Getting started with troubleshooting [Samsung 245BW]

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

First of all, congrats for the great writeup and thorough documentation. Even i don't do that sometimes.

Thanks, I figure if I'm getting free advice, the least I can do to put is put in the legwork to describe my problem and what I've done.

That MOSFET seems to be okay, but you are testing the wrong one. The ones on the heatsinks are what you should be worried about. Trace the path from the resistor that blows up and you will arrive at the cause of failure.

Test QM802 and QM803 for a short.

WOW! Look at all those other MOSFETs that were hiding. Actually, on closer inspection, some of those MOSFETs have diode symbols on them. I'm guessing those aren't tested the same way as QM802 and QM803? There's also a MOSFET on the large heatsink on the far right of the board, lower right hand corner marked ICM851. It's 5-pins, and I can't quite make out the manufacturer. I think it's SK, or SIK, and 2008HFE. I couldn't find a match after a quick Google.

I tested QM802 and QM803 and here's what I got:

Black on S1, Red on D1 - .4 ohms
Black on S1, Red on G1 - 1.6 ohms
Black on D1, Red on G1 - 1.6 ohms

Black on S2, Red on D2 - .4 ohms
Black on S2, Red on G2 - 1.3 ohms
Black on D2, Red on G2 - 1.3 ohms

So it looks like these are bad? Is there anything special I need to do to remove them and replace them, or is it just a matter of desoldering and resoldering? While reading up on MOSFETs, I found this forum thread, and this post in particular caught my attention:

http://www.edaboard.com/thread125626.html#post549583

I don't have an antistatic wrist strap. Should I get one when I buy new MOSFETs?

[Th3_uN1Qu3]

Re: Getting started with troubleshooting [Samsung 245BW]

WOW! Look at all those other MOSFETs that were hiding. Actually, on closer inspection, some of those MOSFETs have diode symbols on them.

If it looks like a duck and quacks like a duck, it's a duck. If it has a diode symbol and tests like a diode, then it is a diode. They are most likely double diodes with common cathode on the center pin. Check those for shorts too. Normally, a secondary side short should make the power supply fail "gracefully" ie it shuts down, but you never know. Use the diode setting of your multimeter. Black lead always goes to cathode, in this case center leg. Since it is a double diode, move the red lead between the outer legs and get readings on both. They should read between 1000 and 100. If they read lower than 100, reverse red and black leads and see what you get. A diode must only conduct one way so if you get a reading the other way round it's bad.

There's also a MOSFET on the large heatsink on the far right of the board, lower right hand corner marked ICM851. It's 5-pins

That's a power supply controller with integrated MOSFET switch. Probably for the standby supply. Since you have standby voltage then that part is OK. Good for you - they are very difficult to get a hold of and a cumbersome adaptation would have been needed if the original part could not be located.

I tested QM802 and QM803 and here's what I got:

Black on S1, Red on D1 - .4 ohms
Black on S1, Red on G1 - 1.6 ohms
Black on D1, Red on G1 - 1.6 ohms

Black on S2, Red on D2 - .4 ohms
Black on S2, Red on G2 - 1.3 ohms
Black on D2, Red on G2 - 1.3 ohms

So it looks like these are bad?

Uh huh.

I don't have an antistatic wrist strap.

Neither do i. This is a non-issue nowadays with most semiconductors. If you read a datasheet, you will notice that they have been tested for ESD (electrostatic discharge). Nothing bad will happen to them if you handle them, unless you happen to be rubbing your feet on the carpet or something.

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

Thanks for the confirmation Th3_uN1Qu3. So, are MOSFETs like resistors - where if all the specs match, they're all basically the same? Or are they like capacitors where I also need to watch for brand and a series?

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

Actually, nevermind. Sadly, my local electronic shop doesn't have them, so I'll have to order it online. I wasn't able to find it on DigiKey (probably because I didn't know the right catagory to narrow it down to), but I found the exact replacement on mouser.com:

http://www.mouser.com/ProductDetail/...R9QRetsMpF3E9k

Since these are connected to heatsinks, I assume I'll need some thermal paste? I have some leftover Arctic Silver that I used on my CPU heatsink. Will that be sufficient?

[retiredcaps]

Re: Getting started with troubleshooting [Samsung 245BW]

Actually, on closer inspection, some of those MOSFETs have diode symbols on them.

Those are probably schottky diodes.

I tested QM802 and QM803 and here's what I got:

Black on S1, Red on D1 - .4 ohms
Black on S1, Red on G1 - 1.6 ohms
Black on D1, Red on G1 - 1.6 ohms

Black on S2, Red on D2 - .4 ohms
Black on S2, Red on G2 - 1.3 ohms
Black on D2, Red on G2 - 1.3 ohms

So it looks like these are bad?

Correct. They are shorted. Desolder them and re-verify they have same readings.

A diode only allows current in one direction. Since these are shorted they are causing the RM801 resistor to blow.

[Th3_uN1Qu3]

Re: Getting started with troubleshooting [Samsung 245BW]

Since these are connected to heatsinks, I assume I'll need some thermal paste? I have some leftover Arctic Silver that I used on my CPU heatsink. Will that be sufficient?

Since they are "live" (at mains potential), there will likely be a rubbery thermal pad between them and the heatsink. Just re-use that. Also make sure that you don't lose the plastic washer on the screw if that's the case. Power semiconductors don't need as good thermal transfer as a CPU does, you can just use cheap white silicon paste if you want to, but in most cases that pad is good enough by itself. Arctic Silver is likely to be conductive (even if they say it isn't).

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

Correct. They are shorted. Desolder them and re-verify they have same readings.

Yup, they have the same readings out of circuit as well. Unfortunately, I may have messed up. Removing these were a lot tougher than capacitors and resistors, and I lifted the trace up slightly on one of the legs of one of the MOSFETs. I attached a picture - though it's pretty blurry. Does that mean this board is done for, or is there something I can do to repair it?

Since they are "live" (at mains potential), there will likely be a rubbery thermal pad between them and the heatsink. Just re-use that. Also make sure that you don't lose the plastic washer on the screw if that's the case. Power semiconductors don't need as good thermal transfer as a CPU does, you can just use cheap white silicon paste if you want to, but in most cases that pad is good enough by itself. Arctic Silver is likely to be conductive (even if they say it isn't).

Thanks, good to know. Hopefully I didn't break my board and I'll put this info to use...

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

I just checked the continuity of that lifted pad by touching the pad with one lead and another component along the trace further down. It still shows continuity. Hopefully that's a good sign?

[Th3_uN1Qu3]

Re: Getting started with troubleshooting [Samsung 245BW]

That's fine don't worry. Been there done that many times. You can always use a piece of wire in case you mess it up permanently. Get yourself a solder sucker and some desoldering braid and you won't lift a pad again ever.

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

That's fine don't worry. Been there done that many times. You can always use a piece of wire in case you mess it up permanently. Get yourself a solder sucker and some desoldering braid and you won't lift a pad again ever.

Hahahaha, I have both, and I still lifted the pad on this one. Guess I need more practice on the bigger joints.

I am actually surprised I got this far before doing something like that, I usually screw up much earlier on. Knock on wood - I'm not out of the woods yet...

Parts are on order - Thanks again guys for all the help!

[retiredcaps]

Re: Getting started with troubleshooting [Samsung 245BW]

It still shows continuity.

If it didn't show continuity, you can always run a wire to something that is in the circuit. And since you already mapped it out earlier, it would be easy to do.

[Rtech]

Re: Getting started with troubleshooting [Samsung 245BW]

Glad you now know why the resistor, burnt out.This liitle Mosfet tester, is easily made from scrap components,and if I am in any doubt at all whether I have one that is good or not..then this quickly sorts it out:-

http://translate.google.com/translat...dor_mosfet.htm

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

Thanks Rtech, I may actually try and put one of those together just for fun.

While I'm waiting for the parts to arrive, I was wondering - should I be looking for what caused the MOSFETs to short out? I ordered twice as many mosfets as I needed, but if there's anything I can be doing now to minimize problems when they get here, I'd like to do that.

[Rtech]

Re: Getting started with troubleshooting [Samsung 245BW]

Unlikely,Mosfets do fail,and regularly take out this resistor.On a Dell with a one board system and usually designated by the letters FPf, 6-8 components can fail,and calls for a rebuild of the Power supply,so you are not doing so bad.Good luck with the repair, and let us know how you get on.

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

My new MOSFETs arrived today, but I had a quick question. On closer inspection, I see that the MOSFETs I purchased are FQPF 9N50CF, but the original broken ones are FQPF 9N50C.

Mouser.com shows the 9N50C as obsolete, so I guess I don't really have a choice, and the datasheets for the two are almost identical. However, there are a few differences, so I'm not sure if I should be concerned:

FQPF9N50CF.pdf

FQPF9N50C.pdf

[Th3_uN1Qu3]

Re: Getting started with troubleshooting [Samsung 245BW]

The ones you bought are a tad weaker. The difference is low enough to be ignored though - in extreme cases it could have appeared as a manufacturing tolerance. Drop 'em in and see what gives.

To make sure you don't blow those too if there's something else wrong, replace the fuse with a 60 or 100W light bulb. It has to be the regular incandescent type, a CFL won't work. The bulb should flash briefly as the main capacitor is charged, then it should light at very low brightness when you power up the monitor. If it glows brightly, immediately disconnect the monitor from the mains and look for other sources of trouble.

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

The ones you bought are a tad weaker. The difference is low enough to be ignored though - in extreme cases it could have appeared as a manufacturing tolerance. Drop 'em in and see what gives.

To make sure you don't blow those too if there's something else wrong, replace the fuse with a 60 or 100W light bulb. It has to be the regular incandescent type, a CFL won't work. The bulb should flash briefly as the main capacitor is charged, then it should light at very low brightness when you power up the monitor. If it glows brightly, immediately disconnect the monitor from the mains and look for other sources of trouble.

Ah ok, thanks for clarifying!

To do the test with the light bulb as you mentioned, I'm guessing I'd need to get a light bulb socket like the one I attached, and some alligator clips to clip to the metal prongs holding the fuse? Also, which fuse would I connect it to since there are 2 fuses?

[Th3_uN1Qu3]

Re: Getting started with troubleshooting [Samsung 245BW]

To do the test with the light bulb as you mentioned, I'm guessing I'd need to get a light bulb socket like the one I attached, and some alligator clips to clip to the metal prongs holding the fuse?

Yes, something like that. You connect the bulb in place of the fuse on the AC line. Circuit wise, it's the first thing on the board after the power socket. Though i usually just hack an extension cord and plug the device into that, leaving the fuse in place - splice one wire of the cord, wire the bulb socket, wrap it in a little electrical tape, et voila you have an universal power supply short tester.

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

Thanks Th3_uN1Qu3, I found some links for how to build a "light bulb current limiter" which, from the descriptions, sounds exactly like what you described. I think this one may be the easiest for me to wire up myself:

http://www.io.com/~beckerdo/guitars/...ter/index.html

[nerdbot]

Re: Getting started with troubleshooting [Samsung 245BW]

Well, not fixed yet, but more progress...

I put in the new MOSFETs with a little silicone thermal paste (not the Arctic Silver, just to be safe) and put a new 1W 0.22 ohm resistor in RM801. I was also able to wire up that light bulb current limiter using an extension cord, so I was able to use that to make sure there wasn't a short somewhere else that would blow the resistor or the new MOSFETs.

I then connected the logic board to the power board, and now I get a steady 5.2V on the S_B pin, even with the logic board connected.

I connected the bezel buttons, and the power button lit up and turned on and off as I expected.

I then connected everything up (logic board to LVDS connector, power board to inverter board, invertor board to CCFL connectors, and attached a VGA cable from the monitor to my laptop. Turned on the monitor, and the display properties on my laptop picked up the monitor. I set it to extend my desktop to the 2nd display, and the monitor status light went from "standby" blinking to solid "on", but the screen didn't turn on. I then set the displays to go to sleep after 1 minute, and the monitor properly went from "on" to "standby", and also woke up properly when I moved the mouse.

I disconnected the logic board from the LVDS connector so I could flip over the chassis and access the pins on the logic and power board. I took measurements on the power board connector that runs to the inverter board (the ones marked "+24V", and they all read 0V.

I'm guessing that means there's still something wrong on my power board, right? At least the SMPS issue is fixed