[SOLVED] Blanking issue on Hameg 203-7 oscilloscope

[shakalnokturn]

After not having any luck here asking about a HM203-6 with blanking problems I didn't bother asking for this one... As nobody here seems to be into oscilloscopes I thought I'd give the conclusion away.

Symptoms were blanking being too slow and showing bright spots to the left of the trace for first 10 minutes on a cold start, slowly improving to normal after that.

Troubleshooting was a PITA because 2kV from ground and heating the PCB to find the problem also causes the same symptoms as when cold. (There's only a small temperature range where the scope behaves normally.)

In the end the SMD rework gun allowed to pinpoint the culprit which was C314 (68pF 3kV ceramic) of course it measures as good out of circuit. I have yet to test it with an isolation tester.

[Greybeard]

Thanks for reporting your repair.

...the culprit which was C314 (68pF 3kV ceramic) of course it measures as good out of circuit.

In your text you wrote the culprit was C314.
In your photo you marked C317, and in the schematic you marked C315.
Each one is 68pF 3kV.

Which one was the defective?

PS: To make a long stort short: C314 

[shakalnokturn]

Thanks for pointing that out... it doesn't make much sense! It must have been late by the time I finished that repair.
Although I no longer have the scope to check I beleive it was the one in the photo: C317.
Hopefully there are no errors between the PCB numbering and schematic although that would give me an excuse for one mistake.
The other way to read my report would be: "change all the 3kV 68pF ceramics" 

[Greybeard]

The other way to read my report would be: "change all the 3kV 68pF ceramics" 

Of course I ordered 10 pieces of them a few minutes ago.
I just wanted to understand...

[Greybeard]

I think the caps C302 and C303 (both 0.22uF / 1kV) may be "a little" underdesigned:
Nominal voltage is 1000V <-> voltage in operation is (450+1972)V/2 = 1211V !
The +500V point is slowly falling from first 490V to ca. 450V after an hour.

What do you think?
Should I  replace C302 and C303 by 1.5kV or 2kV types?

[shakalnokturn]

I'd never really paid attention to that, I'd have to check in one of my scopes, they may have used 2kV parts there on others.
I've never seen any of those fail yet, not sure on the voltage ratings I've come across though.
I'd probably measure them and leave as is if values are close enough.

[Greybeard]

I checked the caps in my HM203-7 and HM203-6:
All 4 caps in the doubler/quadrupler circuit are 1000V types, in HM203-7 and in HM203-6.

Nominal -1980V point (-1972V real) is stable on DMM (2 identical DMMs in series ).
Nominal +500V point (490-450V real, with ripple 30Vpp, f=100Hz) is falling slowly (by self-heating?) and fluctuating randomly about +/-1V  on DMM.
The +500V point is not used as supply voltage, it is just part of the -1980V regulator.

Is this behavoir normal or is something wrong with it?
What is causing the fluctuation?

[shakalnokturn]

I'd say the 30V / 100Hz ripple on the +500V is normal, that's precisely what the regulator acts on to maintain the -1980V stable. Heating would be a good explanation for the slow decrease.
As for the +/-1V not sure could be the variations caused by beam on/off, I doubt it matters anyway.

What is actually wrong with your scope?

[Greybeard]

What is actually wrong with your scope?

Blanking does not work. No blanking while waiting on trigger.
Or no blanking at all (beam return trace is slightly visible).

Input pulse on opto coupler looks OK. I measured many resistors around the opto. Swapped opto, BF199 and BF440 - without success.

Desoldered the three 68pf/3kV caps, measured them OK (>80pF, D~0.06, no high voltage test).
Reseatet the three 68pf/3kV caps in another sequence, because I have no new parts (but ordered them). Error still exists.

The screen looks like yours or this one:



This is the variation at +500V point (take account only of the top peak line, the band below is caused by the 100Hz ripple overlay):

[Greybeard]

Comparison between HM203-7 and HM203-6:
Variation at +500V point looks nearly the same.
Voltage at +500V point is +730V 

[Greybeard]

Difference between HM203-7 and working HM203-6:

HM203-6: opto coupler diode (Pin 1) shorted to GND -> beam blanked
HM203-7: opto coupler diode (Pin 1) shorted to GND -> beam NOT blanked

Blanking is not working at all!

I got my 2.5kV oscilloscope probe today! 

So I tried to measure the HV side of the opto coupler.
Both, emitter and collector is contaminated by much noise, looking similar as the variations at +500V point. I think the cathode and wehnelt voltages are floating (common mode).
I was referencing to chassis GND, what is probably wrong.

I'm wondering why the nominal -1980V point (-1972V real) is rock stable on DMM.
Why these fluctuations when watching it by o'scope?

I need some time for thinking about...

[Greybeard]

I tried replacing the opto coupler again.
Same result as expected: no blanking.

I found out the photo transistor is conducting, independent from driving the opto LED.
So it must be caused by faulty driving the base of the photo transitor (pin 6).
I pulled the opto out of the socket, bent pin 6 out of line, and reseated the opto:
Shorting pin 1 to GND blanks the trace, function OK.

So only one component could cause this: C314.
I removed C314, bent pin 6 back in line and reseated the opto:
Shorting pin 1 to GND blanks the trace.

Measuring C314 did not show any failure:
C~82pF, D~0.06, R>500MOhm, no high voltage test.
The strange thing is, this is the cap that was C315 before, because I intentionally "rotated" position numbers.

May be all of these 68nf/3kV are leaking... (PS: probably they are!)

[shakalnokturn]

Regarding HV probing I assume that DMM was completely floating.
HV probe is linked to a scope with ground reference DUT is also ground referenced, this should not be a problem though as long as probe ground clip is on scope under test's GND.
Also consider that a DMM averages out quite a bit depending on frequency, is it that you're just seeing extra detail now?
As for swapping defective capacitor's position the odds are it could have almost exactly the same effect in another position...

So is the conclusion as in first post: Change C314, C315, C317, no the lot? 

[Greybeard]

Last check: cross-over test

I pulled the 68nF/3kV cap out of my working HM203-6 and installed it into my defective HM203-7:
it works correctly.

Then I installed in the suspicious 68nF/3kV cap into my formerly working HM203-6:
same failure as in HM203-7 before.

C314 failure confirmed.

I'm just waiting for the ordered new parts... 

So is the conclusion as in first post: Change C314, C315, C317, no the lot? 

Now we know... 

Thanks for your help and for your time.

--------------------------------------------------------------------------------------------------------------------
PS: Does anybody know, how to adjust the "wired cap" (= more or less twisted pair of insulated wire) in the timebase circuit?

[CaptDon]

Multiple scopes with the same problem. Sounds like under-rated or inferior components from day one. Think of all these scopes that will fail and just end up as scrap. It is so nice that the Chinese have offered to fill up those unsightly and dangerous holes in the ground we call 'landfills'. I remember that a company I worked for decided to buy the Chinese knock-off of an Agilent scope. It was an obvious clone and folks in the know said it was made in the Chinese factory that made the Agilent product. These were obvious QA/QC failures not acceptable to Agilent that got the knock-off brand placed on them. We had a total of 3 exchanges with not a single one working correctly. We finally gave up and bought a 'Real' scope. To hell with those knock-off turds. Wonder which landfill our three turds ended up in? Probably India!!

[Greybeard]

I think HAMEG oscilloscopes are very reliable, respecting they are > 30 years old.
At least you can repair them using standard components.

[shakalnokturn]

No clue for the wire capacitor, I'm not expecting it to have much noticeable effect on trace either... It looks like a very fine adjustment for sync between TB start and blanking, I expect you'll only see the difference at X mag. x10 and maximum sweep speed at start of trace.

I too think Hameg make pretty reliable scopes, design isn't fantastic for serviceability though.

[CaptDon]

We had some Kikosui scopes that worked well (when they worked) but had board inter-connect failures almost monthly. I finally took out the connectors and hard wired the inter-connects. The scopes were somewhat lower priced with good features and worked decently until the C.R.T.'s finally faded into oblivion. We also had some of the TEK 2213 style scopes and ours missed the recall/service update for the C.R.T. filament overvoltage correction and they faded to black also. Our old TEK digital scopes TDS420 and up had power supply / RIFA capacitor / and bad SMT electrolytics on the signal boards. No where up to the old bullet-proof standards of the tube type TEK gear from the 60's which still work today!!! My old tube type Waterman 3" triggered sweep portable military scope was bought at a hamfest for $25 and has worked flawlessly for about 40 years. I think it has the original tubes in it. C.R.T. is a 3WP1A with a super sharp trace. I think it is a Dumont.

[Greybeard]

Finally I compared old (leaky) and new C315 using my curve tracer:

[CaptDon]

Wow, those V/I curve tracers are far more handy than most people would believe!!! That is a world of difference between the bad part and the good part!!! Nice catch!! We use one of those V/I curve tracers to troubleshoot locomotive circuit boards that receive signals from the trainline or other remote inputs. It has helped us diagnose failures caused by arc-overs, reverse battery jumps and even lightning induced EMP damage when the locomotives are sitting unused in rail yards!! Very nice unit to add to any bench and can easily be made with a signal generator and oscilloscope.