Hameg 203-6 compressed X and Y traces

[ik1zyw]

Hello.

TL;DR
Over the last two weeks I built an emulator of the first videogame in history (1958), "Tennis for Two". It has been shown and offered to play to visitors of an exhibition of retro-videogames as a side-event to the tennis ATP Finals we've hosted in my home town. BTW, Tennis-for-Two and Pong scored the longest player time, regardless of players' age.
The emulator (credits to Windell Oskay - Evil Mad Scientist) uses an analog scope in XY mode as output device, so I dug the Hameg 203-6 out of the "storage for further inspection and repair" to check the circuit at home before bringing it to the exhibition (we used a Tek with my Hameg as backup).

To the point.
After the exhibition I decided to check the Hameg since it was out of the storage, and of course it has a problem besides the usual wonky contacts and potentiometers. Basically X and Y sweeps are "compressed", almost of the same amount: 0.8x on X, 0.7x on Y. On the left side there is a brighter dot, but that should be the electron beam waiting for the "start sweep" command and that usually sits out of the visible area. See the picture where I aligned the test signal traces to the rightmost side. With X10 magnification the trace takes the whole screen, but I think that's the way the X10 trick works. "UNCAL" were turned fully CCW, so "off".

Another little problem is that the brightness cannot be turned down to zero, but that's another story and I think I know how to try to fix it. The picture is at minimum brightness and it is already clearly visibile in daylight.

According to the diagram X and Y deflection circuits use a 150V power supply line, which I measured to be 152V. There's also a line at 130V that sits at 133V. The DVM doesn't detect AC on them. Across R910 and R917 resistors voltages match what is in the diagram (GNDed input, trace centered). I can move the trace above and below the screen with that voltage swinging between 30V and 115V under my control. +5V, +12V, -12V are OK too.

Now, if I understand correctly the workings of a CRT, I do not have a problem in the 2kV HV parts since those are used for beam acceleration and brightness control. The problem lies in the deflection control chain. Is that (probably) correct?

Having almost the same compression amount on both axes lead me to think of a PSU problem such as too low voltage, but those are within spec. The compression amount doesn't change with warmup. I am leaning towards the idea that someone has played with trimmers because, who knows, they needed a specific "full screen" display. But I am sure Your wisdom will bring different options.

Thank you for reading so far.
Paolo


Schematic diagram, on Dropbox, press "Cancel" if it asks for login or sign-up:
https://www.dropbox.com/scl/fi/fz6typw59rl9b2q33oxeu/Hameg-203-6-schematic_complete.pdf?rlkey=k6k05i470vxeok9q9ykulqcfx&dl=0


PS: I am aware of dangers of lethal high viltages and I know the game rules. With rule #1: one hand in the pocket.

PPS: the worst wonky contact is the channel II "invert" pushbutton as every little vibration is reflected on the second trace.

[doktor pyta]

Hi,
You also probably have a problem in the beam blanking part, so definitely check high voltages too.
HV can affect image size on the CRT.
As McGyver would do it take 9 x 10Meg 1% 0.6W THT resistors and with 10Meg input of the handheld multimeter it will create 10:1 probe suitable up to approx 2.5kV.
You can put it into plastic case of a ballpen.
BR

[shakalnokturn]

I second doctor pyta's input.
This would need checking on full schematic but I don't think it should have any noticeable effect as most voltages are regulated, still have you checked line voltage setting is closest to reality?

[ik1zyw]

Hello.
Yes, the voltage selector is correct (220V).

I spent the evening looking for proper resistors to build the HV divider. In McGyver's style I'll use what is available which means I won't get a comfortable 1:10 ratio but something close to 1:7.3 (will be confirmed with lower voltages) still within the instrument max range, and use a calculator to do the math. Also one DVM input is not 10.0 Meg but rather 10.9 Mohm: better check them all.

It will take a few days before I can measure the HV. I'll get back with results.

[jonpaul]

ik1zyw Ciao!

have a lot of experience with vintage  TEK and Hameg scopes.

The 1980s Hameg CRT scopes are wonderful for general use and easy to use., I  have several similar to yours.
Like any old equipment the controls need lube and exercise and   PSU caps checked and perhaps replaced.
Beware of socketed ICs and transistors, and that some PCB use mounting screws for a ckt earth.

1/ HV too  low will give dim and expanded traces X and Y
HV too high gives very bright and compressed traces.

2/ Measurement of the HV is not trivial as the resistors in 10s..100s megs have a tolerance and voltage coefficient.

Few normal ohmmeters can accurately check the HV reisitors.

3/ Suggest to get full service manula for the 203-6, R&S has them in their discontinued archive.
The manual has calibration for HV, V, H etc as well as some troubleshooting info.

4/ Known weakness in all Hameg CRT is the Z axis blanking: The Optocouplers and NPN/PNP transistors are socketed and do fail.

Worse are 3 x 68 pF 2KV ceramic disc caps running at 1800V but rated 2kV.

Symptom is as you see no blanking.

5/ Unlikely that the Z axis blanking and shrunk traces are related.

BON CHANCE,

Jon

[ik1zyw]

Bonsoir Jon

1/ HV too  low will give dim and expanded traces X and Y
HV too high gives very bright and compressed traces.

Spot on. HV too high. The -1900V line sits at -2550V or thereabout.

2/ Measurement of the HV is not trivial as the resistors in 10s..100s megs have a tolerance and voltage coefficient.

Having read this info somewhere else, I tested the 68 Mohm series at 12V, 80V, 200V, 300V and the resulting ratio increased with voltage, like 1:8.01 to 1:8.03. Nothing too serious that would mislead the 2kV reading.

4/ Known weakness in all Hameg CRT is the Z axis blanking: The Optocouplers and NPN/PNP transistors are socketed and do fail.

I'll first fix the 1900V, then think of Z axis if that will still be an issue.

I'm going back to the schematic to understand the PSU circuit and what could cause higher voltage.
Paolo

[jonpaul]

Ciao Ciao  Paolo:

0/ Bravo for the HV test. CRT deflection factor is inversely proportional to anode v,

Thus your V anode increase 1900>>2550V gives a factor of 74.5% less deflection. That corresponds to your screen measurements.

1/ See the HM203 schematic for the HV PS and regulation system, especially the feedback resistors can be drifting higher in resistance over the decades.

2/ Avoid extended operation at 2550 V till the HV is fixed, as the PSU and blanking will be stressed.

3/ Blanking is dependent on sweep rate, the bright left-hand   line and dim retrace will be much worse at fast sweep rates.

4/Blanking is part of the CRT circuit  it may affect the HV voltage.
Since the optocoupler output side  is  floating at the HV, troubleshooting is not trivial 

5/ See the excellent Tektronix Circuit concepts books of 1980s:
https://w140.com/tekwiki/wiki/Concepts_Series

Several great books on CRTs, CRT circuits, PSU, Vert/Hor amps, etc. Despite from TEK, this is general info that applies to all CRT scopes.

https://w140.com/tekwiki/images/6/62/062-0852-01.pdf
https://w140.com/tekwiki/images/a/a1/062-0888-01.pdf
https://w140.com/tekwiki/images/b/b3/062-3252-00.pdf

Bon Chance


Jon

[Greybeard]

May be this helps:
https://www.eevblog.com/forum/repair/(solved)-blanking-issue-on-hameg-203-7-oscilloscope/msg2705480/#msg2705480

[TaylorD93]

I use a HM203-6 which i bought for parts as it was broken.

I had a similar issue that the -1.9kVDC rail was actually closer to -2.5kVDC.

There is a 741 Opamp which regulates the 1.9kVD rail along with a number of BF422 (i think) cascaded in series. Check these are functioning correctly. In normal operation both input of the 741 opamp should be equal within a few mV, if not then the Opamp maybe dead, or the feedback loop is not able to regulate the 1.9kVDC rail correctly, pointing at a fault elsewhere.

Due to the excessive voltage for a prolonged duration, my 203-6 had many faults on the UnBlanking circuit hence why i got it cheap for spares, there are two transistors in a 6 pin DIP Socket, they had very corroded contacts which i was able to clean with a firbeglass brush, perhaps check these too. Tappnig them with an insulated tool, i would find the blanking circuit would occasionally function and then stop. You can tell the unblanking is not functioning because the beam is bright at the start of the trace (where it would normally be off) it is stationary which causes the dot to be bright.

Other than that i had issues  with the 2kVDC 68pF capacitors all getting stressed and failing.

re the brightness, there are three potentiometers across a 68VDC Zener, but all referenced at -1.9kVDC. If that supply rail is excessive, then the plate potentials will be incorrect stoping functions like the brightness control. So if you resolve the 1.9kVDC rail first, it may resolve some of your other issues too.

Good luck,

Thanks, Taylor

[ik1zyw]

A big THANK YOU for all your inputs.

Considering that those three 68pF capacitors are prone to failure, I am ordering replacements. In the meantime I will just check the 741 inputs and then keep the Hameg unplugged to avoid further damages. I will use the delivery time to study Tek books linked by Jon.

Or... why not build an HV capacitor tester? In the link by Greybeard (https://www.eevblog.com/forum/repair/(solved)-blanking-issue-on-hameg-203-7-oscilloscope/msg3891608/#msg3891608) there's a nice screenshot of a V/I curve tracer: is it something that can be built easily at home and shown on a DSO? I think I need a variable HV generator, an XY scope and some high-value resistors, right?

Paolo

[jonpaul]

Paolo: Issue in blankang can be from several issues as mentioned: Optocoupler, NPN, PNP transistors, OR the 68 pG caps.

Ckt runs with NO 68 caps but you loose the HF comp. (just for test)

1/  68 pF disc caps in HM203 are VERY hard to find.

Due to small size and three stacked radially on the PCB with low height.

I have NEVER found an exact replacement all were too low V or too thick or to large diameter.

2/ Testing is easy but not needed.

a/ cjeck cap on any RLC bridge or meter should be 68 pF +/- 20%

b/ Look at current the cap in series with variable HV DC PS.

Leaky cap is soon seen as excess current after charge.

No special equip needed just low current microammeter and PSU 0-2000V

Jon

PS: Pix typical HM103 defective blanking showing retrace and blanking on squarewave probe check

[Greybeard]

... why not build an HV capacitor tester?
In the link by Greybeard (https://www.eevblog.com/forum/repair/(solved)-blanking-issue-on-hameg-203-7-oscilloscope/msg3891608/#msg3891608) there's a nice screenshot of a V/I curve tracer: is it something that can be built easily at home and shown on a DSO? I think I need a variable HV generator, an XY scope and some high-value resistors, right?

You can test the caps statically (2000VDC power supply and microampmeter in the ground wire). I just had a curve tracer so I used it.

If you want to generate a curve you need:
- HV transformer
- HV probe (100:1 or 1000:1) or HV voltage divider
- Shunt (R1) in the return wire of the D.U.T. (for example R=U/I=100mV/10uA=10kOhm)
- Resistor R2 to limit Imax if D.U.T. is shorted (not needed if R1 is 10kOhm in this case)
- Oscilloscope in XY mode. Invert Y channel.

I found this schematic of a simple curve tracer:


Use your power grid voltage primary instead of 120VAC and your capacitors test peak voltage instead of 12VAC.

[ik1zyw]

Good European evening.

Jon I know that those 68pF caps might not be the culprit and I might not begin the fix by replacing them. I have located them on the TB board and I agree that they are not easy to unsolder in that position. Probably I ordered the right parts (blue pills, 3 kV). I think I will have to extract the board which means disassemblying the front panel too, unless I mechanically destroy them. Unfortunately the R&S archive doesn't carry the full service manual for this model, just the schematics, or I have not been able to find it.

The curve tracer circuit is so simple that I'll give it a try. Any AC/AC adapter will work for initial testing. For testing at 220Vac it's as easy as connecting back-to-back two identical transformers. Then maybe a surplus transformer for a valve radio could provide kV AC.

The socketed transistors do worry me, especially because those sockets tend to become loose with time. I'll have a look at them too.

Paolo

[ik1zyw]

There is a 741 Opamp which regulates the 1.9kVD rail along with a number of BF422 (i think) cascaded in series. Check these are functioning correctly. In normal operation both input of the 741 opamp should be equal within a few mV, if not then the Opamp maybe dead, or the feedback loop is not able to regulate the 1.9kVDC rail correctly, pointing at a fault elsewhere.

Across inputs of IC502 - the 741 - I had 455 mV. I reseated it in the socket, no change. I replaced it and there it is, -1900 Vdc and less than a milliVolt across pins 2 and 3. The trace is back to the correct length and expected brightness. I will keep it running for a while to be sure the fault is somewhere else and it kills the 741 on the long run.

Now there's still the unblanking to fix as the trace begins with a brighter dot. Tapping T501 (BF199) and T502 (BF440) makes no difference, so the socket is still providing proper contact. I tried new BF199's from the parts bin and the trace doesn't show up at all. I'll check them with a transistor tester and look for replacements or a fault elsewhere.

Paolo

[Greybeard]

...blue pills, 3 kV.
I think I will have to extract the board which means disassemblying the front panel too, unless I mechanically destroy them.

I replaced them without removing the PCB.
There was not too much space to solder, but somehow I managed it.

The curve tracer circuit is so simple that I'll give it a try. Any AC/AC adapter will work for initial testing. For testing at 220Vac it's as easy as connecting back-to-back two identical transformers. Then maybe a surplus transformer for a valve radio could provide kV AC.

I'm not sure a test at 320Vpeak is enough for a 2kV or 3kV component.
Finding a cheap 2kV transformer will not be easy...

The socketed transistors do worry me, especially because those sockets tend to become loose with time. I'll have a look at them too.

Replace the socket by a good machined gold plated one.

[Gyro]

There is a 741 Opamp which regulates the 1.9kVD rail along with a number of BF422 (i think) cascaded in series. Check these are functioning correctly. In normal operation both input of the 741 opamp should be equal within a few mV, if not then the Opamp maybe dead, or the feedback loop is not able to regulate the 1.9kVDC rail correctly, pointing at a fault elsewhere.

Across inputs of IC502 - the 741 - I had 455 mV. I reseated it in the socket, no change. I replaced it and there it is, -1900 Vdc and less than a milliVolt across pins 2 and 3. The trace is back to the correct length and expected brightness. I will keep it running for a while to be sure the fault is somewhere else and it kills the 741 on the long run.

Stuff like this sometimes happens with high voltage stuff. The bipolar opamp used for voltage regulation in my 10kV insulation tester suddenly failed (in the same way) for no apparent reason a couple of years back. I surrounded the new one with protection diodes (inputs to both supply rails) to be on the safe side and no further problems. Maybe it was a chance flashover or insulation breakdown I don't know. The tester was at least 20 years old (even older when it failed) when I got it with no sign of previous rework, sometimes old bits just fail.

[TaylorD93]

Fantastic, well done!

The Unblanking circuit also relies on the Opto Isolator, the Low voltage signal (from the time base and trigger circuit) forward biases the diode in the Opto.

You could remove it and check the forward voltage of the diode using a multimeter, check it is not conducting in both directions etc.

If i recall correctly the Opto also has a "base" connection to one of the 68pF caps. Now i have not tried it before, but you might be able to do the "usual" diode check on the Opto's output transistor.

That is take a DMM set to Diode test, and check the Base Emitter junction then the Base Collector, typyically the Base Emitter will read a few mV higher than the Base Collector. But i have never tried that with an Opto before, only regular transistors.

[BlownUpCapacitor]

I'm not sure a test at 320Vpeak is enough for a 2kV or 3kV component.
Finding a cheap 2kV transformer will not be easy...

I think a microwave oven transformer will work. I used one in the past for testing a spark gap in an old spark gap tesla coil project. Just a variac to trim the output voltage of the transformer and using an old analog volt meter with a 5kv voltage input rating to measure the output. I used the capacitor and rectifier to make a DC voltage source.

This method might be a bit dangerous though as... well, it's a MOT. I've gotten an adjustable HV power supply than can supply 3kv from an electronics surplus store, so that could also be an option, though I don't think that there is a great chance that you will find one at yours.

[ik1zyw]

Hello.
Little update.


I tested the optocoupler as suggested by TaylorD93 and the readings match with his description.
I tried replacing the optocoupler with a similar part I had on a scrap board (CNY75GC, responds correctly to the diode tester) and saw no difference with unblanking. With or WITHOUT the optocoupler there is no difference in the retrace/unblanking: trace is present in both forward and backward direction.

I also decided that pushing a bit the intensity and setting the timebase to 50 ms/div or more shows the electron pen moving back from right to left, which is a useful visual clue together with the brigher spot on the left hand side.

With this setting I visualised the optocoupler pilot signal on an analog voltmeter (too lazy to connect the second scope through an isolating transformer). The signal is at +2V during trace and blinks to 0V during blanking/retrace.

Conclusion so far: the optocoupler is probably OK and it gets the command signal.

EDIT. The BF199 that came with the scope is bad: Vbc = Vcb = 0.12V. New ones behave properly. END_EDIT.
Without the BF199 trace is present in both forward and backward direction.
With a new BF199 there is no trace at all (tried 3 of them, same result). Why?


I need a BF440 or equivalent. I also found a BF440 datasheet where the pinout is different between TO92-2 and TO92-4 case.    My next step will be to reverse engineer the correct pinout matching the schematic diagram with the actual board. Then I'll try to fit a PNP with proper voltage rating, probably crossing a couple of legs.
EDIT. The pinout is "BEC", with emitter in the centre. This one seems to be good at the "diode test". END_EDIT.

There is a green LED in this part of the circuit. Is it supposedto light up/flash or does it serve as a voltage reference?

Meanwhile I've received the 68pF 3kV capacitors and their size is equal to those on the board. Considering how much I've paid them I'll try first replacing socketed components in the unblanking circuit. And, if I'll replace the capacitors, I'll try to salvage the part vs destroying it for easier extraction.

I agree that getting a 2+ kV transformer to build a component tester will require some patience. It's better to use a voltage multiplier and do a static test at the desired HV using the procedure already pointed out. I will not get the curve tracer effect, but I will be able to step voltages nevertheless.

Paolo

[ik1zyw]

New day, new troubles.

As they've been exposed to 2500V, I have replaced the 68 pF 2kV capacitors around the optocoupler. It's not that hard: give it a bit extra solder, suck it up, pull one leg out and the second will come easily. I found out that it is WAY better to shine a strong light through the PCB instead of illuminating the solder side of the board (see attached picture): you don't create shades with the iron, hands, forehead, ...

Now I have no trace at all. Not even a dot in X-Y mode. With both old or new optocoupler, BF199, BF440, no green dot/line.

While turning on the scope on the bench without the cover I noticed that every time there was a flash on the top RHS. I shot a video with the smartphone so that I could locate the exact point: near/under R525 which is on the collector of the first BF422 in the voltage regulator.

R525 was open so I replaced it with the prescribed 1k (I haven't checked those BF422 yet). No trace.

I am back to the starting point: since -1900V is present, re-check all other voltages.

Is there a way to exclude the blanking with a jumper, somewhere? Or given that I have no electron beam even in X-Y mode, there's something wrong with acceleration? Well, I'll find out as soon as I check +150V and +130V linesI guess.

Paolo

[ik1zyw]

Alright, I fixed it!

Voltages were all correct. Un/blanking signal was present. BTW, in XY and component tester mode there's obviously no blanking. So I opted for checking component values in the area around the optocoupler with K501, T501, T502 all removed from their sockets.

Resistors should be the same value of the diagram, or slightly lower. Diodes should conduct when directly polarised and be open or show a high impedance on reverse (because of some circuitry). I didn't get to testing capacitors. Refer to the piece of schematic diagram I've attached above. I did everything with a DVM after I gave 10 minutes to capacitors to discharge.

We already knew that T501 BF199 was bad.
I found R501 going to the cathode being 220 kohm vs 51.5 ohm. Since it is very hard to reach I chose to solder above it, in parallel, a 56 ohm resistor.
I found D509 conducting in both directions at 0.23V. I chopped it off leaving enough legs to solder a 1N4148 on the component side of the board.

That's it. T502 BF440 and the optocoupler where still good.

The trace is now crisp clear without any sign of un/blanking.

Thank you all for your valuable guidance.
Paolo

[TaylorD93]

Congratulations on fixing the scope!