HP 1653B Logic Analyzer/Scope- PSU in bad shape

[valley001]

Hi everyone,

I picked up a HP 1653B for almost nothing as a project and intro into logic analyzers.   Its in remarkable condition aside from some yellowing on the disk drive.  I have a boot disk but no probes/pods yet I figure ill look into those once I have it working. 

The unit would not power on when testing with a dim bulb tester, no fan, nothing.  I traced the issue to the PSU which was not outputting voltage on any test points.  Disassembly of the PSU module revealed what I am confident is the primary issue....some/many electrolytic caps took a dump all over the board.  Literally its soaked in electrolyte lol.  Its a miracle this is contained to the PSU module, the main board looks totally fine, no electrolyte dripped onto it.

I have researched this supply (PN 0950-1879) and read the threads here, it seems this is a common problem.  I also have the hand drawn schematics. 

Id like to run my repair process by you all, I welcome suggestions on things I should check before applying power again.  I am aware this is a SMPS and there are line voltages present on the board, I take appropriate precautions when testing.  Thanks all.

1) Remove all old electrolytic caps. 
2) Soak the board in isopropyl alcohol, agitate, and rinse. 
3) Potentially repeat step 2 until clean.
4) Install new caps. 

[SoundTech-LG]

Yes, good start. Remember though that these leakers can be insidious...  I had one that actually wicked up into the transformer windings. Both transformers became leaky, and would blow the switching transistor... Replacing them cured that issue. Another had the single pot for voltage adjustment near the outer heatsink  was contaminated and became intermittent. These are some of the other spots to be aware of. I am sure there are others too. I had one leaking so badly the electrolyte was running down the output ribbon cable to the main board.
The switching transistor is not readily available. I substituted a Horizontal output transistor for a multisync monitor of appropriate specs with success. Good luck. You should be fine though.

[SoundTech-LG]

I marked the trim pot location. Also the red area I circled looks strange to me... like it's corroded or something.

[valley001]

Yes, good start. Remember though that these leakers can be insidious...  I had one that actually wicked up into the transformer windings. Both transformers became leaky, and would blow the switching transistor... Replacing them cured that issue. Another had the single pot for voltage adjustment near the outer heatsink  was contaminated and became intermittent. These are some of the other spots to be aware of. I am sure there are others too. I had one leaking so badly the electrolyte was running down the output ribbon cable to the main board.
The switching transistor is not readily available. I substituted a Horizontal output transistor for a multisync monitor of appropriate specs with success. Good luck. You should be fine though.

Man I hope this is not the case.  There is a lot of electrolyte but most of it seems to be in the vicinity of the 1000uf 35v caps, a lot of those look guilty.  2-3 in the cluster of 8 were pretty bad, and in a couple other places.  Some of the 220uf look bad too.  To the extent I can I plan to test the transistors and diodes after cleaning it up.  Its funny the replacement caps are going to cost far more than what I paid for this thing. 

There is a lot of glare in the pic below from the electrolyte. 

All the caps off now, cleaning tomorrow. 

[wn1fju]

I have worked on three of these Boschert power supplies since they were used in several of my vintage HP pieces (scopes, logic analyzers, modulation analyzers, etc.).  I managed to fix two of them - I could never get the third one to work properly.

You are on the right track.  Scrub the board (toothbrush) with alcohol, rinse, repeat several times.  I would also pull the inductors and clean underneath them.  The two main capacitors on the primary side are probably OK, but I would also replace the rest of the caps on the primary side even if they don't appear to be leaking (maybe they will leak in the future).  Spot check the diodes and transistors on the secondary side - the diodes are probably all OK, but the transistors maybe not.

As SoundTech-LG said, the trimpot at the rear of the board (near the three transistors on the heat sink) can easily be destroyed.  I also had a problem with the SCR on the control daughterboard.  And verify the opto-isolator is functioning. 

As a general rule, once you have the new caps installed, if you have a Variac, bring the power supply up slowly watching the current draw.  If things are still amiss, you don't want to pop the switching transistor.

[SoundTech-LG]

Wow!!!

Your latest photo really says a lot! That has to be the worst, or at least, one of, I have seen. Pulling up the transformer is not easy. It has large copper strips that come out and are soldered directly to the ground plane for the group of 1000uF caps that leak.
You will need a 100W+ iron to make your job easier. The transformer has 4 plastic spacers one on each corner... easy to lose! How do I know? 
Before you pull up the transformer measure with your DVM and look for a very high resistance between the HOT and COLD side windings. The higher the better. Far higher than 5 meg for sure. Also, with that level of electrolyte damage, also check the +5 V adjust  trim pot near the transformers at the edge of the board. Where did you buy your caps? I bought a bag of 50 off Digikey that were under 1 USD ea. if I remember correctly.

[valley001]

Its not looking good.  With the T2 still in place I am getting 0.6 and 0.3 meg between the primary and secondary side.   I have not cleaned the board yet so perhaps the electrolyte is corrupting the measurements?  Is there any problem with soaking the board with the transformers in place?  Im using a small plastic storage bin for cleaning, the plan was to fill it with iso until its just covering the PCB, then agitate and rinse/repeat. 

Where did you buy your caps? I bought a bag of 50 off Digikey that were under 1 USD ea. if I remember correctly.

One of the prior threads someone posted a spreadsheet for the caps with digikey part numbers, some not available now, but with subs total comes to around $50 plus shipping for the caps.

[SoundTech-LG]

Yes that seems very low for HOT to COLD side. Sorry to say, but until you find some other reason for that, it is likely you will want to pull the transformer and check it, and then check again with the large choke still on the board. The choke is much easier to remove.

Clean that board up first, the best you can!!! Maybe you will get lucky!

[valley001]

I decided to remove the transformer, which wasn't too bad really.  Under the transformer was a ton of electrolyte that the initial cleaning did not remove.  I removed it, and now between the empty pins hot-cold side it reads in the 5 meg range. 

The transformer itself also measures high resistance between the primary and secondary (hot- cold).  What I am confused by is the resistance readings I get on the windings them themselves, I get a super low resistance less than 1 ohm on the windings.  I am not really familiar with switching transformers, is this normal? 

[valley001]

Some pics after the first round of cleaning.  I am concerned about the corrosion around the daughter board pins, im afraid there could be leakage between them.  Is this a justified concern or should I just measure and make sure there isn't some low resistance bridging?   

[SoundTech-LG]

Don't worry about the low winding resistance that is normal. you need to pull the other transformer and do the same cleaning and testing. I would try to clean more at t daughter pins too, unless you think you can remove it safely. Replace those trim pots.

[drhex]

I had a board wick up the electrolyte into the board itself and start burning on first power up. May make sense to lift the pins and look closely at all areas where the solder mask doesn't cover the board material itself.

[dl6lr]

I picked up a HP 1653B for almost nothing as a project and intro into logic analyzers.   Its in remarkable condition aside from some yellowing on the disk drive.  I have a boot disk but no probes/pods yet I figure ill look into those once I have it working. 

One warning about the disk drive and the boot disk: I would not put the disk ever into the drive until confirmed the drive is working OK. And if possible make a backup of the disk with either Libdsk or ImageDisk or Teledisk on a MSDOS computer first. The image may come handy if you find the drive is borked. I know the TeleDisk imaging is working, as I used some images that were taken with TD, I suppose ImageDisk by Dave Dunfield is working too. LibDsk by  John Elliott should work fine too, I have not verified. Raw data images (that of the sector contents) are not working, you need the sector meta data too, as HP uses some kind of copy protection with wrong sector index numbers or track numbers.
I acquired a -hp- 4952A that uses the same Sony MP-F52W drive. The drive was bad and it scratched the surface from the disk. I found out when trying some unformatted disks in the drive.
The drive can not be replaced by a usual floppy drive from a PC, as it spins with double the speed. The drive is nearly unobtainium, but it can be replaced by a Gotek floppy emulator and some special wiring harness.
See here https://www.rudiswiki.de/wiki9/HxCFloppyDiskEmulator for some information, I have successfully replaced the drive in my protocol analyzer with a Gotek including the disk change and change reset signals, all is working fine with images of the disks on a USB stick. Feel free to ask for more details if you are ready, i am still in the process of finalizing my repair report.

[dl6lr]

I decided to remove the transformer, which wasn't too bad really.  Under the transformer was a ton of electrolyte that the initial cleaning did not remove.  I removed it, and now between the empty pins hot-cold side it reads in the 5 meg range. 

The transformer itself also measures high resistance between the primary and secondary (hot- cold).  What I am confused by is the resistance readings I get on the windings them themselves, I get a super low resistance less than 1 ohm on the windings.  I am not really familiar with switching transformers, is this normal?

I would remove all transformers and put the whole board into the dishwasher before cleaning  with alcohol and drying.

[SoundTech-LG]

I have a couple of those floppy drives if someone is looking. Did not know they were unique.

[valley001]

Hello all,

Thanks for the input.  I may do the dishwasher suggestion as I plan to remove all the transformers anyway.  I am not sure about removing the daughterboard but I may attempt because the electrolyte did flow through there. Its really a shame. 

Results of diode and transistor testing (in circuit):

-The SCR on the daughterboard has very low resistance between all three pins so I believe it is bad.
-All other diodes and transistors seem ok, some had low resistance between some transistor pins but this could be traced to a low value resistor between them.

I wonder would the SCR have taken out any other components? 

Re the trimmers, does anyone have specs on the originals?  The two in question, R40 and R58 I can not discern any writing on them.  I have 1/2 watt multiturn on hand in many values

So moving forward I plan to:

1- Remove other transformers and continue to clean.
2- Possibly remove inductors
3- Possibly remove daughterboard and clean
4- Replace bad SCR
5- Replace R40 and R58 trimmers.
6- Install new caps
7- Reinstall transformers
8- Test 
 

[SoundTech-LG]

Certainly is a lot of work. I may have a spare board to check your SCR etc. for you. Finding a spare module might be difficult, and the price high. I found one received it, and the main heatsink and case had been removed for the "valuable" aluminum. Got a considerable refund, it made the purchase worth it.
Used the transformers. I'll see if it's handy to check...

The other possibility is finding a cheap scope or logic analyzer with a good supply.

[SoundTech-LG]

If you intend to "dishwasher" that pcba, I would suggest an oven drying for several hours to get any moisture out...

[valley001]

I dont mind the work.

If you did find that board and don't mind comparing measurements that would be awesome.  I discovered a tantalum cap wired between two of  the SCR leads so maybe that's what's bad.  It looks like ill be removing that daughterboard so I can properly work on it. 

Certainly is a lot of work. I may have a spare board to check your SCR etc. for you. Finding a spare module might be difficult, and the price high. I found one received it, and the main heatsink and case had been removed for the "valuable" aluminum. Got a considerable refund, it made the purchase worth it.
Used the transformers. I'll see if it's handy to check...

The other possibility is finding a cheap scope or logic analyzer with a good supply.

Re the dishwasher idea, is the electrolyte water soluble?  I've not washed a PCB in a dishwasher before but I've also never had caps leak like this.  It would be desirable to remove as much electrolyte as possible but id not want to introduce other issues from moisture.   

 

If you intend to "dishwasher" that pcba, I would suggest an oven drying for several hours to get any moisture out...

[valley001]

I carefully removed the daughterboard (no lifted traces!), the pic shows what was still left after several scrubbings prior to removal.  It was simply not possible to remove this contamination without removal of the board.  For fun I measured resistance between pins before and after cleaning, and there was quite a difference before and after.

The SCR tests fine out of circuit, there is a 22 ohm resistor between the cathode and gate which is the source of the low resistance.  The tantalum cap seems fine also.

[SoundTech-LG]

Ok...  finally found my spare parts board. R58 I used, so I can't verify that one without disassembly of a complete module. Sorry. R40 (5V adjust) is 100 ohms, it is measuring around 56 with glyptol holding it set in position.  I concur with you on the TO-220 marked A2 on the board and the tantalum. They read short, or low resistance. By the way R58 must be near the calibration or the module won't start. I would set it 1/2 way, then cal. Not sure if the electrolyte is water soluble, especially after its dried, and corroding things.

[SoundTech-LG]

Great work removing that board! I have not attempted that, as no corrosion there.

[wn1fju]

Great work removing that board! I have not attempted that, as no corrosion there.

It's not hard if you have a suitable desoldering gun like the Hakko FR-301.  Just apply a little solder to all the pins to stir up the old solder and hit it with the gun.  Took me about 90 seconds. 

Anyone doing a lot of vintage (non-surface mount) repairs should have such a tool.  I consider it one of the best $300 purchases I ever made.

[valley001]

After seeing how much crud was under the daughterboard I decided to remove the output connector. Same story here, lots of gunk.  Some before and after photos 

Im nearly done with the cleaning.  I'm gong to lift and clean under the inductors and a couple of the film caps, then one more scrubbing and I'm going to call it good. I couldn't find any other issues, I guess the caps going are what killed it    I guess ill find out lol. 

Ok...  finally found my spare parts board. R58 I used, so I can't verify that one without disassembly of a complete module. Sorry. R40 (5V adjust) is 100 ohms, it is measuring around 56 with glyptol holding it set in position.  I concur with you on the TO-220 marked A2 on the board and the tantalum. They read short, or low resistance. By the way R58 must be near the calibration or the module won't start. I would set it 1/2 way, then cal. Not sure if the electrolyte is water soluble, especially after its dried, and corroding things.

Thanks very much for this information, very helpful.  I was going to take the resistance set on R58 and pre set the new pot to that.  Hopefully this is good enough.  I get some 350 ohms on R58 so ill install a 500 ohm 1/2 watt pot there, sound reasonable?   I may just leave them in for now, to limit the variables ive introduced (there are so many now lol), the replace after its working. 

[SoundTech-LG]

You jogged my memory a bit... I do think 500 ohms is the correct value.
I think you will be pleasntly rewarded on your first power-up. What persistence!
You have done more than probably anyone else to get that power supply operational again!

[dl6lr]

Re the dishwasher idea, is the electrolyte water soluble?  I've not washed a PCB in a dishwasher before but I've also never had caps leak like this.  It would be desirable to remove as much electrolyte as possible but id not want to introduce other issues from moisture.   

Most electrolytes are water soluable. You have to use polar solvents to get good results. Water, soap and brush are very effective. Isoprop is effective if it is i.e. 70%. See some remarks here: https://www.eevblog.com/forum/beginners/what-chemical-to-clean-the-pcb-after-leaking-capacitors/

I found the dishwasher to be very effective (but don't let your wife find out), it uses softened water.

[valley001]

I tested all the caps I removed from this power supply and found all but two of the 1000uf 35v caps were fubar, ESR in the hundreds of ohms.  Im guessing those are the leakers. The rest all tested ok.  Still going to replace them all.