Repairing an IPM, replacing 'die' MOSFETS

[Alastair E]

A few guys on some Polish forums have repaired IPM modules by removing the surface-mount, die silicon devices they contain and fitted conventional package MOSFETS in their place.

I'm thinking of giving this a go, and have a MOSFET in mind thats designed for PDP applications, as we have no idea the ratings of the original die component--



Anyone here done anything similar?

[tom66]

Re: Repairing an IPM, replacing 'die' MOSFETS

Seems like a lot of effort - the IPMs are only about £18 right?

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

Well--The £18 ones are so-called, 'refurbished' things--probably culled outta 8 year old sets with dead PSU's!

Seems proper 'new' ones are more like £40 mark....

[tw2005]

Re: Repairing an IPM, replacing 'die' MOSFETS

Well--The £18 ones are so-called, 'refurbished' things--probably culled outta 8 year old sets with dead PSU's!

Seems proper 'new' ones are more like £40 mark....

Yeah, and even so called New are pulled used. Some of those Ebay sellers have a lot of nerve. I did manage a new one once , forgot the seller now.

[vinceroger69]

Re: Repairing an IPM, replacing 'die' MOSFETS

is there a guide on how to test replace the faulty parts in these ipms.

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

Finally got back to this one.

I carefully looked at the way the IPM is made, and tried a component-tester across the MOSFETS.
--This proved rather useless, as the Gate-circuitry/driver stage was confusing the tester.

A careful bit of work with a scalpel cut the tracks to the Gate connections, at a point a few mil from the pad where the silver wire attached the track, leaving the Gate floating.

I tested the Sustain fets first--they were fine, so turned attention to the ER fets.

These tested as capacitors of 2 odd nan. I checked and confirmed the connections etc many times, but still have the same result--Cap, 22000 odd pf....

I am going to try replacing these fets with more conventional package parts, see what happens....

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

Hmm--Research suggests, these devices are not MOSFETS, but IGBT's....

--Confirmed--I think..... as one pair in the IPM has diodes in parallel,--There would be no need for a MOSFET to have a parallel diode, as MOSFETS have that diode as part of their make-up anyway......

[tom66]

Re: Repairing an IPM, replacing 'die' MOSFETS

Basically all plasma TVs use IGBTs for the sustain/ER circuit. MOSFETs are only used for low power (Ve/Vzbias pulse, setup ramp/setdn ramp, pass forward.)

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

Well,--Success!

Ive fitted two FGPF4536 IGBT one in place of each damaged device on the ER side of the IPM.

These are 50A 360V devices made with PDP in mind.

The new IGBT are TO220 and were mounted to the IPM heatsink and wired into the connections originally occupied by the die part.

The set was run for around half hour, and appears as it was before the IPM failed--complete with the red-spot syndrome these LG are known for...

Temperatures of Y-SUS and X-SUS heatsinks appear much the same on both--ie, Hot but not excessively so.....

These things definitely need some cooling assistance--if for nothing else, peace of mind!

[Glitcher]

Re: Repairing an IPM, replacing 'die' MOSFETS

Nice, do you have any photos? Would love so see how you fitted it and might try it out myself as I've been getting a lot of these bad IPM Ysus.

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

TBH--I didn't take any pics, as I fully thought that the set--wouldn't work!

I say this, as I found my component-tester is somewhat erronious in its testing of MOSFETS/IGBT's due to only using 5V for the Gate voltage. Some devices--perfectly good, test as bad--Solved that issue by adding a dry-cell in the Gate test lead, but not before I had destroyed the die parts in question....
--Some devices need maybe 6V to test properly, and I thought I may have made a boob!

To give you a vague idea, I cut away the die part by scraping it off Couldnt get it by heat with a hot iron for some reason. I did notice a hard black coal like layer in its make-up so maybe that could have been the point of failure, not sure.

I also cut away two slots in the sides of the plastic frame of the IPM to allow the wires to the new parts to enter the IPM. I couldnt find smaller devices or S.M. types in my stock/boards pile, that could have been mounted direct where the die devices were.
-Unfortunately the IPM plastic mounting-lug came off but is still usable as it has a lip which still locates over the IPM metal base to hold it to the heatsink.

The leads of the new parts were cut pretty close to the new devices and this allowed them to be mounted close to the IPM--I believe that the distance between the new part and the original place should be as small as is possible--to avoid issues of Gate signal pick-up.
--There's Huge currents flowing in pulse/switching of these things and possibility of pulse pick-up in the Gate circuit must be considered I think,--cant hurt anyway.

The new parts were the type that didnt need any insulator, and were mounted direct to the heatsink, with a silicone pad for good thermal contact and close to the gaps now in the side of the IPM plastic frame.
Running the wires was done in the usual way, but where the original device connections lead to the IPM pins, the wires were connected to the pins themselves, on the IPM substrate end of the pins.

The pins of the IPM were lengthened by soldering approx 1/4 inch of stiff copper wire to them--for two reasons.
--I had mis-calculated the position of a diode on the board and the placement of one new device--It would no longer fit!....
--Also, to allow easier IPM removal when it was found that it didn't work!

In next few days, I'll get back to this set, I'll update the control-board software with that LG jig I have, as I now have access to a Win XP laptop to load the software on!--Wont work with Win7, Doh!.....

I'll botch in a couple of quiet fans on those heatsinks--there's plenty of room in this set and 9 or 12V supply can be from one of the many the PSU provides......

Maybe if its performance/picture is good enough, I'll chuck it on the wall--see how long it lasts--IF it gets that far without blowing up!

[ben7]

Re: Repairing an IPM, replacing 'die' MOSFETS

Wow, good job!

Really, you SHOULD get some pics!
Though, you soldered the IPM back in didn't you? That may be hard to remove again.

Yes I agree with adding some fans. It'll help with reliability, especially when it comes to the capacitors.

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

Hmm--So my plan was to update the firmware on the Control-Board/LVDS/T-Con, (you choose what to call it!) --With the official L.G. 'SVC' kit, its jig and software with included updated firmware files, USB cables and the 15 pin leads to go to the board.

It did NOT go well!

I now have a 42" Brick, --with sound!

I followed the included instructions that came in PDF format, but for some reason the firmware does not install correctly and the 'verify' confirms its no good.

I'll bung a post up--see if anyone else has this kit and knows whats gone wrong....

Ive had to resort to buying a control-board from fleagay--Not happy!

As to removal of the IPM--its now possible to de-solder each pin at the join, so shouldnt present too much of a problem.

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

After much messing round--I found the cause of the update fail,--was the computer....

Tried a different one, loaded first attempt and we back in business.

This turned the red haze of spots into a light drizzle of green spots........

Tweaks to Vs and SET_UP/SET_Dn helped but when cold--still had some green spots especially in one corner.
Looked around, and found a more up to date firmware file from Tom's TV Service-manual site.
My update file from 2006 ended in CE5.bin, while the more recent I used ended in D88.bin.
--Installing this later version has completely cleared the spotting (Mal-Discharge) the set had.

Ive fitted some fans to the IPM heatsinks, and its run now for 3 or 4 hours, the fans keep the heatsinks practically stone cold.

We see how reliable this thing is over the next few weeks......

[johnboy1313]

Re: Repairing an IPM, replacing 'die' MOSFETS

Pictures or it didn't really happen.

[Alastair E]

Re: Repairing an IPM, replacing 'die' MOSFETS

Hmm---

Some of us are too busy to take picturess of everything--The account is accurate, you can believe it or not, doesn't bother me either way!

[tom66]

Re: Repairing an IPM, replacing 'die' MOSFETS

That's a long running internet joke. We believe you...

[johnboy1313]

Re: Repairing an IPM, replacing 'die' MOSFETS

...or do we?

[tom66]

Re: Repairing an IPM, replacing 'die' MOSFETS

You know what John, I'm not even sure if I believe that you exist. There, I said it.

[johnboy1313]

Re: Repairing an IPM, replacing 'die' MOSFETS

Maybe we're both figments of someone else's weird dream about broken TVs.