Mitsubishi Noise Filter Board Fuse Blows

JackW

I have a four-unit Mitsubishi MX4B36NA Mini Split system. I turned it on for the first time today; it ran for a while, then quit. I pulled the cover off the compressor and found the 6.3A 250V fuse on the noise filter board had blown; it was black on the inside. I replaced the fuse with the same rating and turned just one of the inside units on. It ran fine for about half an hour and then quit. I opened the compressor to find the fuse was completely shattered this time (see pictures). Any help with what is causing this will be greatly appreciated.

Jack

pecmsg

Time to call tech support or your service provider!

JackW

Do you know what is causing this?

109A_5

Tech support will probably just keep changing boards and stuff until the problem goes away unless they know the answer from previous failures. If you haven't already, Google search that failure, Make, Model someone may have posted a fix elsewhere. I'm not familiar with that system so the following is ambiguous, just general troubleshooting. There is no overall picture of that board and its connections to it so I can't see any other obvious (to me) problems on that board.

There is a few different strategies or techniques to troubleshoot that situation. One visual inspection, chafed wires, corrosion, other scorch marks, etc. Another is divide and conqueror. Another is to replace the fuse with an appropriate incandescent lamp or lamps (depending on the needed voltage and current, maybe two or three100 watt lamps in parallel assuming it is a 120 VAC feed) instead of constantly blowing more fuses. Since apparently it does not blow fuses instantly, power off Ohmmeter checks may be useless. It appears something over time is avalanching or the system eventually tries to turn on something that draws way too much current like a seized up motor, all blower motors spin freely? The way those fuses blew it is a huge overload. Boards or sections (if any) that are powered separately are probably not the problem.

With the incandescent lamp in series with the input power to that board or being substituted for the fuse I would power up just that board only (if possible) and see what happens. If the lamp eventually lights up brightly the problem is on that board. If not reconnect other parts of the system one section or board at a time and see what lights up the lamp(s) brightly. One problem with this technique is if that board supplies standby power to the rest of the system then on a call some part of the rest of the system may signal to turn on additional circuitry on that board making the troubleshooting logic more tricky and more specialized testing would be needed.

If you are not skilled with working on equipment with lethal voltages call a Tech, it is not worth your life, don't work alone.

JackW

Thank you, 109A-5, for your help. I'll check all the connections, I can do that, I don't know how to do the incandescent lamp test. The mainboard troubleshooting lights are active. The red bulb flashes continuously very fast, the yellow bulb flashes13 times (very fast), then off for about 12 seconds, and then repeats this cycle. A very nice troubleshooting guide on the web uses the lights to diagnose the problem, but it doesn't address these bulbs' flashing. I'll have to give Mitsubishi service a call to see what they have to say. Thanks again for your input; greatly appreciated.

109A_5

JackW, if you have diagnostic indicators my all means see where that leads. I have found Tech Support can be amazing or useless, you never know. Thinking more about the current rating of that fuse there maybe a shorted mag valve solenoid coil or a relay coil causing the overload rather than a bad motor load.

The incandescent lamp test technique is quite simple once you understand it, however it is slowly getting lost to time and politics. It is a very old technique. Basically you are leveraging the properties of a Tungsten filament in series with the load. It will heat up faster than a fuse will blow, providing a visual indication of an overload and it acts as a current limiter also so you don't blow any more fuses during troubleshooting. Cold filament equals low resistance, hot filament equals much higher resistance, limiting the current. Using a circuit breaker type device in place of the fuse may stress part of the circuit that may cause other problems or damage, also it leaves the circuit dead so power on troubleshooting is impossible. The hardest part is determining the proper wattage lamps to use and wiring the lamp sockets (if needed). In this case the total current drawn by the lamps in normal use would maybe be 3.15 Amps or half the fuse rating as a starting point. The lamp(s) should be the same voltage as the applied circuit voltage, 12 for Automotive, 24, 120, 240 for HVAC/R applications. So using 120 Volt circuit as an example 120 x 3.15 = 378 Watts, so three 100 watt lamps in parallel may work fine. Other combinations will work too.

I have found chafed wires in seconds using this method when other techs were just scratching their heads not knowing what to do next. Sometimes the setup takes longer than the discovery of the problem.

I have never tried it with a motor load, that may be interesting.

pedmec

looked at the wiring diagram and the fuse that keeps blowing is the power/communication feed to the indoor heads. i would disconnect the heads electrically and reconnect one at a time to see which head might be causing it. its a starting point. its the f65 fuse.

one time had the liquidtight for the heads buried underground due to the landscaping. eventually water penetrated the liquidtight and caused a short.

ratio

That fuse was hit by a solid short, not an overload. You said it's on the line to the indoor head, look for a pinched/chafed wire that is occasionally shorting out to ground or the other conductor. Remember that Mitsubishis communicate over line voltage, so even S3 is suspect.

the last time I ran across a similar problem, a megger (Fluke 287) found a weak spot in the insulation that had been hiding from the service dept. for several no-cool calls.