Hi all,

Some years ago I installed a small solar system for a friend. At the time I used a Chinese 1500VA inverter which has worked fine for about 3 years... Until recently.

In their attempt to power the house during an extended cloudy period they somehow managed to connect their diesel generator to the inverter's 230V output socket.

When the inverter is powered since then it will blink the red fault LED with 0 output.

The inverter is unbranded. It uses 4 x K30H603 high speed switchings and 12 x IRF2807 power MOSFETs.

Could someone guess as of what could have burst in such a misuse incident?

Thank you in advance,
Pete

 

Most likely protection devices (fuses, breakers and maybe NTC thermistors). Suspect all of the semiconductors in the output stage.

 

Thanks a lot for that!
I see no fuses or breakers on the output but there is a thermistor on the MOSFET cooling element. Could this have blown due to overheating? I thought these where supposed to sense temperature.

 

there is a thermistor

I just desoldered it. It is a 10K/25 and reads 8k at 22°C.

 

Whether that is a sign of a problem depends on where it is in the circuit. Can you tell where it is in the circuit or what its function is?

 

Thanks for the guidance.
After your suggestion to check the semiconductors, I found 3 of the 4 K30H603 to be sorted. I have now desoldered these but I cannot tell whether the damage has stopped there or not. The problem is that I cannot find these locally. I will order them online but for the month it will take to get them delivered I will not be able to check for any more damage. So do you think it is ok to temporarily solder 4 smaller fast switching transistors in order to check the inverter's function in the meantime? Should that work?

 

(Some text removed for clarity). So do you think it is ok to temporarily solder 4 smaller fast switching transistors in order to check the inverter's function in the meantime? Should that work?

It sounds risky but if you can spare the transistors why not try it? Probably best to not have a load on the output when you try this.

 

So if the K30H603s are rated 600V-30A can you guess what small one I could probably use for testing, given that I will not load the output?

 

I think the more important question is what transistors you can easily get your hands on that might work.

 

Thank you so much for your input so far. It seams like I am getting somewhere.

So...
It took a while to find proper replacements for the output transistors but I have now replaced all four K30H603s with new IXGH32N60BD1. To my understanding these are compatible (please correct me if I am wrong).
I now get a completely different behavior from it.
Once turned on it beeps continuously, fault LED flashes and it shuts down (built in protection maybe). Befor it shuts down it gives some kind of output that my multimeter does not have the speed to read (I do not own an oscilloscope). After a few seconds the beep stops, fault LED continues to flash and every 30 or so seconds this circle repeats itself (probably the built in protection rechecks). I am attaching a video of the behaviour with a mains desklamp connected on the output. I am afraid I had to zip the video otherwise it would not let me upload it. I hope that's ok.

Maybe it is useful to say that measured resistance both on output and input terminals are as expected, during the brief on state the input draws normal current (less than an Amp) and when it shuts down I get 0 voltage also in the input side of the board (just after the big capacitor).

Any thoughts?

 

Here's a link to that video too, I think this is a better way:

https://drive.google.com/file/d/1OhstwtZXw3r3ArnoVZqZuhbHbHIP6R5z/view?usp=drivesdk

 

Strange. It seems to want to work. Do you have a manual for the inverter. It might have been buzzing out a failure code.
"1,9,11,8"

I have not used IGBTs myself but the replacement looks compatible, as far as I can tell.

 

Unfortunately the beeping sequence does not match any of the documented error sequences.
Just managed to measure output voltage of 280v instead of 230v during the 1 or 2 seconds that the bulb tries to stay on. Might be even more as it is too quick for the multimeter to read reliably but 280v or more is true for this short time. Do you think it is excessive or is it normal to start off that high?

 

No, 20% overshoot would be unacceptable except in the most poorly specified products.

You are now in the most difficult of troubleshooting phases in that it almost works. Maybe it is time for you to search out expertise in this particular inverter to save some time.

If you are interested in a more accurate reading of the RMS voltage when it first comes on you can capture the peak voltage in the capacitor C using the circuit below and multiply the voltage by 0.070 to find the RMS value. Use a 450 volt capacitor and 1N4007, UF4007 or other diode with a PIV of greater 700 volts or greater.

 

Well... Chinese generic inverter, so no support that I can find.
Anyway, over time somehow the voltage goes to 60v at startup. This lasts for about 2 seconds so I think it is a good measurement off my multimeter. I have no idea what changed... Other than the voltage the behaviour is as before.