Samsung SyncMaster 940T blows inverter fuse

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

Link to Newark's page on the FK 820s (my edit window expired):
http://www.newark.com/panasonic/eeuf...ial/dp/32R9027

[mockingbird]

Re: Samsung SyncMaster 940T blows inverter fuse

Ok then, stevenps suggested the FK. It looks like a good match.


It's perfect actually. ESR is 0.035, should be more than good enough. I'm pretty sure I can put the two 820uF instead of 1000uF.

[Quasar]

Re: Samsung SyncMaster 940T blows inverter fuse

...The good news is that I discovered that Samxon does indeed have low-esr suitable 25v 1000uF capacitors 10x20mm caps so I'll end up probably buying those. Thanks for your help...

You're welcome!

Samxon lies too. I checked Rubycon, and they confirm the other 3 manufacturers info I posted, 680/25 is the top end for 10x20. Above that it's 12.5mm.

If 4 MAJOR manufacturers all match fairly close on their sheets, odds are that those that claim better are full of it. More from less? Something has to give when you pack that much capacitance into that smaller can. I'd bet either OV surge limit or lifespan. There's more foil and less electrolyte.

Q

ISn't FK surface mount?

No. Radial, through hole.

[mockingbird]

Re: Samsung SyncMaster 940T blows inverter fuse

ISn't FK surface mount?

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

If the original 1000uF caps are not really 1000uF, you might be able to get away with 820uF 25v Panasonic FK 10x20. It's at the low end of 20%, but it might be good enough. Newark/Farnell sells them.

[mockingbird]

Re: Samsung SyncMaster 940T blows inverter fuse

I *might* be able to get a 12.5mm in there, but that would be like milking a frog. Things are just very tight over there, and that dried up toothpaste that they used aint helping things. I tried scraping some of it off but I'm afraid I'll end up cutting through the PCB or hitting something - it's some hard stuff.

Besides that, I can't really fiddle with this too much. The underside has lots of little fragile components and the soldering is Delta - very neat and professional - but at the same time very small and the points and traces are too close for comfort. I can't justify the effort. It's a 20" 4:3 LCD.

The good news is that I discovered that Samxon does indeed have low-esr suitable 25v 1000uF capacitors 10x20mm caps so I'll end up probably buying those. Thanks for your help.

Oh, no it's not a customization, that's the PCB out of the chassis, I'm just leaning it against stuff for the purposes of taking a pic .

[Quasar]

Re: Samsung SyncMaster 940T blows inverter fuse

I'd bet #3 & 4 can be 12.5 mm. You'll have to lay #2 over a bit, then #4 will fit. #3 can exceed the board edge and a little bend to the wire going to the resistor and you should be okay. Mount them a little proud of the board so you can get some movement.

2 things with the pics - The first one is a nice top down... on the transformer, so I may not be seeing the cap spacing correctly. Same with the second pic as it's straight on with the board edge, and not the cap tops to show the lack of clearance. It seems you have a few mm's to deal with on height.

Another shot of the foil side over the back of those caps area would be nice.
Sometimes you can move or pivot a component enough by drilling another lead hole. Then all kinds of room opens up. I do this on PSU outputs to get a -real- 3300uF cap in there.

Is this a customization of something? Is that wood I see over the top?

Q

[mockingbird]

Re: Samsung SyncMaster 940T blows inverter fuse

Quasar -
Thank you for your helpful reply. Here are two pictures (attached for archival purposes):



Here are the caps by their numbers in the picture:
1) 25v 220uF
2) 25v 220uF
3) 25v 1000uF
4) 25v 1000uF
5) 10v 2200uF
6) 10v 1000uF
7) 25v 220uF
8) 50v 22uF

Keep in mind, caps 5 and 6 are Taicon PW and the rest (besides the tiny starter cap) are CapXon KF so they all need the low ESR KZE caps. So I assume we can't borrow from caps 5 and 6. Which leaves the 25v 220uF caps. But are they in parallel?

If the capacitors are in parallel I assume that if I raise the capacitance of one of them, I can diminish the value of another (Caps 3 and 4).... I can't answer what the actual voltage is, as I don't have the technical know-how to find out. As for 680uF, that would be great. If I could put 680uF there I'd put KZE as I like UCC and everything else I'm recapping with either KZE or LXZ/LXY.

[Quasar]

Re: Samsung SyncMaster 940T blows inverter fuse

Going by the CapXon KF sheet, that dimension does not exist in the 1000/25 cap. 13x20 is what the sheet says with 45 mOhms and 1650 mA for that -value-. Yeah, it's a "special" order, but it isn't "real" as follows.

Going by what we've learned here at BCN, the "more correct" values for your cap, going by the KF sheet, would be indicated by the can -size- of 10x20mm. That, by the sheet, is the 820/25 (small) and the 680/25 caps can sizes. The values range from 52 to 65 mOhms & 1320 to 1400 mA. That's the values you should try to match. Personally, I'd want to be better than those values if possible.

There's a clue here when you examine other manufacturer's data regarding that can size in that voltage. CapXon is bogus & they're liars. The best FM, FC, HD, HE, KZE, or KY will go is 680uF.

Draw your own conclusions.

What is the *actual* voltage presented to the cap?

Can a pic of the predicament be posted? Perhaps with some indications as to what goes where? Sometimes you can change another cap in value, and still get everyone happy with overall capacitance and ESR.

Q

[mockingbird]

Re: Samsung SyncMaster 940T blows inverter fuse

Since we were discussing CapXon bogus values, I was wondering about the following.

I am recapping an LCD and there are two CapXon KF 25v 1000uF.

The problem is that their measurements are around 10mm x 20mm. The closest I can come to this series considering the dimensions are Nichicon HE or UCC KY.

If I step down to 16v 820uf, I can find Panasonic FC or Nichicon PW.

Yes, I'm aware of the discrepancies in ESR for these series compared to Capxon KF's, but I must have 10mmx20mm caps precisely, or I won't be able to solder them in if they're too wide, and the PCB won't fit inside if they're too tall.

[PlainBill]

Re: Samsung SyncMaster 940T blows inverter fuse

Turns out I can do this by using the software that came with the scope to record for me while I use probe wires to poke at the pins on the ICs. If I wait a few minutes between measurements, the FETs won't overheat. I'm not quite sure what I'm supposed to be looking for though - the frequency of the signal at inverter IC pin 9 (gate signal for the P-channel FET, which provides +13v) is around 190KHz, and the signal from anything else is hard to read at that time scale.

Also, what would happen if I remove both FETs and attempt to measure? The inverter should shut off pretty fast, but that might be enough time to see if the inverter IC is really providing a weird signal, or if the FETs are not switching properly. Would such a reading mean anything?

Removing the FETS may work. I suspect they are the problem anyway.

Here is what I laughingly refer to as my thought process. The 'bad' monitor shows significant differences in drive waveforms. These waveforms are important. When the N-FET is on, the P-FET must be off, and vice-versa. Also, the N-FET is turned on by the gate going high, the P-FET is turned on by the gate going low.

The first step is to get a reference drive waveform. It could be the output from pin 8, pin 9, or even the oscillator pin. Let's use the waveform at pin 8, which I assume is the drive for the N-channel FET. Now we want to see the waveform from pin 9. But we want to see the two waveforms as they exist at the same instant.

This is important. See the crude drawing I made. If you sync on the falling edge of the observed waveform it appears both FETs may be on at the same time, this indicates a problem with the inverter controller. But if you sync on one signal and display both waveforms you see the P-FET gate drops well after the N-FET gate. They are not on at the same time.

Now we've established that the drive signals from the inverter controller are correct. How about the drive signals at the gate of each FET? There appear to be resistors and caps between the inverter controller outputs and the FET gates. So we observe them at the gates. It isn't necessary to move the probe that was on pin 8, we want that signal present as a reference.

While we are at it, we should see the signal at the source pins of the FETs. Remember, the P-FET is turned on when the gate drops below the source. If the drive waveforms only rise to 10 volts and the source voltage for the P-FET is 14 volts, the P-FET is always on.

Last, we want to see the signal on the drains of the FETs. Of course, they are tied together.

One word of caution about the text display on the scope. That represents what the processor THINKS is happening. It's very good at handling sine and triangular waveforms. It's not so good at analyzing square waves with lots of noise spikes.

PlainBill

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

Solder a short lead scrap to solder blob then you will have a place to attach clips.
Can just clip it off later if there is no other reason the desolder the joint.

Turns out I can do this by using the software that came with the scope to record for me while I use probe wires to poke at the pins on the ICs. If I wait a few minutes between measurements, the FETs won't overheat. I'm not quite sure what I'm supposed to be looking for though - the frequency of the signal at inverter IC pin 9 (gate signal for the P-channel FET, which provides +13v) is around 190KHz, and the signal from anything else is hard to read at that time scale.

Also, what would happen if I remove both FETs and attempt to measure? The inverter should shut off pretty fast, but that might be enough time to see if the inverter IC is really providing a weird signal, or if the FETs are not switching properly. Would such a reading mean anything?

[PCBONEZ]

Re: Samsung SyncMaster 940T blows inverter fuse

Solder a short lead scrap to solder blob then you will have a place to attach clips.
Can just clip it off later if there is no other reason the desolder the joint.

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

First, I'm a little confused about the operation of the scope. Are the traces done at the same sweep rate and vertical sensitivity?

The ones that have the same frequency likely are. The settings are at the bottom of each image. I've been using the auto-set button to let the scope give me what it thinks is the best image, except on inverter IC pin 9 where the frequency is way off - it only shows half of a wave there, so I zoom out a little to get a better image.

Next, I'm uncertain about the last two traces in each set - Pin 8 and Pin 9. Are these the outputs of the inverter controller?

Those are taken directly on those two pins of the inverter controller, yes.

The technique I prefer to use is to use two probes. Hook one to pin 9 of teh inverter controllerand sync from that signal. Use the other to probe the following points.

Pin 8
Gate(s) of FET driven from pin 9
Gate(s) of FET driven from pin 8
Drain of U202
Drain of U203?

From these you should be able to compare 'good' and 'bad' monitors; it isn't necessary to post the pictures.

I'm not sure I'll be able to do that. I can only find a place to clip my scope probe on the pins of the transformers, and I can't hold two probes on two solder blobs while operating the monitor and scope at the same time. Also I'm not sure I'll be able to measure the gates of the FETs, as the legs are too close together; I was able to do the two pins of the inverter IC because they're on the end of the chip, and the FET drains are all connected together and attached to the transformers, so I was able to measure there. Do you have any suggestions on how I might be able to get this done?

[PlainBill]

Re: Samsung SyncMaster 940T blows inverter fuse

You are right, massive differences in the waveforms. I'm losing track of the details on a number of these threads. Please bear with me.

First, I'm a little confused about the operation of the scope. Are the traces done at the same sweep rate and vertical sensitivity?

Next, I'm uncertain about the last two traces in each set - Pin 8 and Pin 9. Are these the outputs of the inverter controller?

The technique I prefer to use is to use two probes. Hook one to pin 9 of teh inverter controllerand sync from that signal. Use the other to probe the following points.

Pin 8
Gate(s) of FET driven from pin 9
Gate(s) of FET driven from pin 8
Drain of U202
Drain of U203?

From these you should be able to compare 'good' and 'bad' monitors; it isn't necessary to post the pictures.

PlainBill

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

The pictures from the working monitor look much cleaner. In particular, the output from inverter IC pin 9 on my other monitors looks way off. This monitor has bulging/leaking 820uF caps, but I'm not going to replace them until after I get the others fixed (I don't want to accidentally damage the only one of these I have that works). I also noticed that if I hold my scope probe right next to the CCFL wires, I see a nice sine wave (I didn't check this on one of the bad boards). I'll see about measuring some of the surrounding components tomorrow to see if any are different. Please let me know if any other comparisons would be useful.

Also possibly interesting, people posting at https://www.badcaps.net/forum/showthread.php?t=5136 appear to have the same IP board I do (based on a comparison of the bottom of the board); most of them had a problem with bad FETs blowing the inverter fuse if it wasn't just bad caps causing a power supply failure.

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

I got in all the caps I ordered, and I completed the replacement. See my earlier post for the full list. Here is a new set of scope pictures, for the monitor I used the 680uF 25v caps on. I think I mis-labeled one of the images last time - the one labeled FET drain is actually the ground side of the transformers. They should be correctly labeled this time. I'll post again when I get numbers from my good board.

[stevenps]

Re: Samsung SyncMaster 940T blows inverter fuse

How about the 3193? I'd think that would be a close match. Worth a shot for the datasheet?

Looks like a close match based on what of the circuits I've been able to trace. The 3193 uses dual N-channel FETs though (the 3195G uses P & N channel FETs), so the timings may be different.

?? Posted this earlier. Seems to have been lost by the forum. Not the first time. ??

AF4502C:
http://cgi.ebay.com/ws/eBayISAPI.dll...m=320606366434

$1.10 each

Since they're pretty cheap, I went ahead and bought some. I'll have to do a DigiKey order this weekend to get the ChipQuik I'll need to do the job, and probably some smaller diameter solder as well.

BTW: My suggestion for the eBay board was not as a replacement, rather as a substitute that you could take good readings off of for comparison.

When you were done, you could sell it off.

I hope I don't need to do this since I was able to find a working unit, but I'll keep this in mind if the board I have isn't helpful. Thanks!

[Quasar]

Re: Samsung SyncMaster 940T blows inverter fuse

?? Posted this earlier. Seems to have been lost by the forum. Not the first time. ??

AF4502C:
http://cgi.ebay.com/ws/eBayISAPI.dll...m=320606366434

$1.10 each

BTW: My suggestion for the eBay board was not as a replacement, rather as a substitute that you could take good readings off of for comparison.

When you were done, you could sell it off.

Q

[Quasar]

Re: Samsung SyncMaster 940T blows inverter fuse

I wish we could find a full fledged datasheet for the BIT3195. Or even a source for the IC itself. I could swear those drive waveforms are wrong, but I DON'T have enough experience to be sure. I'll try to do a little digging for more information.

PlainBill

How about the 3193? I'd think that would be a close match. Worth a shot for the datasheet?


Q