HP 54502A repair once again

[vanhouten]

Hi everybody,

my HP-54502A quit from working just a few days ago.
When I got it long time ago it was at least booting and after installing a fresh NVRAM (Dallas-Model) on a socket it turned on again and has been working since with the first auto-calibration.
Now it fails again on me in that in won't even boot 99% of the times.

Here are the symptoms I can recognize:
-> Very sporadically it might turn on normally but only to die again after a few seconds.
-> The fan is slightly moving periodically but will not start spinning
-> thers is a clicking sound coming from near the main transformator of the PSU
-> there is a little less audible clicking from the opposite side near the CRT.
-> voltage test points on the PSU show anything but the values proposed in the service manual
    (which recommends replacing the PSU altogether )

I have read some similar behaviour in another post, yet there was no conclusion as to what the actual reason might be.
Does anyone have had similar behaviour or an idea, where to start diggin on this one?
I would like to avoid selling it at a bargain-price because I consider the scope to be mostly intact besides the mentioned flaws.

Any idea would be very appreciated.
Thanks a lot for yout time reading this...

Best regards,
Fab

[tautech]

From the first post in the Repair thread:
http://www.sphere.bc.ca/test/tek-parts/troubleshooting-scopes.pdf

The clicking sounds like a SMPS failing to start. Does it use a SMPS?
Start digging in the PSU's LV & HV

Any Tantalum caps on the PCB's?

[wn1fju]

Yup, sounds like a power supply fault.  I believe the switching power supply used in the HP 545xx series is a Boschert model.  Someone years ago reverse engineered the thing for the 54503A and the hand-drawn schematics are at http://www.radiocollection.be/images/restaurations_img/HP54503/supplyHP54503.pdf.  This supply is used in a lot of HP models, not just scopes.  For instance, I had one in an HP 53310A modulation domain analyzer.  For what it's worth, I was ultimately unsuccessful fixing the Boschert supply I had, which included replacing all the electrolytics (7 of them had leaked goo) and the main switching transistor (several times, I blew them up).  Finally, in desperation, I bought a "non-working" HP 1650A logic analyzer for a few bucks since it had the same Boschert power supply.  But alas, the 1650A actually worked, so after I pulled the supply and verified that my 53310A was working, I put the good supply back into the 1650A.  Bottom line:  I finally bought two Meanwell switching supplies and built my own power supply for the HP 53310A.  Works like a champ, draws only about 80 watts instead of Boschert's 120 watts and cost me less than fifty bucks.

[KJDS]

It's interesting to learn that the 545xx series uses the same power supplies.

[Wuerstchenhund]

It's interesting to learn that the 545xx series uses the same power supplies.

They don't. The 5450xA/54510A/54520A and some other models use the same PSU (Boschert). The 54522A/C and 54542A/C use a different PSU (Artesyn/Computer Products XL130-3630E, now it belongs to Eaton). And if my memory serves me right then the 54510B/54520B (and maybe some other models) use another model of PSU.

[vanhouten]

Hi all,

From the first post in the Repair thread:
http://www.sphere.bc.ca/test/tek-parts/troubleshooting-scopes.pdf

The clicking sounds like a SMPS failing to start. Does it use a SMPS?
Start digging in the PSU's LV & HV

Any Tantalum caps on the PCB's?

Yes, SMPS it is but I haven't examined the PSU thoroughly yet.
Thanks for the tek-tutorial, this is very helpful indeed! And this way the circle closes again since this story began in the first place when I was trying to fix an old Tek-465 that I had lying around for some time. While trying to do so my HP54502 died on me as well and now I am falling back to a Tek-5403 (which is probably older than me) in order to fix them both 
Embarrassing not having read the repair-thread first, I rather searched for keywords and followed this thread due to a matching post...
Of course this post should go into the Repair-Section as well 

It's interesting to learn that the 545xx series uses the same power supplies.

I can only tell for the 54502A and the PSU parts number is P/N 0950-1879 (Boschert Type 700040-01 Rev. K)
It looks quite similar to the schematic of the HP-54503A that wn1fju provided.

Bottom line:  I finally bought two Meanwell switching supplies and built my own power supply for the HP 53310A.  Works like a champ, draws only about 80 watts instead of Boschert's 120 watts and cost me less than fifty bucks.

This could be a second solution of course, but I will throw at least one try at fixing this one.

A first reading of the power transistor housing looks like the attached scope screenshot.
Settings were: 10:1 probe, x: 2ms/div, y: 5V/div, out-dc
There are glitches about 3-5 times per second.

Update (if only for personal use):
Switching the scope on today just cured the malfunction for about 10 minutes. All supply voltages back to normal and a self-test turned out successful. Hooked up a scope to the PSU's power transistor for comparison and the ramp-up in the wave-form covers only a third of the overall signal swing instead of more than half when malfunctioning. Let it sit there running for a while until the PSU got a hick-up again, switching on and off for a few times.
Eventually the scope will turn on again but showing a black screen. Checking the PSU output voltages gives correct readings for the +12V, +3.5V and +5.2V rails, while failing on -12V and -5.2V, referring to the service manual.

Thanks everyone so far, any feedback is valuable!
vanhouten

[Smith]

I just repaired one a few months ago. Sometimes it wouldnt boot or suddenly restart. There was noise visible on the signal too.  Power supply caps had gone bad and leaked all over the PCB. I knew this happend about 5-10 years before, so you can guess the damage. Took me days to clean and repair. None of the caps measured bad, and I couldnt see where the electrolyte was comming from, so I eventually replaced all caps. Now it works again.

I heard some people having issues with the film caps on the primary side. You can check with a variac and boot it at a lower voltage. If increasing the voltage creates more noise, they've probably gone bad.

Its a really nice scope. I just love the user interface. Hardware isn't all that great.

[vanhouten]

Hi again,

this scope turned into a lottery machine. Each time switching it on brings yet another surprise.
Actually the hand-drawn schematic by wn1fju is a good start and mainly accurate, yet incomplete.
I replaced all of the PSU's secondary side 220uF/50V caps. Some of them had leaked, some looked OK but
the whole batch had gone bad.

Interlude:
Meanwhile I got supicious that somewhere there might have been a breach/short between primary and secondary side, which would have rendered the whole endeavour of repairing the PSU meaningless, since the main-board would have been gone for good.
In order to check if the main-board was still ok, I hooked up the supply rails of an old AT-PSU to the main-board. This provides not exactly but at least almost all the voltages necessary to supply the components. This turned out quite well, yet the self-test fails on AD- and DA-converters. All other functions and menus of the scope are working normally.

Back on the PSU there are now some output voltages that seem to be stable at times. Others vary in mean voltage and have a sawtooth-like waveform and/or show glitches/spikes that correspond to the intervals of the ticking noise that comes from the main transformer (ca 5 times per second).

As I am lacking experience in handling SMPS, I am still stumbling through this but at least somewhere in the right direction. The trick with the AT-PSU made me think of giving up on the original PSU and replace it altogether though.

[Smith]

Selftest probably failed due to the fact the voltages differ slightly (5v instead of 5.2v) PSU noise and most units have to heat up at least half an hour before passing all selftests.

[vanhouten]

Sounds plausible to me as well, the AT-PSU misses the required voltages by 0.2V to 0.5V and the 5.2V rail seems to be the critical one here. I am just glad that the main-board seems to be rather intact. When cold-starting it there is a warning on the screen that the interpolator calibration is wrong and I can see the trace slowly dropping from zero downwards.

Even when I think of replacing all the caps I get goose bumps already but the fact that this fixed it in your case makes it eventually appear worthwhile. Thanks for the feedback!
Do you have any idea what the ticking sound on the primary side might be related to?

Cheers,
vanhouten

[wn1fju]

You said: "I replaced all of the PSU's secondary side 220uF/50V caps. Some of them had leaked, some looked OK but
the whole batch had gone bad."

Very amusing, because those 220uF caps were exactly the ones that were leaking goo all over my PSU too!  Must have been a REAL bad batch!

Well, I'm no expert, but I'm not exactly a novice either, and I ultimately was never able to diagnose and repair the fault with mine.

In case you're interested, I used a Meanwell RT-65B to supply 5, 12, and -12V and a Meanwell RS-25-5 to supply the -5V.  That left me short because I still needed 15V and 3.5V according to the label on the PSU.  It turned out that for my application (the HP 53310A), the 15V was used only for the fan, so I simply used 12V and the fan works fine.  The 3.5V was obtained from the 5V through two diodes, and later I found that 3.5V is unused in my application anyway.  I took the top cover of the existing Boschert supply, did a small amount of sheet metal trimming, mounted everything on the cover and slid it back into the target machine.  The only "delicate" thing was that I had to cut off the cable connector (20 pins if I recall) and individually run the wires to a terminal block I also mounted next to the Meanwell supplies.  Whether my replacement supply was as noise free as the Boschert, who knows.  But at least if it breaks, I know exactly how to fix it.

[vanhouten]

Usually I prefer fixing things over just replacing, mostly because of the challenge and learning experience... well sometimes because it is cheaper as well 
But in this case I tend to go for your second solution and order the replacements you suggested (might be cheaper indeed). Did you have any problems with calibration due to mismatching voltages (5V vs 5.2V etc)?
As I said my machine isn't too impressed by the voltages I offer it to feed from the AT-PSU. I hope that the attenuators and converters are still alive and just being picky about that.
So I will try to poke the Boschert some more and see how far I can get until the PSU's arrive.
Thanks for the input!
(If you have more, let me know  )

[Smith]

Mine had a ticking noise too. It was clearly noticeable while the scope was resetting itself.

[vanhouten]

This seems to be a common issue with SMPSs. In the Tektronix-Troubleshooting-Guide mentioned earlier in this post this is described as an indicator to a conclusion that is not further explained  (page 34)
It just advises to check the resistance of the scopes power supplies in case of ticking noises.
I unhooked the PSU and placed a load resistor (10 Ohm) between the secondary 5.2V-output and common. This way two of the output voltages appear to be stable, which tells me that the defect is probably restricted to the secondary side and related to the individual circuitry of the respective rails (this would fit the bad-cap problem you experienced as well).
I will try to examine the working rails more closely then as a reference for the broken ones... if they still work 

[wn1fju]

You asked, "Did you have any problems with calibration due to mismatching voltages (5V vs 5.2V etc)? "

The Meanwell supplies I used all have adjustment pots.  The 5V supply has a range of 4.75 to 5.5V, so getting 5.2V was not an issue.

On my Boschert supply, I replaced ALL the electrolytics, the switching transformer, the opto-isolator, pulled and checked all the flyback diodes, checked all the FET/BJT regulators they put in the back end, and on and on.  I got close at one point, but never completely fixed it.  I keep it around in plain view to keep me humble.

As I said earlier, the replacement supply works well.  My only concern would be whether HP and Boschert engineered the original thing to have certain noise characteristics or switching transients that I am not getting with the Meanwell supplies and that might make the 545xx scope not perform as well.  Perhaps an expert can chime in on this thread???

[vanhouten]

On my Boschert supply, I replaced ALL the electrolytics, the switching transformer, the opto-isolator, pulled and checked all the flyback diodes, checked all the FET/BJT regulators they put in the back end, and on and on.
I got close at one point, but never completely fixed it.  I keep it around in plain view to keep me humble.

This sounds like a lot of fun... that I am probably going to miss.
Maybe my Boschert will get framed an have its place on the wall of shame as well 
Just went through the spec-sheets of the Meanwell PSUs and they read like the most reasonable solution to me.
The service manual of the 54502A does not mention the 3.5V-rail to be of any particular use either.
I think I am going to order a new fan as well while I am at it.
Thanks again!

[Shock]

Some photos of the board would be nice.

[Smith]

As for accetable PSU noise:
Set the scope on the most sensitive level and watch the noise level. You might check the calibration outputs for noise too. I guess noise figures should be noted on the service / calibration manual. If those are not exceed it shouldn't be any problem.

As for the repair, yes repairing this PSU is a pain in the @$$.  Most repairs I heared about failed and ended up with buying a replacement. Thankfully I have access to lots of professional tools, doing this at home would have been a nightmare.

[vanhouten]

Some photos of the board would be nice.

Indeed, sorry for the shortcoming but my lab-camera is somewhat outdated.
There has been a well documented tear-down of the 54502a in this post including the PSU.
https://www.eevblog.com/forum/reviews/hp-54502a-nearly-vintage-digitizing-scope/

[wn1fju]

As described earlier, here's a picture of my replacement power supply for my HP 53310A.  The general layout of the unit (main board, CRT board, CRT, etc.) should be close to the HP 545xx series.  The only part I re-used from the original Boschert supply was the female connector from the AC filter.  I couldn't find a suitable replacement for the output connector so I just tacked extra wires on the cable going down to the main board (after snipping off the male connector).

My only concern is whether or not all the nice wiring acts like antennas and transmits the switching noise.  I suspect it does a little and I suppose I should try to cover the entire supply with some more sheet metal.

[vanhouten]

Looks neat!
The assembly appears to be almost identical to the 54502A.
Here is what I have come up with to jump-start my board in order to verify it still somehow working 

As Smith suggested it would probably be too much of a pain to fix this thing at home, as I have to.
I guess I will get over being such a lousy sport quite soon and realize once more that cheating is just another way to win.

Thanks for all the help so far, guys!

[vanhouten]

Hi guys,

finally back from holidays and meanwhile the PSU's arrived, so time to get back to work...
@wn1fju: could you quickly comment on how you did the rewiring of the MeanWells to the ribbon-cable?
The +V of the right supply is hooked up to the COM-rail of the left supply in order to get the required -5V.
I can see from the picture that you inserted two diodes as well, yet the wiring is a bit tangled, so I can't
figure out to which terminals they actually attach.

Thanks in advance for any hints!
Cheers,
vanhouten

[wn1fju]

For vanhouten:

Attached is a zoomed in picture of the wiring around the terminal blocks.  The big supply on the left supplies the 5, 12 and -12 and the little supply at the top does the -5.  You are correct in that the little supply V+ goes to ground and V- then becomes the negative voltage.  The ribbon cable had 20 wires and I split pins 1-10 to run to the lower terminal strip (from right to left) and pins 11-20 to the upper terminal strip (right to left).  You can see on the masking tape the legend that I used for the terminals.  I simply jumpered along wherever I needed to pick up the proper voltage.  The only other thing I did was to run double wires from the left supply for +5 and G.  The two diodes pick up one of the 5V connections and route it to the proper 3.5V terminal.  It later turned out that I didn't require 3.5V, but I left the diodes in.

My choice of wiring is certainly not the only way to do it.  Just make sure to leave the end of the 20 pin ribbon cable disconnected from the mainboard when you first test the supplies in case you made a mistake! 

In a perfect world, I would have removed the 20-pin connector from the Boeschert supply and figured out how to mount and use that.  I still have a queasy feeling that all my wires are radiating switching noise and compromising the noise floor of the HP 53310A.  I'd be interested in seeing if you suffer any ill effects with your scope. 

[vanhouten]

Thanks so much for taking your time!

I will see how it works out for me and give feedback as things go on.
I think this is probably the most convenient way for anyone having trouble with this series of scopes and
there might be still quite a few of them around.

[vanhouten]

Hi again,

unfortunately installing the replacement supplies did not cure my scope.
Power-up is flawless, yet in selftest I'm being told that the AD/DA section
is broken somehow (see attached picture).
This doesn't look all too good to me and I am close to giving up on this
machine... what a shame.
Did anyone have a similar problem with this scope and in the best case
an idea where to go from here?

Best regards,
vanhouten

[Smith]

Looks like the nvram battery is dead. It' s quite common after over 20 years. You can replace the module or connect an external battery.  I did replace mine with a new one. Google for hp 54502a nvram and there are lots of photos and info to be found.

[vanhouten]

Hmm, I replaced the NVRAM around 3 years ago when I got the scope, so I assume it might still be alright. After a self-calibration it has been working ever since until the PSU gave up on me.
I recall that one can even do a successful self-cal with a flat NVRAM-battery, it just looses its calibration data when the power is turned off.
In calibration mode the display tells me that the connection between the DC-Calibration-terminal at the back and the channel input might not be properly established (which I can rule out).
There is a static notice at the top of the screen "failed interpolator cal"... if that provides any clue to anyone more familiar with this scope than me.
Attached are screenshots of the display in normal operation after power-on and in calibration mode.

Is there a way to check if the input attenuators are still working at all?

Thanks again for any feedback!

[vanhouten]

Update:
I double-checked every connection and found merely by accident, that the 3.5V-rail of my PSU-assembly wasn't stable at all times. Turned out that the ribbon-cable socket had loosened up just enough so that the 3.5V-line was slightly wiggling.
After properly reseating the cable everything looks much better  .

So despite not being mentioned in the service manual at all, on a 54502A all these symptoms are directly related to the 3,5V-rail:
-> failed interpolator cal
-> static faulty offset of scope-trace
-> AD/DA-sections fail in self-test
-> Analog Trigger fail in self-test
-> Acquisition-RAM fail in self-test
-> Time-Base fail in self-test

Just ran calibration and self-test once again and all turned out well!
So for the time being I consider this case closed...

Thanks a lot wn1fju & Smith for your support and sticking around!

Best regards,
vanhouten

[Smith]

I'm glad you got it working. I really like these scopes. Technically they aren't that great compared to modern equipment, but I love the interface.

[wn1fju]

For vanhouten:

Glad it works.  I would be interested in knowing if you see any noise problems attributed to the replacement supplies (especially if you used the same two Meanwell supplies I did in my 53310A). 

If you could cap off the inputs, set the timebase for something reasonable for a 50 kHz to 100 kHz signal, and turn the Volts/div settings to their absolute minimum, do you see any transient spikes?  You probably won't be able to trigger on any of them.

In my 53310A, which is a modulation analyzer, not a straight oscilloscope, I swear I am seeing some artifacts in certain low frequency scenarios.  But since I didn't have the unit operational with the original Boschert supply for very long, perhaps they were also there and I just didn't notice them at the time.

[manneke]

Hello,

If there is still aneed for 5450x diagrams then a free set can be found at  www.radiocollection.be

search thru the repair reports ot the author

good luck
Gilbert