Hameg 203-7 CRT beam problem although everything seems fine

[cascador]

Hi everyone,
I have a Hameg 203-7 built in 1992, with a beam problem.
Documents and service manuals can be found here:
- https://cdn.rohde-schwarz.com/hameg-archive/HM203-7_deutsch.pdf
- https://cdn.rohde-schwarz.com/hameg-archive/HM203-7_english.pdf

Before posting, I have been doing a fair amount of troubleshooting, including:
- I tested all high voltage CRT input pins with a high voltage probe,
- I traced the X1,Y1,X2,Y2 CRT input signals on another oscilloscope (i.e. after amplification, but I also tested the signal before amplification)
- I checked all the voltages, the sawtooth signal, tested the transistors (those in sockets) on a transistor tester, the blanking circuit, and many other things for almost 2 days
Conclusion: everything is fine !! Nice X sawtooth, blanking at the right time, Y voltage has the same shape as the input signal, Y1 and Y2 change accordingly in opposite direction, brightness and focus are ok, etc. But the beam jumps all other the place...

I just posted a video on yt so that you can have a look at the beam problem with the square calibration signal as an input on channel 1 : https://youtu.be/X0FR12LhiH4
If you have any idea I would be grateful.
For the moment, my conclusion is : dead CRT although I can't subscribe to this...

Thanks !

[GGMM]

hello,

Looks like the sweep is falling off like the trigger’s on?
Hard to say, see what happens on opto can be.

cdt
 

[shakalnokturn]

Holá cascador.

I just don't know what to think of your video... It looks a right mess.
Do you know the scope's history?

Can you give some details on how the scope behaves in X-Y mode?
Can spot be centered? Be focused? Stable? Arc noises?

Sometimes when all else fails leave it in a corner for a couple of weeks, when you return to it, it becomes obvious the CRT had started letting air in 

[cascador]

Thanks for your remarks.
The trigger does not seem to have any influence on the beam...
I tried XY mode, and I could get a point close to the center, or even a vertical line appearing sometimes. Not very stable nor straight though...
It's hard even impossible to stabilize the beam.... I retried, and could not get it another time....
Each time the beam appears there is a "knocking" noise when time is slow (higher duration per division).
Focus works.
I tried to let it in a corner for many months before and nothing changed unfortunately !
I bought it used I don't know its history... But no rust, no dust, it's clean


 

[jonpaul]

Bonjour cher monsieur!

A Paris les vides greniers et marche au puces j'ai achetée similar EU 5..25.

1/if every PSU, HV and blanking is OK, then  Most likely is intermittent CRT socket pins, scratchy INT or FOCUS pots, or failing CRT eg if unit was previously dropped, hidden damage.

2/ Did you exercise and clean every switch and control, esp beam finder, X10 H, focus int?

3/ J'habite Paris 6eme si vous ne pouvez pas le dépannage, nous avons  le possibilité RDV.



Bon chance,

Jon

[shakalnokturn]

From my experience a knocking noise related to equipment using a CRT is due to a high voltage breakdown.
In many cases this can be seen in the dark as an arc (sometimes in the CRT itself) that should show up one way or another when high voltage probing.
I assume you have something like a 30kV probe that you used to measure DC voltages with a DMM. Bandwidth won't be great but hooking it to a scope instead may give you more clues.
Have you tried running it with no deflection at all, lifting the inputs to the X&Y plates?
I agree that none of this sounds good for the CRT...

[Wallace Gasiewicz]

Sometimes there is corrosion inside the CRT base pins. This can cause all sorts of unusual problems depending on which pin or pins are affected.
If you can desolder the pins and take off the connector on the tube and clean everything then resolder sometimes the tube works again.
Possibly you could remove the solder on the pins and add more good solder without actually removing the tube base.
Before removing the tube it is a good idea to discharge the tube pins to ground thru a resistor just to be careful.

[cascador]

Thanks for all your remarks.
I will try to run it without any deflection at all, good idea.... !
There is indeed a possibility of a hidden damage following a drop...
When the knocking noise occurs, before and after, there is no input voltage change on the DMM. It's constant, whatever happens...
There is a connector to the CRT that is removable, I removed it and everything is clean.
Still investigating...

[cascador]

I tried to disconnect the X1 X2 Y1 Y2 pins from the CRT back. I then shorted X1 / X2 and Y1 / Y2 together. And I could manage to get a (slightly moving) beam point (sort of) in the middle of the screen for 1 second then it disappears and never comes back.
To get it I have to switch off the oscilloscope, wait a few seconds, then on again with the intensity to 0, then increase it. At some point, the beam appears for 1 second, then it's over.
BTW I checked the optocoupler and the associated transistor, it's fine. I get +23 volts at the cathode at each blanking pulse.
I figured out that the voltage that should be 1890/1913 falls to 1877/1900 within a few seconds. Still 23 volts of difference between the two states, but could it have an impact on the beam .. ?
Focus, intensity, astigmatism and geometry voltages are fine.

[shakalnokturn]

Does the voltage drop you note coincide with the loss of the spot?

I don't think that much change in H.V. is important in itself, what is important though is that the differnce between cathode and G1 voltages doesn't change.

At least you know D313 and D314 are working.

For the sake of troubleshooting (at least confirming the CRT is OK) I'd lift R3037 and tie cathode to -1913V, getting the whole blanking section out of the way so you can have better confidence in H.V. supply / CRT bias.

Low voltage testing the opto is one thing but in this case it is also worth testing isolation by throwing a high voltage between LED and phototransistor and measuring leakage current. Isolation can fail with age. (1992!)
I and others have run into poor isolation over C314, C315, C317 causing various blanking problems.
T307 is often a good suspect too, I've run into a number of not so easy to corner faults due to BF199's.

[cascador]

No the loss of spot occurs this way: 1) power off 2) power on: no spot 3)  I turn the intensity up 4) the spot appears for 1 second, and goes away (moving sideways) 5) end of story I have to power off again - no other way
This occurs whatever the voltage level is, before or after the drop.

As suggested, I have connected the cathode to straight 1913 V. Nothing changes (still 1 second spot). For validation I then reconnected to default 1890 via R3037, while removing the optocoupler, and no beam appears at all.
I am afraid the CRT itself is dead. I might next try to change a few capacitors, opto and transistors here and there, but I am not very optimistic about the result.
Anyway thanks for all your suggestions, much appreciated !

[Andy Watson]

Have you checked the trace rotate coil? Easy to disconnect it and eliminate it.

[cascador]

Just disconnected it, same results.
I reconnected X and Y, and by playing a bit with all the buttons, I got some strange behavior, with beam flash lines starting down in the middle then up to the actual horizontal sawtooth location.
It seems the oscilloscope want to become a game console. Short video here: it may definitely prove the CRT is bad, because I can't explain how the circuits could even generate such a signal....
https://youtu.be/CmdbrZpp6D8

[George Edmonds]

Hi Cascador


I have been designing and servicing oscilloscopes for over 65 years now and have seen the problem that you have with a gradually or rapid fading out CRT display numerous times.  It is 95% of the time caused by the resistors in the high voltage dividers, in your case the 2kV, supplies to the CRT electrodes.

There is a little known problem with high value resistors, say over 500K, they slowly go high in value due to the applied voltage, the older the resistor the more pronounced the change in value with applied voltage is likely to be.  The moment the applied voltage is removed the resistor temporarily goes back to its original resistance and hence the display returns.

Please keep in mind that although the CRT uses some 2kV around the Cathode and some Grids it only takes a couple of volts change to some of the supplies (importantly Cathode to G1) relative to the cathode to turn the beam off, hence no display.  It is not a good idea to just strap the deflection plates together, it is more important to check there is no voltage between Y1 and Y2 and X1 and X2

It normally take some 10 to 20 minuets for the resistors to change in value, but I have seen much shorter times, due to the applied voltage sufficiently to turn off the trace.

The replacement resistors MUST be high working voltage one such as Vishay VR25 or VR37, now days most resistors have a maximum working voltage of 350VDC or lower.

G Edmonds

[cascador]

Interesting...
I just spotted the high voltage resistors, R3002, R3004, R3005, R3011 and R3012 on the diagram, they are mentionned as such.
The first 3 of them are used in the cascode to generate the +500V to stabilize the high voltage with the IC301 feedback.
Before I can get a hand on the appropriate replacement parts, I tried to quickly test by replacing high voltage resistors by a set of low-voltage ones in series, with the same resistance sum.
Although it does not solve the problem there is a noticeable improvement with a bit less erratic beam....
I'll try to replace them with high voltage ones and I'll keep you posted.

[GGMM]

Hello,

I have a Hameg 312-8 in progress.
The X  is unstable, (Slight intermitent shift of 2
 or 3 mm ) the trace return is visible.  ---> so no good values on the CRT?
Then the THT control transistors jumped.  ---> THT to hight??
The design may be different from your model.
In conclusion, I find the THT resistors  track  as interesting

cdt

[cascador]

If the trace return is visible, you may want to check the blanking circuit first. Set the lower X speed and even with a voltmeter without a high-voltage probe (be careful to clip the probe before measuring and don't touch them after that, these are DC high-voltage lines but the difference is small), you can see the voltage variation on the cathode, on my scope it's +-23 volts.

[coppercone2]

What kind of resistors does it have for the HV circuits? I think I would replace everything I can in there with metal oxide film or ceramic composition resistors (if you have the budget). IF you saw a improvement from changing those, then chances are its a good avenue to take.

A cracked resistor can spark intermittently and cause all sorts of problems. I saw this with a HV power supply recently, one of the supply resistors was invisibly cracked until I saw a spark under neath it while working late in darker conditions. You can also try to run it in a dark room and monitor it. Clicking noises = sparks. When I replaced everything with metal oxide, it was fine.


But this is for circuits that have series parts, it sounds like you need to heed the warning of the Vishay parts if there is only a single resistor biased to such a high voltage. In my case they were no more then 300V of tension on each resistor, so only a problem once things fail short.

[coppercone2]

Hi Cascador


I have been designing and servicing oscilloscopes for over 65 years now and have seen the problem that you have with a gradually or rapid fading out CRT display numerous times.  It is 95% of the time caused by the resistors in the high voltage dividers, in your case the 2kV, supplies to the CRT electrodes.

There is a little known problem with high value resistors, say over 500K, they slowly go high in value due to the applied voltage, the older the resistor the more pronounced the change in value with applied voltage is likely to be.  The moment the applied voltage is removed the resistor temporarily goes back to its original resistance and hence the display returns.

Please keep in mind that although the CRT uses some 2kV around the Cathode and some Grids it only takes a couple of volts change to some of the supplies (importantly Cathode to G1) relative to the cathode to turn the beam off, hence no display.  It is not a good idea to just strap the deflection plates together, it is more important to check there is no voltage between Y1 and Y2 and X1 and X2

It normally take some 10 to 20 minuets for the resistors to change in value, but I have seen much shorter times, due to the applied voltage sufficiently to turn off the trace.

The replacement resistors MUST be high working voltage one such as Vishay VR25 or VR37, now days most resistors have a maximum working voltage of 350VDC or lower.

G Edmonds

This is fascinating, I suspected some resistors of doing a few things sometimes but I never could find any problems with them really, but logically they were the most obvious cause of a problem.

[Runco990]

So I have had a few Hamegs and love the component tester most have. 

Stupid and simple question:

Have you checked for AC on all the voltage rails?  I've had one with similar weirdness and it was caused by an open cap in the power supply.

As always, I look at the simplest things first, so please don't hurt me...   

[GGMM]

hello,

Good evening,

The cloaking circuit consists of a CNY 17-11  and some transistors.
It is an emitting diode and a photo transistor in a 6-pin Dip case.
I think the diode and transistor should respond to the multimeter as a diode and a "normal" transistor.
But I think when I removed the CI  from its support it didn’t look good. As I have no spare parts, no test possible and I have to order other parts.
 Just confirm what I’m saying about the multimeter test to be sure.

cdt

[cascador]

Hi guys !
This afternoon I bought a HM 203-4 for €20 which has a time base problem. But I took it to tear it down and keep the CRT which is not the same as the 203-7 but is compatible according to the datasheet.
I then removed the 203-7 CRT and put the other one, and guess what, it worked the first time I switched it on !
Problem solved, it really was the CRT. I don't know what could happen to it however.
Thanks for all your good advices