Takasago KX-210 power supply

[Asuka]

I promised to post this before the end of January, so here goes. Please see this thread for background.

I'm trying to repair a Takasago KX-210 power supply which I bought as broken last month. I have pulled the power board out of the device, and I am working my way from the power input through the device. So far I have checked the mains power connection, the power switch, the mains rectifier, the filter capacitors, the mosfets, and the transformer. The capacitors all seem to be OK, but when I got to the mosfets (Toshiba K2953), it seems that one of the mosfets is acting as a diode between the gate and the drain. Can anyone confirm whether this is possible? When I put the black terminal of the multimeter on the drain and the red terminal on the gate, and turn on the diode tester, one of the mosfets reads .575 and the other one reads nothing, even though they are identical units. I also checked the transformer for continuity etc. and I cannot see anything wrong with it.

I was going to post full resolution photos but I couldn't sign up for an imgur account for some reason today, I'll try again later.

[Asuka]

Here is an image annotated with the various legs, since it might not be clear what is going in where. The on/off switch is a mechanical switch and this board is two-sided, so some of the tracks are on the top and some are on the bottom. I've checked that the switch is functioning.

[Armadillo]

In situ measurement, particularly without a schematic, cannot be taken as conclusive because you may be measuring charging capacitor or other nearby resistor elements.

I don't know how confident and safe you are in measuring the circuit while powering up? But you do at own risks, you know the voltage is high.

The isolation transformer [not absolute] depending your level.

Absolute must: Shock Protecting Gloves, a pair of protection glasses, 1X60W incandescent light bulb in series method properly socketed, and safety first.

If I am not mistaken, the symptoms, when you turn power on, No Sound, No Picture [no display of leds, lights etc], nothing.

Firstly, Clear by troubleshooting the power input section, switch, voltage select switch, until the DC bus section by measuring the DC capacitor voltage when you applied power through the light bulb in series method.

If you find it difficult to understand me and wish to troubleshoot without power on, please wait for others to advice to you, OK! Cheers;

Please take photo vertically perpendicular, so that the traces and component layout are obvious and easy to see.

******* REMEMBER TO DISCHARGE THE CAPS ***************

[Electro Detective]

Check out these three PSU videos by Mr. Carlsons Lab     the gent certainly knows his stuff   

youtube.com/watch?v=pffOJdCQ7kw

youtube.com/watch?v=qsrMkde23s0

youtube.com/watch?v=mwl74jk8ZyU

[coromonadalix]

Remove the mosfets and test them outside circuit,  they dont have any diode in them, there are tutos on the web to help you test them ... testing them in diode test  will arm the mosfet, IE it will act as a switch if i may say

[Asuka]

The isolation transformer [not absolute] depending your level.

I'm still looking for an isolation transformer. After I got the advice in a previous thread, I bought a 600 VA 100V to 24V transformer at a bargain shop to rewire it as isolation transformer, but it seems a shame to destroy the transformer. I'm not sure whether it's a good idea to try running the 24V into the system from the current transformer I have, or whether that might do some harm.

If I am not mistaken, the symptoms, when you turn power on, No Sound, No Picture [no display of leds, lights etc], nothing.

Yes, it is doing nothing but it is drawing a small current from the mains.

[Armadillo]

to rewire it as isolation transformer, ....., or whether that might do some harm.

Is my opinion that you don't add complexity to your repair now. You cannot doubt your own foundations and your own tools during repair and that will include the isolation transformer.
Transformer winding is an Art. It sounds easy but is not. You will have to de-laminate the plates one by one. You don't want to build it and doubt it because there will be ultimately big voltage drop even at 1A. You may have broken the insulation and instead of safety, it becomes more dangerous to you. The existing bobbin may not catered for the extra windings. And how about humming or buzzing sounds after build.?
You may be able to do it, but just not now.

I suggest you return it and get a proper 100V to 110/120V 600VA~1000VA isolation transformer properly fused in and out and socketed and metered.   
I'll show you my build.

[Asuka]

to rewire it as isolation transformer, ....., or whether that might do some harm.

Is my opinion that you don't add complexity to your repair now. You cannot doubt your own foundations and your own tools during repair and that will include the isolation transformer.
Transformer winding is an Art. It sounds easy but is not. You will have to de-laminate the plates one by one. You don't want to build it and doubt it because there will be ultimately big voltage drop even at 1A. You may have broken the insulation and instead of safety, it becomes more dangerous to you. The existing bobbin may not catered for the extra windings. And how about humming or buzzing sounds after build.?
You may be able to do it, but just not now.

I think this transformer can be rewound relatively easily by removing the outer layer and replacing it with thinner enamel wire. I've attached photos so you can confirm. The 24V output is the outside layer of very thick wire. However, it seems a shame to break this. Perhaps I will need 24 V AC for something, someday.

I suggest you return it

I cannot return it but I only paid 500 yen (about $5 US) for it so it doesn't matter.

and get a proper 100V to 110/120V 600VA~1000VA isolation transformer properly fused in and out and socketed and metered.   

I'm looking for one at the moment. I promised to start this thread by the end of January so I had to start without the transformer (also without earth wiring, another story).

[Armadillo]

I am surprised that Japan is selling used instrument at very very reasonable price. $5 is a give. I am amazed. Maybe will look at Japan for these used parts.

It don't look any easy to me. But at $5,  you can try and pick up the fun out of it as another interesting project. 

[Asuka]

I am surprised that Japan is selling used instrument at very very reasonable price. $5 is a give. I am amazed. Maybe will look at Japan for these used parts.

This is from "Hard Off", www.hardoff.co.jp, as you can see in the photograph. It's from the "Junk" section which means no return, no guarantee. The particular store I bought this from has tons of rack-mount video equipment from the 1980s / 1990s selling for 500 yen or 1000 yen, for example.

It don't look any easy to me. But at $5,  you can try and pick up the fun out of it as another interesting project. 

I don't think it's necessary to unscrew the plates, just cut the old wire and thread the new wire through the hole. However, as I said, I don't want to destroy this so I am going to continue looking for a better option.

[glicos]

Based on the datasheet of the MOSFET, there's a diode connected across the drain and source (reason why you have a diode response one way of your DMM). Need to touch your + probe to gate to turn ON the MOSFET. You should have a reading (both ways across the Drain and Source). Then touch gate and drain with your finger to discharge and turn OFF your MOSFET. You should have a diode reading one way as your previous test...
But i suggest test MOSFET using a DC power supply and some resistors to test it properly.

[Armadillo]

This is from "Hard Off", www.hardoff.co.jp,

It ask for "Prefecture". Searching through the whole list of Prefectures... I cannot find another KX-210. LOL...  . You got the one and only. Lucky you. ebay USA is at it at crazy price.

I don't think it's necessary to unscrew the plates, just cut the old wire and thread the new wire through the hole. However, as I said, I don't want to destroy this so I am going to continue looking for a better option.

I can tell you, it's not easy, fortunately you are looking for better options.   

[Asuka]

Can you post a close up photo of the area surrounding your MOSFET? I want to see how the gates are driven.
Normally, a blown MOSFET shows dead short across all 3 terminals, and upon power up, spectacular magic smoke releases.

Maybe they are not blown then. I have attached pictures. It was hard to get these photos because my camera doesn't focus very well on small items. The scratches on the reverse are the result of my clumsy use of the multimeter probe. The "fault" I am seeing is that there is a measured "diode" from gate to drain on the left side mosfet. However, I have not removed them from the circuit yet so it might be the result of the other components in the circuit.

[Asuka]

This is from "Hard Off", www.hardoff.co.jp,

It ask for "Prefecture". Searching through the whole list of Prefectures... I cannot find another KX-210. LOL...  .

Ah, sorry, I sent you on a wild goose chase. You cannot find these kinds of things on their web page. You have to go to the physical store and then go to the "junk" section. There are two sections of the shop, the main section where they sell second-hand goods which they offer a guarantee on, and the "junk" section where there are just things thrown into boxes or piled on shelves and you have to search through them for bargains. The guaranteed items are often almost as expensive as new ones, and frankly I would just buy a new one instead. If there is anything on the web page it will be the guaranteed items. The junk section is randomly priced, usually very cheap, but sometimes more expensive than a new item. It's not that rare to find some item from Daiso or an item which costs less than 100 yen new sold for 100 yen. They really do not care about the items, they won't go as far as cataloguing them, and you cannot take it back to the shop if you find it doesn't work. There is also a place in the shop with a power outlet where you can try turning the item on.

You got the one and only. Lucky you. ebay USA is at it at crazy price.

The original price of the item is fairly high, but you need to go there and look. The same shop I bought the power supply at has a tektronix 2467b oscilloscope on sale for 30,000 yen, but again it's in the junk section. You won't find it in the hard off web site.

I don't think it's necessary to unscrew the plates, just cut the old wire and thread the new wire through the hole. However, as I said, I don't want to destroy this so I am going to continue looking for a better option.

I can tell you, it's not easy, fortunately you are looking for better options.   

It's a bit tempting to try to prove you wrong but I will end up regretting it on the day I need 25 amps of 24V AC.

[Armadillo]

It's a bit tempting to try to prove you wrong but I will end up regretting it on the day I need 25 amps of 24V AC.

At $5, I think you only need to prove to yourself, on your original intention to modify the transformer when you bought it in the first place. It don't just land itself in the junk section for nothing, I suspect.
I have seen many,    theory-vision-intention, theory-vision-intention and that's it, no Execution but only revise reasoning for excusing initial buying action.
No worry, you are not the only one. Things will pile up eventually.
No action, nothing learn. That's what I want to say. Positive thinking OK. thanks.   

[Asuka]

At $5, I think you only need to prove to yourself, on your original intention to modify the transformer when you bought it in the first place. It don't just land itself in the junk section for nothing, I suspect.

As far as I know it works correctly, here is a photo you can check using a 100 W incandescent bulb. I don't have a proper insulation resistance tester so I haven't tested it beyond what's possible with a multimeter, but there are no obvious big problems.

I have seen many,    theory-vision-intention, theory-vision-intention and that's it, no Execution but only revise reasoning for excusing initial buying action.

You might be right that I am justifying after the fact, but as I said it seems to be in OK shape. If I really cannot find an isolation transformer, maybe I will modify this one. As far as the $5 goes, the thing must have enough copper in it that the scrap metal value is beyond that. This transformer weighs 11 kilograms. Look at this web site, they offer 715 yen for one kilogram of enamel wire:

https://www.ohata.org/pika.html

No worry, you are not the only one. Things will pile up eventually.

Actually since I started learning electronics, I got some confidence to fix quite a few other broken things, so stuff is actually starting to get fixed, not pile up.

No action, nothing learn. That's what I want to say. Positive thinking OK. thanks.   

Thanks for your inputs. Next job is to start on a schematic for the power supply.

[Asuka]

Based on the datasheet of the MOSFET, there's a diode connected across the drain and source (reason why you have a diode response one way of your DMM).

The diode response was from the gate to the drain.

Need to touch your + probe to gate to turn ON the MOSFET. You should have a reading (both ways across the Drain and Source). Then touch gate and drain with your finger to discharge and turn OFF your MOSFET. You should have a diode reading one way as your previous test...
But i suggest test MOSFET using a DC power supply and some resistors to test it properly.

I think I should draw a schematic of the power board to try to make sense of what it's doing before any more tests.

[Armadillo]

By the look of it, only the high side mosfets are transformer driven.

I doubt even the isolated auxiliary low voltage power supply is available.

Surprisingly there are no hazard warning logo anywhere, no clear delineation between cold and hot sides, neither.
I would say don't even trust resting your hands onto metal heatsinks even though I can see thermal pads.

Only when you are sure of what you are doing and ready, no rush into things. 

[Armadillo]

Isn't that weird? Pri-sec leakage inductance will introduce delay, hence if not all 2 switches in the half bridge are driven by the same xfmr, you will have shoot through or you need very high dead time.

I suppose ultimately there must be some form of feedback current winding/signal, which should be converted to voltage ramp to the controller which will use it to terminate conduction and properly timed and insert necessary dead time regardless of the psychical tolerances of the external devices [to limits] and prevailing circuit loading.

We are given a small "peek hole" to the whole of the circuit otherwise, i would want to think that it is full-bridge, so I am all guessing. 

[Asuka]

So far I have got to the following schematic. There is some error in my understanding of the 100V / 200V switch since I have the 120K ohm discharge resistors for the capacitors splitting the 100 V in this diagram, which obviously doesn't make sense. The tracks on the bottom are labelled A, C, D, etc and the top tracks are alpha, beta, gamma, etc.

I'll keep working on this and post a better version later. This does not include the control circuits yet.

[Asuka]

Can you double check if anode of D29 is indeed connected to drain of the left MOSFET? It looks weird to me as an active clamp forward converter.

Yes, you are right, thank you for correcting me. I have also checked D22, which was correct, and added the missing beta track which is carrying the 100V power into the rectifier.

What is the function of RL1 in this circuit? I found the data for this component:

https://www3.panasonic.biz/ac/e_download/control/relay/power/catalog/mech_eng_dk.pdf

[Asuka]

Today I started on the sub-power board. This has leads coming in from the rectified unfiltered high voltage DC from the main power board I showed last time, which come in through the black and white twisted leads at the bottom of the photo of the board, and then it has its own switching power supplies on this board, which go to, for example, the fan, which runs on 24V DC and was not working when the unit was switched on initially.

I spent some time tracing out the high power leads on the reverse of the board, and I noticed an odd phenomenon with this board where there was a quite big track leading to the large transformer which looks like it should be carrying power to or from the transformer, yet the bottom lead of the transformer, circled in pink on the attached photograph, has no continuity with any of the other transformer leads.

After looking at the transformer, one thin piece of enamel wire on the leads has snapped. You can see in the photograph below. Although it is difficult to see in this photograph, the backlit enamel wire which seems to have snapped is the same type of wire as the wire on the right of the transformer, and then the thicker enamel wire on the left side of the transformer also has a pair opposite the apparently broken wire which is just visible in the photograph coming round the plastic box, so it's not connected to that wire.

It seems likely that the wire was broken in a shock of some kind, or perhaps overheated, and the way to fix this is to desolder the transformer then put a jumper over the broken wire.

(Unfortunately, after a lot of effort, this was the best photograph I could take of the broken wire, since my camera will not focus on small items.)

[Armadillo]

You will need to have the auxiliary power supply working, otherwise the unit will appear dead. I think you should concentrate on here as the priority.
The big cap looks like it bulk on the top [maybe camera effect]. I suggest you take it out and measure the ESR. Don't measure in circuit. I would say never for suspected cap.
If you suspect the transformer, then take it out of the board to inspect it closer and measure its primary and secondary inductance as well as continuity test. I have never seen transformer wire "burn" like this. 

[Asuka]

You will need to have the auxiliary power supply working, otherwise the unit will appear dead. I think you should concentrate on here as the priority.
The big cap looks like it bulk on the top [maybe camera effect]. I suggest you take it out and measure the ESR. Don't measure in circuit. I would say never for suspected cap.

Oops, sorry, I thought someone might assume that when I posted that. I should have pointed out that that was an effect in the photograph. Please see the attachment.

If you suspect the transformer, then take it out of the board to inspect it closer and measure its primary and secondary inductance as well as continuity test. I have never seen transformer wire "burn" like this. 

Sorry, it is not "burnt", the look of the photo was due to the circumstances of taking the photo. The previous photograph was backlit using a light bulb behind the circuit board because it was so hard to get a photo of the wire.  (I might need a camera lens for photographing very small objects if I keep posting here.) The attached photo in this post is the very best one I could get after quite a lot of fiddling.  The broken wire is the one on the right. The fact that there is no continuity from that giant track into the transformer to anywhere else in the circuit board (see the pink circle in the previous photograph) seems very suspicious too.

[Armadillo]

Can't tell from the picture if it's visually broken.
You sure you didn't measure on the enamel of the magnet wire, yeah?

[Asuka]

Can't tell from the picture if it's visually broken.

No, you are exactly right, you can't tell from that photograph. That's why I shone a light bulb through the circuit board for this photograph:

https://www.eevblog.com/forum/repair/takasago-kx-210-power-supply/?action=dlattach;attach=392503;image

The point is, I don't have a camera which is good enough to take a photograph showing what I can see with my eyes, which is that the enamel wire is broken.

You sure you didn't measure on the enamel of the magnet wire, yeah?

Yes, I am very sure, because I was able to measure continuity from the transformer lead which I indicated with the pink circle on the back side of the board to the plus lead of the large capacitor, but not to any other leads on the transformer or any other parts of the circuit board except for that large power track which goes from the plus lead of the capacitor towards the transformer. That was the measurement which I thought was very odd, which is why I had a very good look at the transformer and found this. As for measuring on the enamel wire itself, I'd like to see someone try it.

Next stop is to desolder the transformer, and try to fix that tiny little wire.

[Armadillo]

You know, for example switching transformer EMI screen is a screen [copper plate or wire] with 1 lead, normally connected to ground and nothing else.
But you can take it out and investigate.

Just don't investigate "too long" because these wires maybe fragile. 

[Asuka]

You know, for example switching transformer EMI screen is a screen [copper plate or wire] with 1 lead, normally connected to ground and nothing else.

You are absolutely correct again. Please see the attached photographs.

But you can take it out and investigate.


Just don't investigate "too long" because these wires maybe fragile. 

You are always very helpful and concerned about everything, I feel glad to get your attention.

[Asuka]

I was able to remove the transformer from the board after quite a lot of effort with soldering wick. The through holes seem to be able to swallow a lot of solder and it was rather hard to get them out. I have attached two pictures of the broken wire. It is about 0.2 mm in diameter I think. At first I was planning to repair it using heat sink tube and 0.8mm diameter solid core wire from a doorbell, which you can see in the third photograph. That doorbell wire was just too thick to solder on. In the end I used 0.35 mm enamel wire which you can see in the fourth photograph. Wrapping the other end of the enamel wire around the transformer's leg was a rather delicate job but fortunately the wire and the solder joint survived that. I was able to resolder the transformer to the circuit board and I have confirmed continuity of the connections, as well as non-continuity of the things which aren't meant to be connected. The newly-soldered wire has a resistance of 9 ohms, which seems like a lot, but it is extremely thin wire. Unfortunately removing the solder with the wick left a lot of mess on the back side of the board, so I'm going to clean this up and then reassemble the power supply tomorrow. I don't have anything suitable for cleaning the board at the moment.

[Armadillo]

  Well done, you found the fault.
With the transformer removed, can see the broken wire clearly.

[Asuka]

Well done, you found the fault.
With the transformer removed, can see the broken wire clearly.

Thanks very much for your help.

If it starts working with this fixed transformer, I might write to the manufacturer and ask for a replacement for the transformer just to be on the safe side. I haven't replaced it back yet, since it's been a long day. I'll post an update when it's all back together.

[Asuka]

I put the power supply back together today and tried powering it up. Unfortunately the unit did not operate. I noticed the following symptoms: LD1, the pink circle on the left in this picture, lit up, and HS3, the heat sink behind the FET on the right, also circled in pink, got extremely hot, so that a smell was noticeable.

It's possible that the transformer failed again due to my repair breaking, or that I should have cleaned the back of the sub-board more carefully, or that I damaged something putting it all back together, or I failed to connect something properly, or there is another problem in the item somewhere.

[Armadillo]

Did you manage to get a current draw this time? or the brightness of the lightbulb?

If its hot and smell, something is shorted. This is a easier repair than other types.
I would dismantle the connectors to all the outputs [power input is needed].
Check all the mosfets, diodes and caps.

The next time you want to run it, make sure the lightbulb in series method is used. When the light is bright, turn it off immediately.

[Asuka]

Did you manage to get a current draw this time? or the brightness of the lightbulb?

I didn't have a light bulb in series, so I've spent some time rigging up an isolation transformer and current limiter.

If its hot and smell, something is shorted. This is a easier repair than other types.
I would dismantle the connectors to all the outputs [power input is needed].
Check all the mosfets, diodes and caps.

Yes, I will dismantle the unit, then clean it up, then check the continuity.

The next time you want to run it, make sure the lightbulb in series method is used. When the light is bright, turn it off immediately.

Got it. I have rigged up the light bulb as shown in the picture. Isolation transformer also made from back-to-back audio amplifier transformers.

[Armadillo]

Ehmmmmmm....... You would need to know the KVA and the voltage drop. I don't know but I assume you did your maths.

[Asuka]

Ehmmmmmm....... You would need to know the KVA and the voltage drop. I don't know but I assume you did your maths.

I did the maths of matching the transformers and checking the resistance of the coils, and I tested it out with a 12V AC power supply.

If I have a 100 W light bulb in series with the transformers, the most amperes that could possibly flow is 1 amp, and these are big audio amplifier transformers which take mains inputs. Both weigh about two kilograms with outputs around 50 volts, so I don't think there is any problem.  The output of the transformers is exactly the input voltage, 100V, if the bypass is switched on, and reduces to 95V when the light bulb is in series. I think it's adequate. I've also put a 3 amp fuse in between the transformers and the mains input.

[Armadillo]

So, if its supplying 1A then theoretically zero volts is left for the DUT???
What is the Watts rating of the KX-210?

Edit: But if your objective is just to test for shorts and bypass thereafter, I suppose it should work for you.   

[Asuka]

So, if its supplying 1A then theoretically zero volts is left for the DUT???

I am surprised that the same person who suggested using this is asking this question. The bulb acts as a 100 ohm resistor in series, so the maximum current which may flow is 100V / (100 ohms + load), which is always less than one amp. It only goes to 1 amp through the light bulb if there is a short circuit.

My design is based on the one in this video:

What is the Watts rating of the KX-210?

Maximum input is 5.5A depending on the load.

http://www.takasago-ss.co.jp/english/products/power_electronics/dc/kx/kx.html

That would require a 600 W light bulb in series. I'm not planning to run this under load anyway.

[Armadillo]

I am surprised that the same person who suggested using this is asking this question.

Don't be surprised. We are just trying to guide.   

The important thing is the Operating Voltage of the Takasago. What is the minimum voltage that it can operate so that you can check the circuit.
For example 0.5A X 100 ohms = 50V, so 50V remains for the Takasago. Will the transistor switch at 50V?
And how about at other starting or operating currents?
But always, never try never learn, aside from that the transformer will also drop some voltage other than the bulb.
I am not good at 100V supply. 

Edit: You can always parallel another 100W incandescent bulb if not enough. 
But i am not sure of your audio power transformer back to back configuration especially effect on voltage drop.

[Asuka]

But i am not sure of your audio power transformer back to back configuration especially effect on voltage drop.

The output voltage is 100 volts, identical to input voltage. I have measured it with a multimeter. Anyway, I ran it for a half an hour with a 100 W light bulb and the correct-way-round transformer didn't even get warm. The wrong-way-round transformer got a little bit warm after half an hour. I don't think it's going to be any kind of problem.

Today I had some time and took the thing apart again. I cleaned up the mess I made last time, and started measuring continuity. There is continuity right across the gate and source of the mosfet which was getting so hot before. The part number of this MOSFET is 2SK1808, which seems to be this:

https://www.renesas.com/en-eu/products/power-mosfets/mosfets-for-switching/device/2SK1808.html

(link to English-language page). The gate of the mosfet is only connected with a 200k-ohm and 22kohm smd resistor,


so I don't see how it can be short-circuiting like that. Perhaps I misunderstood something. According to the data sheet there is a function like two zener diodes between gate and source, but I just get continuity with a voltage difference of less than one volt. I also measured the resistance, which was much less than one ohm. The package doesn't seem to be physically damaged at all.

[Armadillo]

You are right to use ohms to measure shorts. if it is less than one ohm both ways [red and black probes reverse] it is then a definite shorts.
This warrant you to take out [desolder] the mosfet and confirm it externally that it is dead sure shorts before you order things around.
If you are able to take out the transformer, I am sure you can take out the mosfet. Just remember to retain any mica insulating sheet or the plastic screw sleeve.

Is this mosfet part of the auxiliary power supply? Seems rather huge?   

[Asuka]

You are right to use ohms to measure shorts. if it is less than one ohm both ways [red and black probes reverse] it is then a definite shorts.

It's about 0.5 ohms on this multimeter, but it's not an EEVBlog multimeter but just RadioShack.

This warrant you to take out [desolder] the mosfet and confirm it externally that it is dead sure shorts before you order things around.
If you are able to take out the transformer, I am sure you can take out the mosfet. Just remember to retain any mica insulating sheet or the plastic screw sleeve.

Hopefully the mosfet will be easier than the transformer.

Is this mosfet part of the auxiliary power supply? Seems rather huge?   

It's the same one as in the picture before, where I circled the heatsink in pink.

If the mosfet is faulty, it seems like I need to find a new replacement since this is no longer in production.

Putting V_DSS, V_GSS, I_D into Mouser I find STF4N90K5 from STMicroelectronics and FQPF4N90C from ON Semiconductor, and a few similar things at DigiKey. Is that a good way to choose a replacement, I wonder?

[Armadillo]

I would agree with the STF4N90K5 in TO-220FP package.

I suggest you remove the mosfet and measure again to confirm no other shorts.
From any DC capacitor after the rectifier, you measure ohms to make sure that the capacitor charges up to high enough ohms and not shorted or low ohms. If there are still shorts, the DC capacitors will show a short, meaning cannot turn power on. But still cursory measure each mosfet to be double sure, no more shorts.

Have you already study the auxiliary power supply at least it can output the voltage at the DC capacitor after the regulator or rectifier of the auxiliary supply?
You can isolate the main switching side by taking out some fuse or connector. Just test the auxiliary power supply for function.
Now you have the isolation transformer ready, you can use it to test it.
Be safe and take your time.
I always will forget to unplug and discharge capacitor after test. I think I need to tie a red string on my thumb for reminder.   
Wear a rubber glove and a safety goggle to add confident level.