Transistor replacement for L240AS-00

[Winston Chadwick]

I've been repairing Dell Optiplex 790 PSUs, and I'm finding a lot of dead LS240AS-00 rev. B models. The main input cap dries out which blows Q100 and the fuse.

Q100 is an Infineon Technologies 6R199P.

It looks like Summit Electronics and SemiconductorPlus carries the 6R199R in the US. Has anybody ordered from these sites? Are they better than Aliexpress?

It looks like Lite-On put a ferrite bead around the base pin of Q100. The bead is 4mm in diameter. Does anyone know where I can get replacements for these?

[R_J]

Re: Transistor replacement for L240AS-00

It looks like the actual number would be IPP60R199CP, then digikey has them http://www.digikey.com/product-detai...SA1-ND/1282065
or so do others
If you need isolated case use IPA60R199CP

[Marianelforum]

Re: Transistor replacement for L240AS-00

It looks like Lite-On put a ferrite bead around the base pin of Q100.

It is the "gate", and not the base, because it is a mosfet transistor, and not a bjt one, it is good to avoid making this confusion in the future.

[Winston Chadwick]

Re: Transistor replacement for L240AS-00

It looks like the actual number would be IPP60R199CP, then digikey has them http://www.digikey.com/product-detai...SA1-ND/1282065
or so do others
If you need isolated case use IPA60R199CP

Thanks. I didn't know that was the full ID for the part. Should I reuse the ferrite bead around the gate if I get the isolated case? This is filtering noise right?

It is the "gate", and not the base, because it is a mosfet transistor, and not a bjt one, it is good to avoid making this confusion in the future.

Noted. Thanks for the correction.

[R_J]

Re: Transistor replacement for L240AS-00

Yes use the ferrite bead on the gate, be careful removing it they can break easily (cut the leg below it and slide it off, don't use pliers to pull it off), If the original did not use the isolated case then you don't need to either. If the original was mounted with a mica insulator you could use the isolated tab version, but best to stick with the original as it could use the tab as part of the connection to the circuit.

[Winston Chadwick]

Re: Transistor replacement for L240AS-00

Thanks for the clarification R_J. It was mounted with insulation, but I'll get the non-isolated version to keep it factory spec.

It doesn't look like the manufacturer used any thermal material between the MOSFET and the heatsink.

I have some Arctic MX-4 thermal paste. Should I put tiny dot on the front and back of the insulator layer for better heat conductivity?

Edit: It looks like the heatsink screw is insulated on both sides of the heatsink.

[budm]

Re: Transistor replacement for L240AS-00

It uses Thermally Conductive Silicone Interface Pad.
http://solutions.3m.com/wps/portal/3...Pads?N=8704982

[R_J]

Re: Transistor replacement for L240AS-00

Like budm said, that silicone sheet is both the insulator and thermal conductor and must be used to prevent the fet tab from contacting the heatsink

[etnietering]

Re: Transistor replacement for L240AS-00

My company has a few dozen Optiplex 790s with these L240AS-00 power supplies. 6 or 8 of the PSUs have croaked over the last year or two. Always the same - the PC just turns off, and won't power back on. I've snagged a couple before they could be junked, and have fixed two.

The two I've fixed exhibited the same symptoms as mentioned above - blown fuse. Upon disassembly, I've found that Q100 is failed short and the primary bulk cap reads very low (2uF or lower).

Unfortunately the IPP60R199CP transistor is marked as "Not for new designs", and when I was working on these, out of stock at digikey with a very long lead time. I did some looking through the other available TO-220 power MOSFETs, and bought a couple different ones to try. In both of the units I've fixed so far, I used the IPP60R180P7XKSA1. The characteristics are about as similar as I could get them. Once I have a chance to try any of the others, I'll post back with the results.

In addition to replacing the dead components, I decided to do a full recap on these. The factory caps are an unimpressive mix of TAICON and Ltec, and these PSUs are nothing special efficiency wise - not even 80 PLUS standard! And the fan barely spins, even when things heat up.

First, a sticker shot. Surprisingly low current available on the +3.3V rail. The motherboard in these PCs must not use that rail for much.


Next, two overview shots. One with just the cover removed, and one with the board out of the case.



Here's a close-up of the primary end. The big heatsink holds the main switcher, and the PFC transistors and diode. The little heatsink holds the bridge rectifier.


The components that have to come out are the fuse, primary bulk cap, and the big primary heatsink. All that goop in there makes it a bit of a chore to get some of this stuff out. Fortunately, it's not the crap that turns conductive as it gets old!


And they're out! The blown transistor is on the left-front of the heatsink.


Here they are with their replacements. The fuse is a 250VAC 6.3A slow-blow. I opted for the ceramic cartridge over a glass one, but didn't bother heat shrinking the whole thing. The replacement cap is a Nichicon CP series, rated for 10,000 hrs @ 105C. It's also a little shorter than the original - there's not much clearance for a taller one.


Here's the heatsink with the new transistor. I reused the ferrite bead, and attached it with a little drop of superglue. I also bent the center lead out a bit, so that it'd match up with the holes in the board. The better it's lined up with the leads of the diode, the less wiggling it takes to get the heatsink back in place.


This is all that's strictly necessary to revive one of these (assuming it has the same problem). If you want to do the full recap, however, now is the time to get the little guy on the primary side of the daughterboard. It's a 22uF 50V cap - I replaced it with a Panasonic EB series. Before:

And after:


To be continued...

[etnietering]

Re: Transistor replacement for L240AS-00

And here we have the primary side reassembled.


Next I replaced the 8 secondary side bulk caps. The originals were:
3.3V: 1500uF 6.3V TAICON HD
5V: 2200uF 10V Ltec LZG -> coil -> 1200uF 10V TAICON HD
12V: 2x 1200uF 16V TAICON HD
5VSB: 3300uF 10V Ltec LZG -> coil -> 1000uF 10V Ltec LZG
-12V: 220uF 25V Ltec LZG

The 3.3V rail only has 1 cap because it's derived from the filtered 5V rail with a MOSFET. This provides very clean power without much filtering.

Here's the originals:


And the secondary side with all the caps gone:


And the replacement caps lined up below their originals:


My replacements are a mixture of Panasonic FC, FM, and Nichicon HE. Basically, whatever I found in my stash of parts that fit.
3.3V: 2700uF 6.3V Panasonic FC
5V: 2200uF 10V Panasonic FC -> coil -> 1000uF 10V Panasonic FM
12V: 2x 1500uF 16V Nichicon HE
5VSB: 2200uF 10V Panasonic FC -> coil -> 1000uF 10V Nichicon HE
-12V: 680uF 16V Panasonic FC

Here they are, installed. Unfortunately I didn't get a picture from the same angle as the other two. I did have to scrape out some of the goop in order to find the lead holes from the top. And make sure the positive side of the -12V cap goes to ground.


Only one more, I promise...

[etnietering]

Re: Transistor replacement for L240AS-00

The next replacements were the 3 caps in the feedback/control part of the primary side, in the center of the board. Here we have:
22uF 50V Ltec
470uF 35V TAICON
1uF 450V TAICON

I didn't write down the series for any of these, I'm afraid. Here they are:


And removed:


The replacements are:
22uF 50V Panasonic EB
470uF 35V UCC KYB
1uF 450V Panasonic NHG

Here are the new ones lined up with the originals:


And the new ones installed (after more goop removal):


The two final electrolytic caps are on the secondary side of the daughterboard. I call these the li'l bastards, because they are quite difficult to get to. The proper way would be to either remove the daughterboard completely (30+ solder connections and goop), or remove the secondary heatsink (~20 solder connections to huge traces, and goop). So what I did was get out my tweezers and my cheat sheet of four letter words. Half an hour later, both caps have been swapped with Panasonic EB. I was unable to get a good (or even passable) picture of the caps themselves. They're just below the trimpots. There's a 10uF 50V, and a 4.7uF 50V. This picture attempts to show where they are from the solder side of the board. The 4.7uF cap is the one on the left, and the solid bars represent the negative leads.


I also made an improvement to the fan control circuit, in an attempt to keep things cooler. I added a 10Kohm resistor in parallel with the temperature sensor (an NTC thermistor) to make the fan run faster. I took some rather unscientific measurements of the thermistor's resistance, using my hot air station to heat up the heatsink. At room temperature, I measured about 1850 ohms, with the heatsink "warm" I measured about 1600 ohms, and with the heatsink very hot I measured close to 1300 ohms. With the addition of the 10Kohm resistor, the readings changed to about 1550 ohms cold, and 1150 ohms hot. The thermistor forms the top half of a resistive divider circuit, whose center voltage is fed through an op-amp into a 2SB772 PNP transistor to power the fan. This circuit was quite difficult to try to analyze, being split over two perpendicular circuitboards, the tops of which were blocking each other pretty effectively. Also, I have not done any measurements to try to determine whether or not the fan is actually spinning any faster. If I ever do, I'll be sure to post the results.

Here's a picture of the 0603 SMD resistor I soldered across the legs of the thermistor:


And that's everything. I've had both of these PSUs attached to my lame 85-ish watt dummy load, and neither have had any trouble after several hours. Not all that stressful (~35% load), but I have not had a chance to test them in one of the Optiplex 790s, or on a better dummy load. I'll be sure to post an update once I have anything new. Hope y'all have enjoyed my pictures!

[momaka]

Re: Transistor replacement for L240AS-00

Wow, great post etnietering, and glad to see you back in BCN!

These small odd-ball PSUs from Dell have also given us a lot of trouble back when I worked at a repair shop last spring. Unfortunately, the repair shop was more of a parts swap place, so I didn't get to analyze what was wrong with the old PSUs. But it should suffice to say that many of the new Dell Inspiron machines still come with similar odd-ball PSUs like this (actually even more than before), and they have even more problems it seems.

Keep us posted how the repairs work out for these PSUs. I fear that many modern PSUs with APFC simply are not designed very well and thus they end up stressing their primary filter caps more than they should. Combined with cheap unreliable non-Japanese brands, I wouldn't be surprised to see more of these and other similar new APFC PSUs go bad.

[stj]

Re: Transistor replacement for L240AS-00

Q
why didnt you get the old rubber compound off before fitting the new caps?

[etnietering]

Re: Transistor replacement for L240AS-00

Wow, great post etnietering, and glad to see you back in BCN!

Thanks! While it was nice having all my equipment working without issue, I'd forgotten how much I enjoy making these kind of repairs.

These small odd-ball PSUs from Dell have also given us a lot of trouble back when I worked at a repair shop last spring. Unfortunately, the repair shop was more of a parts swap place, so I didn't get to analyze what was wrong with the old PSUs. But it should suffice to say that many of the new Dell Inspiron machines still come with similar odd-ball PSUs like this (actually even more than before), and they have even more problems it seems.

At work we've been trying to use thin clients wherever possible. In the places that do need an actual PC (due to special I/O, GPU requirements, etc.), we've been getting mini towers instead of the little SFF jobs. Of course, in the most recent batch (I think they're Optixplex 3040s), the regular ATX sized PSU cutout has an adapter in it so that one of these dopey long skinny SFF PSUs will attach. Haven't had any trouble with those, yet...

Keep us posted how the repairs work out for these PSUs. I fear that many modern PSUs with APFC simply are not designed very well and thus they end up stressing their primary filter caps more than they should. Combined with cheap unreliable non-Japanese brands, I wouldn't be surprised to see more of these and other similar new APFC PSUs go bad.

I did read one of your long posts on how these APFC circuits can stress the primary caps. You suggested getting a small film cap to put in parallel, to handle some of the ripple. Unfortunately, there's no room for something like that in this PSU. It's pretty solidly jammed full of components. And the primary cap is on the hot end, of course. That is why I spent a little extra for the 10,000 hr rated cap, and tried to speed the fan up a bit. We'll have to see if that makes much difference.

Q
why didnt you get the old rubber compound off before fitting the new caps?

Two reasons for this: 1) it's a pain. It's adhered to the board pretty well, and when I to pull it out with tweezers, that part I've got hold of just rips off. I did try being a little more aggressive with it, but 2) I found that in places there are components completely buried within the goop. The last thing I wanted to do was accidentally damage one or tear one off completely, and then have to guess where it went, and/or attempt to source a replacement. Between the 2nd and 4th images in my last post, you can see just below where the brown UCC cap is, I partially uncovered a small glass diode of some variety.