Tektronix 7603 Oscilloscope - Repair Log

[David Hess]

One thing I did notice is the power rails in the module are down a bit (+14.5, -14.5V, +49.7V, -49.9V), I'm not sure what they are meant to be, but there are series resistors on the input to the module so maybe it's normal.

The voltage levels are not critical as long as they are stable.

The issue with the readout still remains so I'm starting to suspect the readout board again...

There are three signals which matter for this coming out of the readout board:

- X/Y shutdown which disables the X and Y inputs
- The readout vertical signal
- The readout horizontal signal

The X/Y shutdown goes first to the horizontal channel switch on the horizontal amplifier board, and then to the vertical interface circuit on the interface board where the plug-ins connect.  The readout vertical signal goes directly to the vertical CRT amplifier.

What is weird is that both are failing because they originate and terminate in different areas, and the X/Y shutdown apparently *is* working for the horizontal signal.  That implies two different failures which seems unlikely unless something on the readout board could cause both.

Tekwiki has a cleaned up PDF of just the schematics that makes following these signals easy.

[sean0118]

There are three signals which matter for this coming out of the readout board:
The X/Y shutdown goes first to the horizontal channel switch on the horizontal amplifier board, and then to the vertical interface circuit on the interface board where the plug-ins connect.  The readout vertical signal goes directly to the vertical CRT amplifier.

Thanks, I hadn't seen the X-Y inhibit coming from the readout board on the schematic. Just in case it's useful for anyone with similar issues it comes out of the U2126 timer along with the Z axis signals.

It turns out the plug (P38) for the X-Y inhibit input to the main interface board wasn't plugged in properly, must have come loose when I removed the rectifier module earlier. With this fixed the readout display doesn't move with the trace anymore, but the vertical of the readout is still collapsed.

I've been probing around the readout input into the vertical amp board and I'm not sure what is going on. I can see the signal at the inputs to R414 & R408 (see below) but their outputs are at constant -15V -8V. Both sides of R429 are at -15V DC. Is -15V the correct bias or should it be closer to -50V?

I've measured most of the resistors in circuit and they seem okay, PNP transistors Q407 & Q415 (2N2907A) also test okay with a multimeter, I substituted them for general purpose parts too but no change. Could it be the hybred IC U450? I can measure the diode drops in the transistors between pins 1 & 2 and 4 & 5.

I feel I must be missing something obvious, otherwise I might need to start removing components to test them...



R414:


R408:



edit: I found a schematic in the 7623 user manual which shows the DC levels, mine seem roughly correct. I'm measuring -14.4V instead of -14.1V on the input of J431, but I assume that's a schematic error? Otherwise Vbe of the transistor internal to U450 has to be 1V...

[David Hess]

Does the vertical work at all?  Does the position control work on the trace as expected while the readout stays along the bottom?  If so, then the hybrid part is working fine.

[sean0118]

Does the vertical work at all?  Does the position control work on the trace as expected while the readout stays along the bottom?  If so, then the hybrid part is working fine.

Yep the vertical for the trace does work fine, that's a good point. Looking at the vert amp circuit again some of those resistors are very small in value, so there probably isn't meant to be a large voltage signal there. I'm starting to think the problem is actually on the Readout board.

Probing around the Readout board I found the Y Deflection line going into pin 7 of U2284 Format gen to be only 4V. It seems the line is lowest at U2284, the other side of R2279 is close to 5V for example. With it removed the line is no longer pulled down.

Also the channel address lines (pins 8, 1, 15) are higher in voltage than expected, 2.5V with a multimeter and with a scope I can see they are getting pulled up close to 5V. The wavefroms on one of the schematics I have suggests they should be getting pulled up to around 3V, but there are 1k pullups to +5V so might be normal...

I'm starting to strongly suspect U2284, might also try replacing U2250 because it's just a 7493 counter...

[David Hess]

I'm starting to strongly suspect U2284, might also try replacing U2250 because it's just a 7493 counter...

It would be too bad if U2284 failed but I remember that happening in the past.

Many of these boards used Texas Instruments edge wipe sockets which have poor reliability, so the socket may have failed.  Reseating suspected ICs should be done.

[sean0118]

It would be too bad if U2284 failed but I remember that happening in the past.

Many of these boards used Texas Instruments edge wipe sockets which have poor reliability, so the socket may have failed.  Reseating suspected ICs should be done.

Yeah you're right, from looking around it does seem the Readout ICs might be prone to early failure:
https://groups.io/g/TekScopes/topic/7659141#142266

There's replacements for U2284 on ebay for a reasonable price so I might just buy one to see if it fixes it. I'm tempted to try adding a second resistor to pull it up harder first, although I doubt it will work.

Thanks for the suggestion, I have reseated all the Readout ICs and transistors but didn't see any change.

[David Hess]

Yeah you're right, from looking around it does seem the Readout ICs might be prone to early failure:
https://groups.io/g/TekScopes/topic/7659141#142266

The failures I have seen are consistent with contamination through the plastic package, which would make it a matter of age and the packaging.

[sean0118]

Okay, I finally got back around to this project. 


I bought a replacement 155-0020-00 for U2284 from ebay (seller: rcflying-hobby). It came in the original packaging which was pretty cool, but the foam had degraded and started sticking to the legs, but it cleaned up okay using some DeOxit.

I swapped it over with the defective one and the readout is working again, so does confirm U2284 was the issue. One interesting thing, the readout is now at the top of the screen, was at the bottom before. I don't think the readout vertical position can be adjusted that far? Maybe it's a different revision IC?

Everything else seems to be working okay, but I still need to cal the horizontal timebase module...

[dave_k]

Great work on the repairs, keep up the good work!

[David Hess]

One interesting thing, the readout is now at the top of the screen, was at the bottom before. I don't think the readout vertical position can be adjusted that far? Maybe it's a different revision IC?

Each plug-in slot gets readout positions at the top and bottom, and can use one or the other or both.  Not all plug-in types are consistent; early timebases use different readout positions than later ones.

The original IC must have been broken in that respect unless something else is broken now.  I do not know which readout positions the 7B50 uses because I only have later timebase plug-ins.

[sean0118]

Okay turns out the 7B50 timebase module has a "SWP CAL" trimmer on the front panel, so the horizontal cal was actually super easy. 

It's a little bit out once you get around the 50/100ns range, just the first few div on the left of the screen, maybe some small non-linearity somewhere either in the scope or timebase module? But pretty good considering 50ns is the limit of the 7B50 module:


1ms:



1us:




100ns:




50ns:

[David Hess]

It's a little bit out once you get around the 50/100ns range, just the first few div on the left of the screen, maybe some small non-linearity somewhere either in the scope or timebase module? But pretty good considering 50ns is the limit of the 7B50 module:

How do you know that the horizontal non-linearity is not from the 7603?

[sean0118]

How do you know that the horizontal non-linearity is not from the 7603?

Actually yeah I think you might be right, I swapped the modules into the 7623A and it looks better (maybe still out a little bit?). Looking at it again I'm not sure if it's non-linear, might be able to cal it out easily.

7623A:


7603:

[sean0118]

Okay I went through the horizontal amplifier calibration steps, seems to pass the "low frequency linearity" check. Low frequency horizontal gain adjustment seemed okay too.

While looking at the high frequency timing cal instructions I realised the 7B50 horizontal module can actually go to 5ns because of the x10 magnifier mode (although my time mark generator seems to struggle below 10ns).

Interestingly the 10ns high frequency adjustment capacitor (C588) is already at a limit so can't be adjusted further. I could adjust the 5ns adjust capacitors (C566, C586) though, which did improve the 10ns timing (slightly out on right side of screen still), although this did not improve the 50ns timing.

It's a bit of a mystery to me where the problem is, but seems to be getting near the limits of the scope?

10ns (before cal):


10ns (after cal):


50ns (after cal):