Tektronix 7603 Oscilloscope - Repair Log

[sean0118]

Following on from the success of the Tek 7623A repair I thought I'd have a look at this 7603. This one came with a 7D01 Logic Analyzer plugin installed.

This thing is super dirty and crusty in some places inside, so I think one of the first things I'll do is clean it up a bit. I checked the voltage rails and they were out by maybe 1% but trimming the -50V brought them back within range.

Although, it seems this scope does have some major issues. The trace is doing some very strange things with and without plugins installed. Turning the Intensity knob has no effect, but turning up the Readout Intensity knob causes a strange pattern I don't like the look of, if you know what causes it let me know. 





With no plugins:



With plugins from 7623A installed:



With the Readout Intensity turned up:









Board locations:



Power rail test point Locations:



A5 Vertical Amplifier:



A6 Horizontal Amplifier:





A8 Z-Axis Amplifier



A3 Readout Board

[sean0118]

Okay, I made a little bit of progress. The vertical amplifier outputs are way off from what I would expect. It's not a square wave, and it's way too slow (period should be ~1ms). The period doesn't change when I change the time base div either.

But, turns out there's massive 100Hz ripple on the -15V! it's pretty much dropping out to 0V then going to -20V. The multimeter was measuring about 14.8V which I thought was good enough, just shows I should have connected a scope to the power rails earlier.

I suspect the main bulk storage capacitor for the -15V may have gone bad, which would explain why it's 100Hz (rectified from 50Hz). But I guess there could be a faulty component pulling down the -15V too.

Vert Amp Outputs:





-15V rail:

[wn1fju]

I've been burned many times by voltmeter readings when checking power supplies, since the voltmeter will just give you an average value.  I always look at the power rails with a scope even if the voltmeter readings seem perfect.  In fact, even if a piece appears to be working correctly, it's always a good idea to consult the oscilloscope.

[bd139]

This will be a fun one. Whole power supply probably needs recapping. Then there’s the game of hunt the tantalum capacitor to play when it still doesn’t work. This journey is well worth it though. Good luck. 

There’s a good thread here regarding the power supply and using adapter boards for the capacitor https://www.eevblog.com/forum/testgear/tektronix-76037613-repairs/

[David Hess]

My first suggestion was going to be massive ripple from the power supply because of worn out input capacitors; I have seen it before.

[sean0118]

This will be a fun one. Whole power supply probably needs recapping. Then there’s the game of hunt the tantalum capacitor to play when it still doesn’t work. This journey is well worth it though. Good luck. 

There’s a good thread here regarding the power supply and using adapter boards for the capacitor https://www.eevblog.com/forum/testgear/tektronix-76037613-repairs/

Yeah that's what it really needs. I'm actually planning to auction these off for charity here on the eevblog though (they are from work so can't sell them). I don't want to spend much on them because I won't be able to recover the cost, but hopefully whoever wins the auction gives it the full re-cap it deserves. Thanks for linking the other thread, it's good to have it here as a reference.

My first suggestion was going to be massive ripple from the power supply because of worn out input capacitors; I have seen it before.

Yeah it makes sense. I'd like to know what causes that pattern when the Readout Intensity is turned up. At first I was worried it was a problem with the CRT itself, but it almost looks like transistor Vds / Ids curves rotated 90 degrees, might just be a coincidence though.



I replaced C811 which directly follows the rectifier on the -18V line. I only had a 10,000uF which is really only a temporary fix as C811 was 14,000uF. I'm not sure if its ripple current rating is adequate for the long term either. I also added a wire to join the two -18V planes on the PCB that were connected by the multiple pins of C811. Unfortunately I forgot to take a photo of this while the module was out.

The ripple has improved a lot, now only ~1V on the -15V rail, probably due to the replacement being too small. The CRT intensity still doesn't adjust and the trace is still off centre etc. But the Readout Intensity adjustment doesn't produce that strange pattern anymore.

But, the -15V is now at -20V, so I assume the excessive ripple killed the regulator. Hopefully the -20V hasn't damaged anything.



A10 Rectifier:



A12 Signals out:

[sean0118]

Okay, I made some more progress. 

The main pass transistor (Q958) for regulating the -15V tested okay. But the smaller one that drives it (Q956) did not, not a complete short, but didn't pass the 'diode test'. Vbe and Vbc were both about 0.47, but were 0.8V in reverse instead of open.

Q956 is a 2N5859, but I replaced it with a 2N1893 because it's all I had in that package. It's not a good replacement, 0.5A vs 2A max collector current. I won't be running it long until I get a proper replacement.

Replacing Q956 was enough to bring it to life. However, it seems the oscillator on the reference board is dead, even though it's set to 1kHz only 4V DC is coming out. The photo taken below is using the 1kHz ref from my Rigol. The horizontal scale seems a bit off, might be a calibration issue, or maybe I still have some voltage rail ripple.

[bd139]

That’s looking pretty hopeful 

[harrison@cs.unc.edu]

I had similar problems with one of my 7603's. No high voltage led me to the rectifier board. No +5 or +130. I tested all the filter caps and
found that 5 out of 7 were leaky or just open. I replaced all of them with newer versions that are 105c tolerant. I tried to buy low ESR, but
the proper values were back ordered. After replacing all the caps, the +130 was still missing. I had checked for the raw 130 from the xformer
and it was correct. Still no +130.  The bridge rectifier was testing good while in the circuit, but still no +130. The problem was that one of the
xformer taps was intermittent. i had to remove one of the replaced caps to get to that connection. I resoldered and had all the voltages
present at the regulator board.
I didn't have a trace because no plug ins were installed. I put in the 7B53A and powered up again. The magic smoke left the scope!!
Two tantalums on the vertical and one on the 7B53 gave up life. I replaced them all and a few others. Calibration did not suffer and it has
been a good scope for me. I have 3 more on the way to repair and hopefully sell later.
nChapelHeel

[sean0118]

Okay I narrowed down the horizontal issue a bit more, I think the problem is actually in the 7B50 horizontal module.

One thing that threw me off a bit is the 4V ref is 1kHz +/- 20%, so it's not good as a reference for the horizontal. So I used the 1kHz ref from my Rigol instead, which should be the most accurate source I have.

I went through the calibration procedure for the horizontal gain of the 7603 (as best I could without a calibration module). First I set the 4V ref to DC and calibrated that to my multimeter, I set it back to 1kHz and swapped the vertical and horizontal modules. This gives vertical traces spaced horizontally by the 4V ref, which allowed me to calibrate the horizontal gain independent of the horizontal module.

Except when I swap the vertical and horizontal modules back to their correct locations the horizontal scale is still incorrect. I put the modules into my Tek 7623A  and it's off by the same amount. I should be getting 10 divisions at 0.1ms/DIV but I'm only getting 8.4 divisions:


Calibrating the horizontal gain (4V 0.5V/DIV)


Calibrating the horizontal gain (4V 2V/DIV)


Tek 7603 1kHz 0.1ms/DIV


Tek 7623A 1kHz 0.1ms/DIV

[David Hess]

I went through the calibration procedure for the horizontal gain of the 7603 (as best I could without a calibration module). First I set the 4V ref to DC and calibrated that to my multimeter, I set it back to 1kHz and swapped the vertical and horizontal modules. This gives vertical traces spaced horizontally by the 4V ref, which allowed me to calibrate the horizontal gain independent of the horizontal module.

Standardization is only required if you expect to swap modules between oscilloscopes which was never really intended.

Before I got a standardizer plug-in, I modified my 7A16A amplifier with a couple of test points at the interface so that I can connect a multimeter to measure the DC and AC signal voltage at the interface, which as I recall is 25 millivolts per division but someone should check me on that.  Many multimeters have plenty of resolution and accuracy at that level to calibrate the mainframe better than required.  AC calibration can be done with a low frequency square wave and either AC RMS or AC averaging with a correction factor of 11%.

[sean0118]

Okay it died again! 

I left it on for a bit then came back to find the trace was gone. Measuring the power rails I found they were all much lower than usual, which pointed to an issue with the -50V rail as it's the reference for the others.

When turning the scope on I found the -50V rail increased very slowly and never reached -50V. I suspected the 5.6uF electrolytic capacitor C889 in the reference circuit had gone leaky, however I lifted one leg and found it to be okay.

I then found R890 (5.49k) had gone open. It biases the zener diode reference, so without it the zener voltage was much lower than normal and C889 charged very slowly.

I replaced it with a CPF15K4900FEB14 (5.49k, 1%, 25ppm, 1W), it's much smaller than the original resistor, however the service manual specs it as only needing 0.5W. I also re-soldered some of the regulator pins as they looked possibly cracked and some had way too much solder. One or two of the pads were already lifted slightly so it may have been touched up before.

With R890 replaced the scope works again and all the power rails are in spec.   

[youngwug]

What is the measured value after removing the R890(5.49k)?

[sean0118]

What is the measured value after removing the R890(5.49k)?

Hey sorry I just saw this, I think I measured that but I can't recall what the value was. I just measured the voltage over it as 40.39V (-49.32V to -8.93V), that's only 7.4mA / 0.3W so I think it's okay, I'm not sure what caused the original resistor to fail.

I think something else is wrong with it now though, the trace keeps fading out, disappearing completely at times. The trace intensity adjustment seems to have no effect. The display intensity for the 1ms label etc does work fine though...

Anyone have ideas what I should look at first?

edit: I forgot to mention, the power rails (-15V, +5V, +15V, +50V, +130V, -50V) look good, both DC and ripple. I'll have a look the the manual to check the CRT bias etc.

edit2: Now that it's had time to warm up the trace isn't disappearing but it's brightness is unstable and the trace intensity knob still has no effect. Seems to be a similar issue to this:
https://www.eevblog.com/forum/repair/tektronix-7313-trace-disapears/

[David Hess]

The DC restorer which levels shifts from low to high voltage in the z-axis circuit path could be leaking before failure.  You might inspect it in a dark room and see something.

[sean0118]

The DC restorer which levels shifts from low to high voltage in the z-axis circuit path could be leaking before failure.  You might inspect it in a dark room and see something.

Thanks, that seems quite possible, is the DC restorer in the HV module?

Since posting I have probed the Z-Axis test point with a scope (see pic below), I was trying to follow the 'Adjust CRT Grid Bias' instruction in the manual (pg 3-6). It seems there's meant to be a voltage offset at that point which is varied by the trace intensity control, however I only measure a constant 0V. But, varying the readout intensity does produce voltage spikes at that test point.

I think this might point to an issue on the Z-axis amp board? As a fault in the HV module would load down the line and effect the readout intensity too?

I've also ruled out a few other things, the plug-in modules work in the other scope, the intensity trim pot seems to be good (resistance measured at end of cable) and the CRT grid bias adjustment still varies the intensity.

[sean0118]

I can't catch a break with this scope, I accidentally left it plugged in over night, but with the front switch off. Went to turn it on but the mains fuse was blown, put a new fuse in (bad idea I know) and it blew that one in a few seconds...   How is it possible for one scope to fail in so many ways? 


edit: One of the +18V rectifier filter capacitors is short, either C813 or C814

[sean0118]

Okay, I got it powering up again, it was actually rectifier CR811 that was shorted, but I had already cut out C813 and C814 so replaced them too. They probably needed replacing anyway.

CR811 seems to be a D-72 package, I had some similar spares that were very close but 1mm too high. It gets sandwiched between the PCB and enclosure so the extra 1mm meant they wouldn't fit. I eventually found a suitable replacement from Vishay which is in the list below. I was under the threshold for free shipping so I ordered long life Kemet caps for C813 and C814, I also ordered a 15,000uF replacement for C811.


Vishay   VS-KBPC602PBF   200V   6A
KEMET   ALC70A153CD040   40V   15,000µF
KEMET   ALC80A103CD040   40V   10,000µF
SIBA   70-065-65/1.6A   Time Delay   1.6A

It powers up now but I still need to fix this z-axis / trace intensity issue...     

[sean0118]

Okay sorry to bump the thread again but I think I found the cause of the z-axis issue...   U99C is dead!

It's part of a CA3046 transistor array, the other transistors in the array test okay (just measuring BE, BC diode drops with a multimeter).

The exact same IC failed in my 7623A, maybe it's a common issue?
https://www.eevblog.com/forum/repair/tektronix-7623a-oscilloscope-repair-log/

[sean0118]

Okay, replacing U99C fixed the intensity issue, well it still seems a bit unstable but the intensity knob does work now.

But I left it running for an hour and something else has failed... this is fault #7, must be close to a record...

The readout display has now collapsed vertically, and it moves with the trace position (I don't think it's meant to do that?). If I feed in the 4V ref the trace is the correct amplitude but I can't get it to trigger. 

No Input:


4V 1kHz in (1ms/Div):

[David Hess]

The readout, and other traces, will shift with trace position because of thermal balance issues but it should not follow it.

[sean0118]

The readout, and other traces, will shift with trace position because of thermal balance issues but it should not follow it.

I'm starting to think the fault might be in one of the plug in modules (7A18N, 7B50), I can't get it to trigger when installed in the 7623A either. The 7623A doesn't have a readout board installed so makes it difficult to troubleshoot.

I'm guessing the readout might depend on the trigger system somehow?

[David Hess]

I'm starting to think the fault might be in one of the plug in modules (7A18N, 7B50), I can't get it to trigger when installed in the 7623A either. The 7623A doesn't have a readout board installed so makes it difficult to troubleshoot.

I'm guessing the readout might depend on the trigger system somehow?

The readout has different modes selected by a switch on the readout board.  One of the modes is free running where a readout is displayed at the time and also multiplexed during sweeps.  There is another mode where the readout is only generated after a sweep, which works better for photography and when the readout multiplexing would be annoying.

[sean0118]

The readout has different modes selected by a switch on the readout board.  One of the modes is free running where a readout is displayed at all time.  There is another mode where the readout is only generated after a sweep, which works better for photography and when the readout multiplexing would be annoying.

Yes good point, I tested that 'gate triggered' mode and it does work in that the readout is only displayed at the end of the sweep (but the display of the readout is still collapsed vertically etc).

[sean0118]

Okay, I opened the 7B50 horizontal module, tested and re-seated some of the transistors in the triggering circuit (they were okay) put it back in and the triggering works again...  might have been a poor connection somewhere?

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