC932 right below it might give you some insights, it's basically the same thing but for the outdoor fan. I would expect similarities between them.

IIRC most/all Mitsubishi mini splits have some LEDs on board. Are any of them lit up? They might even be flashing an error code.