Topics

Tektronix 7854 Digitization Issue

Kyle Rhodes
 

Hi all,

I've got a wonderful 7854 acquired from a fellow list member, but I'm
having an issue with it. The digitization / storage function is not
working properly. Allow me to explain:

It acquires waveforms, but it's as though it's acquiring two different
waves, one at the correct frequency, and one approximately 1/5 the period.
I posted a video to YouTube showing an acquisition, you can clearly see
another superimposed portion of the wave across the entire sweep.

https://www.youtube.com/watch?v=vUq_e4OoAPA

The samples average out a bit if you run 10 or 100 acqs, but obviously it
doesn't result in a nice digitized waveform.

Also, the rendered text gets wonky during an acquisition. I'm not sure if
this is normal, or maybe it's related to the issue?

It's as though the time axis is not correlated properly with the Y axis
sample. Has anyone else encountered such a problem? This is a complex
unit, and before I spent many hours digging through the documentation and
learning how it works, I thought I'd ask for some starting points. Time is
money, and while I love working on this stuff, my day job involves using my
collection of equipment to produce other goods, so the less time spent
troubleshooting, the better. :)


Thanks,
Kyle

 

It acquires waveforms, but it's as though it's acquiring two different
waves, one at the correct frequency, and one approximately 1/5 the period.
I posted a video to YouTube showing an acquisition, you can clearly see
another superimposed portion of the wave across the entire sweep.
The "low frequency" that you see is actually the retrace (return of beam to left of screen). Have a look where it "ends" at the right side of the screen: Always where the real trace stops, so it's actually the starting point for the retrace. It should normally be invisible but blanking isn't set correctly for the digitisation.
It's likely that adjusting the "grid bias/blanking" for the digitisation part will solve your problem so check and if necessary adjust Z-axis levels. The Z-axis circuits are on the board mounted at the back, where the GPIB and battery backup connections are. It's not a dangerous adjustment in the sense that you may/will lose calibration.

Also, the rendered text gets wonky during an acquisition.
That is perfectly normal during acquisition.

Raymond

 

The "low frequency" that you see is actually the retrace (return of beam to left of screen).
At the beginning of your video, the retrace is quite visible as the low-frequency waveform. Subsequent acquisitions show artefacts caused by the fact that the digitisation system wasn't designed to show the retrace as a separate trace so it's just interfering.

Raymond

Albert Otten
 

On Tue, Sep 17, 2019 at 03:03 AM, Raymond Domp Frank wrote:


It acquires waveforms, but it's as though it's acquiring two different
waves, one at the correct frequency, and one approximately 1/5 the period.
I posted a video to YouTube showing an acquisition, you can clearly see
another superimposed portion of the wave across the entire sweep.
The "low frequency" that you see is actually the retrace (return of beam to
left of screen). Have a look where it "ends" at the right side of the screen:
Always where the real trace stops, so it's actually the starting point for the
retrace. It should normally be invisible but blanking isn't set correctly for
the digitisation.
It's likely that adjusting the "grid bias/blanking" for the digitisation part
will solve your problem so check and if necessary adjust Z-axis levels. The
Z-axis circuits are on the board mounted at the back, where the GPIB and
battery backup connections are. It's not a dangerous adjustment in the sense
that you may/will lose calibration.
---
Raymond
Hi Raymond,

I agree with your interpretation, retrace samples appear in the acquired waveform. But I think your suggested remedy is not working. The Z-axis blanking is itself a result of internal signal. The only logical output of the Z-axis circuit is the GSF signal which inhibits digitizing/storage of extreme Y values. During retrace there must be some "inhibit" signal active which prevents digitizing/storage during retrace.
It's won't be easy I think to pinpoint the cause. I spent a few minutes reading the Theory of Operation and looking at some diagrams, but oh boy!

Albert

 

On Tue, Sep 17, 2019 at 01:40 PM, Albert Otten wrote:


But I think your suggested remedy is not working. The Z-axis blanking is
itself a result of internal signal. The only logical output of the Z-axis
circuit is the GSF signal which inhibits digitizing/storage of extreme Y
values. During retrace there must be some "inhibit" signal active which
prevents digitizing/storage during retrace.
It's won't be easy I think to pinpoint the cause. I spent a few minutes
reading the Theory of Operation and looking at some diagrams, but oh boy!
I'd expect what you mention, Albert, but I have to confess that one of my 7854's was exhibiting the exact same behaviour when I first got it and I'm sure slightly adjusting one of the pots on the Z-axis pcb did the trick... Of course, I don't remember exactly which one... Z-axis had to be adjusted anyway and the problem was gone at the same time.
It'll be interesting to hear from the OP.

Raymond

 

Albert,

On Tue, Sep 17, 2019 at 01:53 PM, Raymond Domp Frank wrote:


Of course, I don't remember exactly which one...
It's the GSF (Good Signal Flip-flop) trimmer (U102) on the Z-axis board (dwg 13). See description of Digitiser Control Logic > Bus Control > Internal Acquire Mode, page 2-61 in my manual (070-2874-01), Jan '82. Also, see Table 2-8 on page 2-122 and are F,G-2 in dwg 21.

Raymond

 

On Tue, Sep 17, 2019 at 03:21 PM, Raymond Domp Frank wrote:


It's the GSF (Good Signal Flip-flop) trimmer (U102) on the Z-axis board (dwg
13)
The associated circuitry is in location B-3 on dwg 13.

Raymond

 

Of course, the pointing-out-where wasn't aimed at you, Albert but to the OP, although I directed the post at you.

Raymond

Albert Otten
 

Hi Raymond,

You are right! I misinterpreted the function of the GSF (a little bit stupid since the vertical signal I referred to is not at all involved here). Apparently the blanking signal is included already by the Z-axis controller and not a separate signal to the Z-axis amplifier.
Well, this will be a relief for the OP.

Albert

On Tue, Sep 17, 2019 at 03:21 PM, Raymond Domp Frank wrote:


Albert,

On Tue, Sep 17, 2019 at 01:53 PM, Raymond Domp Frank wrote:


Of course, I don't remember exactly which one...
It's the GSF (Good Signal Flip-flop) trimmer (U102) on the Z-axis board (dwg
13). See description of Digitiser Control Logic > Bus Control > Internal
Acquire Mode, page 2-61 in my manual (070-2874-01), Jan '82. Also, see Table
2-8 on page 2-122 and are F,G-2 in dwg 21.

Raymond

Kyle Rhodes
 

I've had some success! Adjusting the Good Sample Flip Flop seems to be the
ticket. In B5 on page 4-50 of the service manual it has one adjust R102 by
reading some voltages to start with -- this did not work for me. But,
attaching the scope as starting in step 'L', I was able to see the waveform
and I adjusted R102 until the signal looked good and triggered well.

However, I think I was a bit over-zealous in putting the covers back on, as
while the waveform acquires much better now, I am seeing some strange
behavior at the end of the waveform -- namely, it tends to zero out for
0.5-1 divisions. I suspect I can fix this with some more careful
adjustment of the GSF signal pot, as I don't recall it doing this in the
past.

I started from the beginning of the Z-axis calibration on 4-47. My initial
voltage was 30V high, so I got that taken care of.

Though, in B4 "Adjust Z-axis Transient Response) on 4-49 I had some
trouble. I could not for the life of me get C150, R150, C155, and R155 to
do much of anything. They all already were pegged as well, so not sure if
someone else tried in the past too. Not sure if this affects digitization
stuff at all, or just analog display? I need to look at the schematics in
more detail...

Furthermore, it appears I need to look into calibration, as when I was
paying close attention to signal amplitudes and periods during testing, I
noticed some things are off.

And, shall I say, this scope is a beast! What incredible engineering went
into it....


Thanks,
Kyle

On Tue, Sep 17, 2019 at 10:49 AM Albert Otten <aodiversen@...>
wrote:

Hi Raymond,

You are right! I misinterpreted the function of the GSF (a little bit
stupid since the vertical signal I referred to is not at all involved
here). Apparently the blanking signal is included already by the Z-axis
controller and not a separate signal to the Z-axis amplifier.
Well, this will be a relief for the OP.

Albert
On Tue, Sep 17, 2019 at 03:21 PM, Raymond Domp Frank wrote:


Albert,

On Tue, Sep 17, 2019 at 01:53 PM, Raymond Domp Frank wrote:


Of course, I don't remember exactly which one...
It's the GSF (Good Signal Flip-flop) trimmer (U102) on the Z-axis board
(dwg
13). See description of Digitiser Control Logic > Bus Control > Internal
Acquire Mode, page 2-61 in my manual (070-2874-01), Jan '82. Also, see
Table
2-8 on page 2-122 and are F,G-2 in dwg 21.

Raymond


 

On Tue, Sep 17, 2019 at 01:40 PM, Albert Otten wrote:


.....the GSF signal which inhibits digitizing/storage of extreme Y values.
This reminded me of the ancient but quite nice Philips PM3320A/3323A/3340A digital sampling 'scopes: They have separate adjustments to suppress excursions of the trace beyond defined left, right, top and bottom positions by blanking.

Raymond