Hitachi V-202F: deformed beam on CRT

[EduardoLM]

I have finally fixed the HV supply of an old Hitachi V-202F. In fact, I have replaced the whole HV circuitry with a small PCB with an oscillator + mosfets + ferrite transformer, generating -1500V for the CRT (picture attached).

I'm very happy the CRT shows something (thanks a lot to oldway for all the help on fixing it)! Now I will work on the X and Y sections. However, something still bothers me, possibly still related to the HV supply. I have attached a picture of the scope screen at the slowest sweep speed, with both channels on.

Please consider the beams are moving, and the camera blurred them horizontally, so each "dot" is in fact half the width shown in the picture. The problem is that the beams are not round, but short vertical lines! The same thing happens when just one channel is turned on.

What can be the cause of this? Can it be caused by problems in HV? One thing is still have to fix in the HV section is that it should be -1900V, but it's generating only -1500V. Could it be the cause?

Since I know not much about analog scopes, there's another thing I would like to ask: is it normal that, when both channels are on, both traces to cross the screen together, like my scope does? Shouldn't the beam on each channel alternate after a complete sweep ends?

I mean:
sweep 1 -> channel 1 on, channel 2 off
sweep 2 -> channel 1 off, channel 2 on
sweep 3 -> channel 1 on, channel 2 off
sweep 4 -> channel 1 off, channel 2 on

If this is correct, instead of both traces crossing the CRT together, this may be the cause, because the CRT has only one cannon, and tracing both channels simultaneously would force the beam to be constantly changing Y axis to do its job, and even small imprecisions would make a vertical blur visible, just like I see.

Well, any help will be greatly appreciated, thanks!

Eduardo

[EduardoLM]

Sorry, I should have done my homework before asking: I have to look into Alt and Chop modes.

Reading the manual, I have found that my scope automatically uses Alt mode for time/div of 1ms and lower, and Chop mode for time/div bigger than 1ms.

So, it's in Chop mode in the attached picture. But the question that remains is: is it normal to have a vertically distorted beam in chop mode?

Thanks!

Eduardo

[JFJ]

... But the question that remains is: is it normal to have a vertically distorted beam in chop mode?

As you've already found out, at slow time-base settings it's normal for the oscilloscope to rapidly chop between the two channels (giving the appearance of both traces being drawn simultaneously). However, the spot should not be vertically distorted in chop mode.

Whilst some distortions could occur if the beam was not being properly blanked as is it moves from one trace position to the other, the cause of your scope's distortion might simply be that its spot geometry needs adjusting. This is best done in X-Y mode (which displays a single static spot on the CRT). The astigmatism control (RV1035) is then adjusted, in conjunction with the focus and intensity controls, to achieve the best small round spot.

[EduardoLM]

Thanks for the help, JFJ. I will look into that, but before putting the scope in XY mode, I will have to fix the Y circuitry, it's not working, no signal I input to CH1 or CH2 makes any difference, all I see is an horizontal line.

It seems this old guy still need some more work before it's functional.

Thanks!

Eduardo

[oldway]

First of all, congratulations on your perseverance and courage. 

Unfortunately, I must tell you that to produce the high voltage is not enough.

It is necessary that the voltage is correct (-1900V), that it is stabilized (no changes with the brightness setting) and that it is appropriately filtered.

In fact, the sensitivities x and y depend on the acceleration voltage.

If these conditions are not met, no calibration x and y is possible, making the oscilloscope virtually unusable.

Before looking for other failures, you need to solve these problems.

Sorry to "cool" your enthusiasm. 

[EduardoLM]

Thanks for your reply, oldway!

About filtering, it's ok, this part of the circuitry was not replaced, I plugged the new transformer secondary to the same point the old secondary was connected. So the original rectifier and filtering caps are all the same as in the original.

The voltage is not stabilized indeed, but I can add that by adding a voltage divider to the output voltage. That way I can feed a voltage sample back to the MCU and adjust the PWM duty-cycle to keep a constant -1900V. I think this will solve the issue!

By the way, tonight I have fixed the -1500V and increased it to -1960V. This is the closest I can do without proper regulation. I will add regulation to that before moving to the Y circuitry.

Thanks again!

Eduardo

[jh15]

It could be vertical amp noise or ripple, try running up your horizontal speed to see if there is some kind of ripple pattern.

Otherwise ground your vertical input or make sure there is no vertical deflection and make sure you can get a focused dot. you do have an astigmatism pot.

[EduardoLM]

jh15, thanks for the suggestion, and indeed, I think you're right. With faster sweep speeds, I can see pretty severe noise on Y. It's shown in the first picture.

Then I unplugged +Y and -Y from the Y amplifier and grounded both. This is shown in the remaining pictures. With them grounded, the noise is gone, and it seems the focus control is behaving like it should: I can stretch the beam horizontally and vertically, or make it round on mid position. The next 3 pictures show how the focus control changes the beam.

The only problem is that when I adjust focus to have a round dot, it's a fat dot! Changing astig control makes the dot blurry, the best position is on one end of the pot range, I can't go further.

So, there's  indeed noise on Y circuitry, and there's probably something wrong around the astig control, since it's range is not sufficient for proper regulation. I will look into that. By the way, HV is about -1960V now, which is good according to the service manual (between -1800V and -2000V).

Thanks again!

Eduaro

[Andy Watson]

The upper trace on the first picture shows a definite pattern to the waveform - can you turn-up the time base to resolve this pattern (it should give you an idea of where it is coming from)? Also, try triggering on the "Line" setting with the timebase set to 5 or 10mS - this will indicate if the "noise" is ripple related.

[klunkerbus]

I'll bet the frequency of the ripple is the same as the switching frequency of the HV supply. 

[EduardoLM]

klunkerbus, it really can be!

Tonight I'll do what Andy suggested (I have only 1 hour per night to work on the scope). The HV PSU switches at 50kHz, and there's another detail to take into account: it uses square wave on the primary, which can lead to lots of harmonics. The original one had a more sinusoidal waveform, so I think the new one  is a lot noisier. Perhaps adding capacitors to the right spots can solve the problem.

Thanks!

Eduardo

[jh15]

I wouldn't add capacitors, but check the ones that should be soing the job. Unless you meant replace rather than add.

Some progress is good, you will be learning a lot perfecting your 'scope.

[EduardoLM]

jh15, I meant adding (or increasing) capacitors to filter possible noise or ripple caused by my HV power supply, not on the original scope circuitry. However, I don't think anymore that HV switching is the ripple cause. Last night I made some tests and measurements on both channels front-end, and this is the result:

Channel 1: completely dead (I couldn't follow the signal past the first two transistors), and it shows almost no ripple.

Channel 2: showing heavy ripple, unrelated to the input signal. When voltage scale rotary switch is changed, ripple frequency also changes! And the frequency is quite different from HV switching frequency. And I noticed it has been tampered with by a previous owner (non-original transistors).

So, I'm not sure yet, but it seems the HV switching is not causing the ripple, as suspected yesterday. But I know one thing for sure: both channels need some work on them, and I think it's better to spend some time on them before moving to other things, like the astig pot or HV regulation.

I'm surely learning a lot, and this is the main purpose of all this. This old scope would not be worth all this work and time spent, if my interests were only about fixing and selling it.

Eduardo

[EduardoLM]

I'm checking both channels input circuitry for bad capacitors or transistors, and have already found some faulty ones. One transistor in particular has been replaced by an equivalent with a different pinout, but the guy didn't change leads to go into the correct PCB holes, so the thing would never work anyway.

I could find all the faulty parts locally, but one: the 2SK106C. It's a N-channel JFET. Does anyone know about an equivalent? I have chacked the web, but the only one I could find was the NTE458, and I can't find this one either. There's one aliexpress seller that has it for sale, but it takes 2 months to arrive here...

Thanks!

Eduardo

[jh15]

While you are waiting, use the least good channel for donor parts and get the other channel going. If you still can't, then you need more troubleshooting. Rather than wait for a part one at a time.

This is if it is easy to swap parts, and your desoldering skills won't hurt donor parts during removal.

Also, especially since others have been rooting around in there, be sure all ground connections are sound.

Might be loose or missing screws and possibly shields removed.

[EduardoLM]

Hi jh15,

Had good progress this weekend! Channel 1 is working! I did what you said: used all good parts from CH2 to make CH1 work. Please see the attachment. It also shows the new blue acrylic I bought to make the scope look better, the original one was very damaged.

About my order, since I could find almost all the parts at a low price, and since they are discontinued, I bought all transistors used in channels input circuitry and Y deflection, as a stock for the future. I have also replaced all electrolytic capacitors of the scope, and all electrolytics of 10uF and under, I replaced by MLCC ceramic ones, this should make my scope less prone to bad capacitor issues.

At least part of the noise I have reported earlier was indeed caused by my HV PSU oscillator. I could see random spikes of hundreds of mV on the stabilized 5V, +12V and -12V lines. After improving the filtering of my PSU, the spikes have been reduced to 60-80mV.

Now, all noise I can see is 60Hz. But it's expected, I'm testing the scope with all case ground connections loose to make it easier to work on it. Once I fix both channels, I'll ground it properly to see how bad the noise really is.

I have also finished a new HV PSU PCB, this time with feedback to make it stabilized. I will try to install it this week.

Thanks again!

Eduardo

[EduardoLM]

After some more work, the scope is almost perfect. I had to fix some broken contacts on the attenuator rotary switches, and replaced faulty transistors on channel 2, and did a rough calibration, and I can say it's working quite well!

Now I can focus on the original problem that started this thread: noise. With the other problems sorted out, now I can clearly see the noise is indeed caused by my switching HV PSU, as some of you already guessed. The attached picture shows it pretty clearly, it's a distorted square wave. Both channels are affected, but channel 2 is worse, probably because the HV PSU is nearer it than channel 1.

Please note the scope is in the most sensitive scale (5mV/div) in the picture, so the noise is really small. The picture shows the noise on AC coupling, but's the same on DC. On GND, it's much lower, but still present, and that leads me to conclude the HV PSU noise is polluting the whole ground plane.

I'll look into that now, and I would like to ask one more question: what level of noise is acceptable for a scope like this? The specs doesn't mention it's noise parameters, but I would like to know what noise amplitude could be accepted as "ok" for this low end and old scope.

Many thanks!

Eduardo