Author Topic: Rohde & Schwarz Universal Radio Communication Tester CMU200: PSU problem  (Read 45129 times)

0 Members and 1 Guest are viewing this topic.

Offline Ericmax

  • Contributor
  • Posts: 14
  • Country: fr
If I trace the 12V STB properly, I arrive on 2651MTC-ADJ regulator circled in red in the picture.
I tested all the capacitors around and i disconnected  the diodes to isolate the input and output of the regulator.
after measurements between the input (Vin) and output (SW) I feel that the Internal MOSFET is dead but this remains to be confirmed.
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
If I trace the 12V STB properly, I arrive on 2651MTC-ADJ regulator circled in red in the picture.
I tested all the capacitors around and i disconnected  the diodes to isolate the input and output of the regulator.
after measurements between the input (Vin) and output (SW) I feel that the Internal MOSFET is dead but this remains to be confirmed.

Looking at the spec sheet for that chip, assuming the circuit here is similar.

a)
Power off.
Can you measure the resistance of Vin, AVin (the chip internal power supply), SW, PB and PGND with respect to the 12V STB ground pin?
Only the PGND should be a short.
b)
Power off.
Also from your comment I assume the two zeners nor the input and output caps (found in the "typical application" section in the spec sheet) are shorted.
c)
Power off.
Can you measure resistance between SW and FB as well as FB and PGND to find if the mosfets are shorted.
Power on.
Finally if these two mosfet are driven to turn on both at the same time and stay on it would also look like a short to load the line, but not with the power off. Any voltage on these pairs of pins?
d)
Could it be that the problem is on the output side of the 2651 and the chip is trying to drive a short elsewhere? Particularly if test a) and b) are OK should the output side look like a short? That could overdrive the 12V standby psu before it 2651 cutoff protection happens?

Hopefully that chip is still available if needed.

Unfortunately my unit is fully back together so I could not take a quick look at the circuit, unlike the last time. However, if you need it I am willing to dismantle the front module again.
OK it is past midnight here. I'll chime in tomorrow morning if I think of anything else.
« Last Edit: July 24, 2016, 04:53:25 am by richnormand »
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Offline Ericmax

  • Contributor
  • Posts: 14
  • Country: fr
Hi richnormand.

Thank you for your help and your proposal but do not disassemble your unit now works again  :-+
The LM2651 is available from Mouser, I'll order it tomorrow, will be easier.
all capacitors, inductors, and diodes are correct.
I made all feasible measures on inputs outputs by disconnecting peripherals that maximum components and regulator appears short circuit V in, SW and PGND on.
at different power with old TOP200, I measured on a 12V STB a 18Vdc briefly.
the maximum for the LM2651 is 14V, I think he's burning at that time.
I have confirmation when I replaced the LM2651.
I would tell you if it can help other members to troubleshoot their front panel.
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Hi richnormand.


at different power with old TOP200, I measured on a 12V STB a 18Vdc briefly.
the maximum for the LM2651 is 14V, I think he's burning at that time.


That is an interesting observation Ericmax. You would think that such a gross over-voltage would not happen, even in a no load situation. The PSU should have such a protection via a crowbar zener or limiter of some sort. It might be worthwhile to investigate the PSU behavior with the new TOP222 to make sure this does not happen before installing the new 2651.
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Offline Ericmax

  • Contributor
  • Posts: 14
  • Country: fr
since replacing the TOP200 by the 222, did I do many measurement with and without front panel.
of course, with the front panel, impossible to measure because it is safe mode. but without the CPU board I have always had a very stable U around 12,23V.
may be there is there you had other causes of the destruction of the LM2651.
Above all I have to replace to confirm my observations.
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Just had a look in one of the PSU's i've got, and the 6 1ohm resistors you have are different to mine, cause one of them is a diode.

I'll have to have a look at the other psu later, just in case it's a difference in the versions.
 
The following users thanked this post: richnormand

Offline Ericmax

  • Contributor
  • Posts: 14
  • Country: fr
if you have several PSU to compare, look at the date code and PCB revisions, this may give indications of eventual corrections.
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Can't see anything on the labels to show there's a revision number, everything looks the same apart from the serial number, and what might be a yy-mm-dd date code.

Had a poke around with the multimeter, and it looks like B1 is hits 10v, then dies, which takes A1 with it.

Also the quick check leds show that 5.2v and 3.3v aren't there, it did start up once or twice, but it quickly clapped out.

It's 4am here so i'll have a look at it later on.

Pics http://imgur.com/a/IHBNy
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Just had a look in one of the PSU's i've got, and the 6 1ohm resistors you have are different to mine, cause one of them is a diode.

I'll have to have a look at the other psu later, just in case it's a difference in the versions.

Well... |O

I took the PSU off the CMU200 and, just like you found out, I could not find an indication of revision. The rest of the markings are identical to your unit except for the serial number and manufacturing date.

I started to work on a different theory and spent an hour sifting through my bench garbage bin.

Guess what I found  :-DD  :wtf:  |O. Five failed one ohm resistors and one failed (open) diode..... That is what happens when you work late at nigh. So I probably went away to see what I had in smd resistors, duplicated the batch seen at the bottom of the picture and completely forgot about the diode!

Since the PSU started to work perfectly I never gave it a second thought. Looks like its a pretty generic diode similar to the others on the board. If it was a zener or a special diode there is no way to find out now anyway.Any markings on the diode on your board? I could no see any on the failed one.  I'll measure the voltage across the resistors and try to figure the diode polarity (OK from your photo) and see if it is forward or back biased.

I'll put an errata on my original post not to lead people astray.

EDIT: using Pieh0's photo as reference I replaced the bottom resistor with a general purpose diode. From the photo I assumed the cathode band is on the right and that would make it back biased as I get a positive voltage on that side. Comparator protection of some sort?

Thanks for posting the photo Pieh0.
 

« Last Edit: July 25, 2016, 07:51:54 pm by richnormand »
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Just poked the outputs with the scope to see what was happening, and here's the result.

Started with the mains turned off, mains on, then B1 to ground, and then mains off in 10 ish second chunks, and like i said, 3.3v and 5.2v are dead, and A1 starts up, then dies, what i found interesting was the 12v line going to where it is, then crashing to ~8v.

Ignore the little wobble on the 12v just as it went high, fat fingers and tiny wires don't work well :).

3rd picture is the same test, but with a working psu.

So at least from that we know that the 12v dying isn't due to a fault, but something is stopping the psu causing 12v to drop like it does on the working psu.

Time for a cold drink and a think. :P

**edit**

Well i've checked the resistors, and they're 1ohm each, so time to try and find the 5 and 3v lines.
« Last Edit: July 25, 2016, 08:23:51 pm by Pieh0 »
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Just poked the outputs with the scope to see what was happening, and here's the result.

Started with the mains turned off, mains on, then B1 to ground, and then mains off in 10 ish second chunks, and like i said, 3.3v and 5.2v are dead, and A1 starts up, then dies, what i found interesting was the 12v line going to where it is, then crashing to ~8v.

Ignore the little wobble on the 12v just as it went high, fat fingers and tiny wires don't work well :).

3rd picture is the same test, but with a working psu.

So at least from that we know that the 12v dying isn't due to a fault, but something is stopping the psu causing 12v to drop like it does on the working psu.

Time for a cold drink and a think. :P

**edit**

Well i've checked the resistors, and they're 1ohm each, so time to try and find the 5 and 3v lines.

A1 I think is the "power good" line. Its high with the 12V standby (power off) and it looks like it goes down a few seconds later  indicating a problem on the bad supply and stays up on the good one when you power up.
B1 is the soft start.

There are three 12V rails I think. 12V standby generated by the TOP200, 12V for the unit and a 12V for the fan.

Can you monitor what the 12V standby (D24 should be on all the time) does on the bad/good supply. Perhaps the 12V standby goes down and takes the rest. (Unlikely)
Can you follow up the "signal good" line to see where it takes its inputs? that might give you an indication of where the control circuit for the 5V is (intermediate board 2 perhaps)
Can you see voltage on the transformer for the 5V line? The 3V might be derived from it on board 1?

There are two wires connected to a quick disconnect on board 3 going to the big transformer on the opposite board. I think that might be the 5V circuit drive if memory serves.

I assume all measurements were done with the PSU out and without any load? Mine worked OK in that situation (not always true of all PSUs).

« Last Edit: July 25, 2016, 09:18:18 pm by richnormand »
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
The 12v trace was from D24, just checked all 3 12v lines and here's what i got.

The standby is ok, stays up even when A1 drops to 0v.

I'll have a look at the other bits later, the psu is too crowded to be poking around at the moment :)

And yes, no load.

Also probed abcd 1, c&d1 don't goto ~5v, so i think it's clear it's that side of things. I'll have a poke around later.

Thanks,

Pieh0.

**edit**

Followed the "signal good" line until it went into a CD4093BC on board 2.

So now the fun begins trying to find where 5v comes from on board 1.

**edit2**

5v is totally dead, it's not even switching. There doesn't appear to be a 5v tap on the transformer, but basically every other voltage is there (28v & +/-12v).

Traced it back to that little add on board on board 1, it has a UC3843AN smps controller on it, and it's close to where the 5v seems to be coming out of on the working psu.
« Last Edit: July 26, 2016, 03:30:15 pm by Pieh0 »
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Found it!

It seems to be the UC3843AN on the add on board, or something else on it (op amps), cause when i swapped them around, the working psu dies the same way the broken psu did.

So, off to RS to order a pack of them, just in case :).

Wonder what caused it to die.

Going to compare the signals going in and out of both psu's so hopeful everyone else can just shove a probe on pin x to determine if that's the board that's died, seeing as it's buried deep inside the psu.

https://www.fairchildsemi.com/datasheets/uc/uc3843.pdf

There's the comparison of the traces of both board.
« Last Edit: July 26, 2016, 04:59:11 pm by Pieh0 »
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Found it!

It seems to be the UC3843AN on the add on board, or something else on it (op amps), cause when i swapped them around, the working psu dies the same way the broken psu did.

So, off to RS to order a pack of them, just in case :).

Wonder what caused it to die.

Going to compare the signals going in and out of both psu's so hopeful everyone else can just shove a probe on pin x to determine if that's the board that's died, seeing as it's buried deep inside the psu.

https://www.fairchildsemi.com/datasheets/uc/uc3843.pdf

There's the comparison of the traces of both board.

 :-+
Nice to have a comparison between the good and bad unit! I'll be intriged to find out what RS wants for it.
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Found it!

It seems to be the UC3843AN on the add on board, or something else on it (op amps), cause when i swapped them around, the working psu dies the same way the broken psu did.

So, off to RS to order a pack of them, just in case :).

Wonder what caused it to die.

Going to compare the signals going in and out of both psu's so hopeful everyone else can just shove a probe on pin x to determine if that's the board that's died, seeing as it's buried deep inside the psu.

https://www.fairchildsemi.com/datasheets/uc/uc3843.pdf

There's the comparison of the traces of both board.

 :-+
Nice to have a comparison between the good and bad unit! I'll be intriged to find out what RS wants for it.

I mean RS Components for the UC3843AN :P, £5 for 5.
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Found it!

It seems to be the UC3843AN on the add on board, or something else on it (op amps), cause when i swapped them around, the working psu dies the same way the broken psu did.

So, off to RS to order a pack of them, just in case :).

Wonder what caused it to die.

Going to compare the signals going in and out of both psu's so hopeful everyone else can just shove a probe on pin x to determine if that's the board that's died, seeing as it's buried deep inside the psu.

https://www.fairchildsemi.com/datasheets/uc/uc3843.pdf

There's the comparison of the traces of both board.

 :-+
Nice to have a comparison between the good and bad unit! I'll be intriged to find out what RS wants for it.

I mean RS Components for the UC3843AN :P, £5 for 5.

Ha! that makes more sense now. Not being familiar with RS Components my only two options for RS were Rohde and Schwarz or Radio Shack :-//
Hopefully you will get that PSU back in the game.

Interesting, reading the whole thread, is that at least three different issues were identified in what does look to be a well designed PSU, with quality components no less.

My next task will be to make a copy of the hard drive, just in case the software is coded for a particular unit and options!
« Last Edit: July 26, 2016, 10:51:12 pm by richnormand »
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 28141
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Found it!

It seems to be the UC3843AN on the add on board, or something else on it (op amps), cause when i swapped them around, the working psu dies the same way the broken psu did.

So, off to RS to order a pack of them, just in case :).

Wonder what caused it to die.

Going to compare the signals going in and out of both psu's so hopeful everyone else can just shove a probe on pin x to determine if that's the board that's died, seeing as it's buried deep inside the psu.

https://www.fairchildsemi.com/datasheets/uc/uc3843.pdf

There's the comparison of the traces of both board.
Don't jump to hasty conclusions.
These are a common SMPS IC, yes they do occasionally fail but more often it's componentry around them.
The datasheet you linked for it is probably the worst I've seen for lack of application detail.  :--
Here's a better one:
http://www.ti.com/lit/ds/symlink/uc3843.pdf
P8 list a typical application, similar is often used by manufacturers.

Depending on the exact topography used in your application, normal points of failure include:
HV dropper resistor/s (startup supply)
IC supply bulk cap      (IC V+ supply smoothing)
Flyback diodes             (run supply)
Check each of the above.

Here's TI's documents page too, with a heap of application notes that may help too.
http://www.ti.com/product/UC3843/technicaldocuments

 
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
I swapped the board's around with the chip on and the faulty supply worked, so it most likely something on the board its self, or something further up the line killed it, only one way to find out :).

Putting in a new chip will be able to confirm that or not, but if it isn't that, then it will be something else on the board, that's when the fun will start cause i'll have to trail wires out of the psu to feed it while i probe around.

Thankfully there's no mains voltage going into there.

http://imgur.com/a/G5liH pictures of the board.
« Last Edit: July 27, 2016, 03:15:14 am by Pieh0 »
 
The following users thanked this post: tautech

Offline PartialDischarge

  • Super Contributor
  • ***
  • Posts: 1611
  • Country: 00
Hi Pieh,

if for future reference you could measure and post a table with the correct voltages at that connector with the unit in standby and turned on that would be of great help to other users of CMUs/CRTUs whose psus turn faulty  ;)
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Well i replaced the UC3843AN, and that didn't do it, but 1 board works and the other doesn't, as before.

So, the other chips on that board are an LM393M, a CD4001BCM and a TLE2022c.

If it wasn't for the fact that swapping out this mini board made the psu work, i'd be looking further up the line. Hopefully it isn't one of the many transistors on that board.

**edit**

Well it isn't any of them, so not sure what else to look at. Time for a think :)
« Last Edit: July 27, 2016, 06:28:54 pm by Pieh0 »
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Well i replaced the UC3843AN, and that didn't do it, but 1 board works and the other doesn't, as before.

So, the other chips on that board are an LM393M, a CD4001BCM and a TLE2022c.

If it wasn't for the fact that swapping out this mini board made the psu work, i'd be looking further up the line. Hopefully it isn't one of the many transistors on that board.

Since you know which one is bad I would be tempted to put both beside each other, in the same orientation. 
Use the DMM on the diode ohm scale and systematically probe each transistor and diode looking for a difference between the two. Reverse polarity if you don't find a difference and repeat.
Similarly on the input, output and Vcc of the opamps. Just finding a difference may be of help.

Finally do a board to board comparison of resistors and other components.

I realise these are still in circuit so its not really a test of the component at this stage but there "should" be a difference between the two boards somewhere lest the flaw is hidden deep in one of the ICs.



« Last Edit: July 27, 2016, 06:34:21 pm by richnormand »
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 
The following users thanked this post: Pieh0

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Will do, knowing my luck i wont find any problems, then the head scratching will start :)
 

Offline Ericmax

  • Contributor
  • Posts: 14
  • Country: fr
Good evening.

small post to confirm my troubleshooting CMU200.

Probleme: abrupt shutdown of CMU, as if a fuse had burned.

1: Troubleshooting PSU: 12V STB LED flash on backplane> replacing TOP200YN by TOP222YN. ( look on Mouser )

Testing the PSU removed CMU: B1 shunt to GND> PSU works well.
test the PSU in the CMU> 12V LED flash STB> problem on CPU board.
The 12V STB enters directly on a  LM2651MTC-ADJ regulator.

http://www.ti.com.cn/cn/lit/ds/symlink/lm2651.pdf

This regulator is short circuit.
2: Replacing LM2651MTC-ADJ (Mouser or Digikey)

Testing the CPU directly connected to the CMU ( Without LCD, PCMCIA and keyboard ) > CMU works well. ( motherboard post boot BIP)

reassembly of the front panel and replaced on the CMU.

everything is Ok my CMU working again.  :phew: :-+

If it helps users CMU200 or CRTU, hoping to be useful.
« Last Edit: July 27, 2016, 07:19:39 pm by Ericmax »
 

Offline richnormandTopic starter

  • Supporter
  • ****
  • Posts: 674
  • Country: ca
Good evening.

small post to confirm my troubleshooting CMU200.



Testing the CPU directly connected to the CMU ( Without LCD, PCMCIA and keyboard ) > CMU works well. ( motherboard post boot BIP)

reassembly of the front panel and replaced on the CMU.

everything is Ok my CMU working again.  :phew: :-+

If it helps users CMU200 or CRTU, hoping to be useful.
That is great news.  :-+ Enjoy.
On my unit I also had to oil the small fan in the CPU enclosure.
You might want to check it before closing the box.

« Last Edit: July 27, 2016, 09:15:24 pm by richnormand »
Repair, Renew, Reuse, Recycle, Rebuild, Reduce, Recover, Repurpose, Restore, Refurbish, Recondition, Renovate
 

Offline Pieh0

  • Contributor
  • Posts: 34
  • Country: gb
Since you know which one is bad I would be tempted to put both beside each other, in the same orientation. 
Use the DMM on the diode ohm scale and systematically probe each transistor and diode looking for a difference between the two. Reverse polarity if you don't find a difference and repeat.
Similarly on the input, output and Vcc of the opamps. Just finding a difference may be of help.

Finally do a board to board comparison of resistors and other components.

I realise these are still in circuit so its not really a test of the component at this stage but there "should" be a difference between the two boards somewhere lest the flaw is hidden deep in one of the ICs.

Had a probe around like you said, and found a transistor and 2 diodes giving different readings, so i swapped the transistors around, and while it doesn't stay on, it does pulse some voltage through the 5v/3v3 line.

Doesn't help that the diode's aren't marked, and the transistors are in code, but they're simple npn/pnp transistors, nothing special. The one that seems to be dead is marked with 1Bs (
), which seems to be a BC846B.

are the locations of parts.

Looks like the diodes are simple switching diodes, so a bunch of BAQ34's should do it.

« Last Edit: July 28, 2016, 06:12:30 am by Pieh0 »
 
The following users thanked this post: EB5AGV


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf