HP 54510B - SMPSU repair /strange problem after recapping

[Electro_Alex]

Hello,

I have (another) strange problem with a Borschert power supply and I cant find a starting point.
I bought a nearly functioning scope on Ebay(just one defective attenuator) and the scope worked for one week.
Suddently it cycled on and off for five seconds and I realised that this would be the known problem with the power supply. I disconnected it from 230V mains as soon as possible.

As I opened the SMPSU i could smell the typical electrolyte smell and I found two small places where the electrolyte leaked into the board.
Not bad like some people showed, so I decided to do a full recap(31 caps/ Panasonic).
The fuse was not blown and there was no visible damage at other parts.
Under C27 there was a dark coal like spot, so I drilled this spot out until i could see that the board material looked normal again(10mm). The drilled out trace under C27 was later reconnected with a wire.

After I was ready with the full recap (including the caps on the controller board), the power supply now does not work normal. The voltages are totally off. I can hear a strange clickling sound (like a bicycle wheel) which gets slower when i unplug the mains connector. It really sounds like if a bicycle wheel with a freewheel is spon, but much more quiet.

It does not really make a difference if I load it with 2Ohm/25W load on the 5V rail.
The voltage above the 2Ohm load drops to 0,9V permanently, but nothing gets hot and the fuse is intact. The 0,9V seems to be a pulsating voltage, so I guess I maybe have a problem with the controller board.

I got the following voltages(unloaded)
4,9V/8,0V/-5,4V/0,0V/2,2V/2,2V

When I connect the psu to the scope, nothing happens and I get an even louder clicking, which seems to come from the motherboard. Maybe the speaker clicks in the same rhythm as the psu cycles.

I measured the voltages right after I bought the scope and everything was fine:
11,9V/12V/-11,55V/-5,25V/3,525V/5,183V

I also checked the bridge rectifier and it seems to be ok.
I can measure arround 150V over the filter caps.
On CNY21N (optocoupler) i can measure 4,9V DC/ 0V unstable/pulsing from C to E (unloaded) and 4,2V pulsing with a load.
There is 0V from A to C.
On LM339AN i can measure 5,8V from Vcc to GND and 5,2V with a load.


Hopefully somebody can help. I really want to try to save this nice scope.

Best regards,
Alex

[xavier60]

There is a link to a reverse engineered schematic here, https://www.eevblog.com/forum/repair/help-needed-to-find-why-hp-54510a-oscilloscope-won_t-boot-or-stay-on/msg765869/#msg765869
SMPS ticking is usually because of out-of-range rails causing shutdown or crowbar protection.
Or controller section not being able to transition from start to run state.

[DeepLink]

The "bicycle wheel sound" could be an indication of an overload situation - hic-up mode, PSU trying to start up into a short

The -5,2 Volt rail is measured to 0 Volt
As the 5,2 V rail has it's own regulation (not relying on any of the other voltages) you might want to look there for a short

Re-check all capacitors for correct polarization

Don't know if the attached schematic is correct for your model, but this PSU was used in many HP instruments of that age

[Electro_Alex]

Thanks for your replies. It helps a lot with not giving up this thing:)
I found the reverse engineered schematic and followed the -5,2V rail.
I removed CR38 and CR27 and tested again.
The clicking changed to a louder and now sharp interrupted clicking with long breaks when switched off like a stopped „bicycle clicking“ with breaks of 0,5-1 seconds.
It seems to come from the big main transformer T2.

The voltages are:
0,8V/8,5V/-4,4V/0,0V/0,7V/2,8V
(Left to right from TP7 to 1, same sequence like before)
The LM339 has now  no supply voltage (0,3V).
It does not make any difference if the 2 Ohm load is conected or not.

I can measure 0,07Ohm DC resistance between the pins 5 and 6 of T3(CR38 and CR27 removed) and 1,1V AC ouput voltage. Are these values ok?

[xavier60]

I cant see what powers the LM339. Check to see if the crowbar SCR is being fired.

[Electro_Alex]

Ok, where and how can I check if the crowbar SCR is fired?
Is the crowbar SCR on the controller board?
I can see more than one thyristor (I guess the thyristors have a silver casing? they are all marked with CR including all diodes).

It seems that there is a short to GND on the -5,2V rail, because I can measure 0,07 Ohm to GND (Pin 11/18 to Pin 5/6/7/8). Is this plausible?
Where should I look for the short?
Q14 seems fine, CR23 too and even if not, this would not explain a short on the -5,2V ouput, I think.

I cant find Q11 on my board. I have a later version of the Borschert psu and there are some differences, for example no adjusting potentiometers on the controller board.

[xavier60]

The SCR is shown as A2 in the schematics, likely a power device like TO-220. Best way to check is looking at the Gate with a DSO if a working one is available.
First, the short on the -5,2V rail needs to be found.
Look for solder bridges. May need to remove all capacitors on the rail one at a time.

[Electro_Alex]

Thanks a lot! Now it seems to go into the right direction.
After removing C15 and C16 the short was still there, but after removing C19(50V/100uF) the short was gone. I may introduced the short be myself as I reconnected the GND lead of C19 because I lost a very thin solder ring on the solder side of the PCB. Iam still not sure what is going on here, but without the short on the -5,2V rail and with C15/16/19 removed I got the following voltages(the same with and without load).
0,8V/8,4V/-4,4V/-0,15V/0,7V/2,9V
I removed more of the silicone glue on A2 and I saw that it is a C122F1 silicon controlled rectifier (TO-220).
The C122F1 now gets warm (46 degree celsius after two minutes).
Is this a sign that the crowbar is active?
On pin 3 (gate) I can measure a pulsing signal.
I just connected the signal wire, because I didnt know to what GND I should connect the Mini DSO to.
This way I got the best signal.

 

[xavier60]

The crowbar seems to be repeatedly firing. Use the 0V out for the probe ground.
When a rail has a low voltage, it could be because it's not coming up to the correct voltage or going over voltage briefly then dropping again.
Check all of the rails with the DSO. The +15,5 rail should show an instant drop as the SCR fires.

[xavier60]

If  C15/16/19 are still removed, put them back and check for the short again.

[Electro_Alex]

Yes! The scope is alive again.
After plugging in with the load plugged in, the crowbar switched immediately off and all voltages where nearly there on point.

My mistake i self introduced was really easy to overlook.
There was a copper ring that came off after desoldering C19 and there was another copper plane directly near the ring that came off. It looked like that these copper planes originally were attached to each other but they were not.
I made photos of the pcb so so could check it.
After finetuning the voltage to 5,0V the scope cane back to life.

Thank you all for your help!
I think the extra motivation helped to solve the problem.😀

Unfortunately, the scope now does not show any input signals on both channels.
I know that one attenuator had a little noise and that this could be caused by a bad input relay connection which may could be fixed by desoldering the coils and cleaning the gold contact spring.
But now I got completely no input signal (on the bad and good channel).

I tried a the cal factor unprotect and key down power up. No success.
If I try to do a probe calibration I still get the error „Attenuation too high or bad connection.“
Selftest still passes completely.
What could be going on here?

Is there a way to measure how far an input signal gets trough?
(maybe like feeding an input signal directly to the attenuator input headers, nearly directly to the A/D hybrids?)

[Electro_Alex]

Both outputs (DC cal and AC) are working, so on this side everything is ok.
The 0V and the 5V DC voltages were perfectly fine on DC cal out and there was a nice 482Hz (500Hz) square wave output signal on the AC output.

I thought about a dead lithium coin cell and removed the mainboard to check.
One of the 6 330uF caps had a little shiny drop under the side/edge where it sits on the pcb.
I checked continuity to GND and saw, that under this cap both pins (+ and GND) had continuity.
One pole 0 Ohm and the other 17 Ohm.
I desoldered this cap and the 17 Ohm between both poles were still there (no visible damage on the pcb/ cap was a sealed cap and not the standard rubber sealed cap type).
Does anybody know if this is normal and can verify this?
Maybe there is a resistor (wire resistor?) above these poles which could explain this?
The cap is located just above U72.

[Vovk_Z]

I don't see anything suspicies on this photo.
If you desoldered a cap then meadure its capaciance value.
I would measure its ESR too if there is a tool.