Tek FG502 fails at high frequency end of dial - help needed!

[gcwills]

Hi all,

I have recently inherited several Tek TM500 series plugins and have managed to repair a few faulty units. A range of faults from a broken resistor at the input of a DC509 counter, a shorted tantalum cap in an AF501 bandpass filter and several others with blown indicator lamps replaced with LEDs.

However, I have a FG502 function generator which works in most settings, however it has an unusual issue with the output dropping out at the higher end of the frequency dial. The pulse and sawtooth output settings work fine with output being sustained across the extremes of the frequency dial. However when the square, triangle or sine functions are selected the output signal disappears when the frequency dial is advanced beyond 5-7 of the range. To make it even odder, this signal dropout only occurs when the frequency multiplier is set for X10, X10², X10³ but on other multiplier settings there is no issue. And to add, at lower settings of the frequency dial when the generator works fine the output frequency is accurate across all ranges of the frequency multiplier.

I have confirmed all rail voltages are correct with no ripple evident, have extensively cleaned the function and multiplier cam switches, have confirmed that the DC levels set by the frequency set pot and following opamps are correct, have tested the buffer FET Q200, Q230A and B, Q290 and Q292 and all of the diodes associated with the diode switches.

Here is the manual for the FG502: https://w140.com/tekwiki/images/1/11/070-1706-00.pdf

An unusual fault indeed! Any bright sparks out there who can throw light on this conundrum?? Cheers!

[BlownUpCapacitor]

Unlikely that this is the issue, but try reseating all the op amps. It's an easy check so why not do it.

If that didn't fix it, check the resistors in the image provided.

If those check out, probe the gate of Q200 with an oscilloscope and see if the triangle wave oscillation error happens there, or if the function of that is flawless.

If the error does appear at the gate of Q200, check U175, U170, and U140. U140 and U175 are the negative and positive current sources for the oscillator.

If those check fine, check Q140, and Q175., though at this point I feel like nothing else can really be checked.

As a last ditch check, check the timing caps in the provided image.

If none of these solve the problem, then I really, don't have much of an idea.

[gcwills]

Thanks very much for taking the time to respond and for your suggestions.
I did try reseating the opamps and also replacing them all to no effect.
I did check the resistors associated with the multiplier and they were all spot on apart from R150 and R186 which read a little over 4M. Replaced them with 3.9M 1% resistors with no effect.
And yes, the signal dropout can be seen at the gate of Q200 so the problem occurs in the circuit before Q200.
I did replace U140 and U175 and tested Q140 and Q175 - all good.
I did find that the associated carbon comp resistors R140 and R175 were reading around 575 ohm - higher than their value of 510 ohm and were replaced with 510 1% metal film - no difference!
I did check all of the timing caps and thought I had a leaky C154 but it turned out to be fine. The caps were tested for value and for leakage with a very sensitive leakage tester - all good.
So the hunt for the elusive issue continues!

[BlownUpCapacitor]

Perhaps something is not biased properly in the oscillator? But I wouldn't know how to check that as that would be INSIDE the op amps.

Check if a trigger output signal is present while oscillating, and during the failed oscillation.

Also use the VCF function to push the oscillator into its failure zone without using the dial.

Report back.

[gcwills]

Thanks again for your thoughts!

The trigger out signal is present while the oscillator is working and fails when the oscillator fails.
Good suggestion re the VCF function. I set the frequency dial to 1 and applied voltage from my bench supply to the VCF input. As expected, the frequency increased as the voltage increased and when the frequency approached the equivalent frequency set by advanced setting of the frequency dial the oscillator stopped. So yes, applied voltage into the VCF input produced exactly the same result.

I have spent quite some time on this issue today and have tested just about all the transistors and diodes in the oscillator circuit to no avail. The fault does not seem to be related to the DC settings as the tests with the VCF reflect. However, one possible clue to the fault was discovered - the collector voltage at Q292 which switches the ramp voltage on and off measured low (30-140mV depending on the multiplier settings) to measuring -3V when the oscillator stopped. The challenge with fault finding this kind of oscillator circuit is it is a closed feedback system so any issue in that loop is reflected everywhere in the loop. I found a very helpful resource on the Tek Wiki here: https://w140.com/tekwiki/images/2/2f/Repair_FG502_English.pdf   that refers to the closed loop scenario.....

[BlownUpCapacitor]

the collector voltage at Q292 which switches the ramp voltage on and off measured low (30-140mV depending on the multiplier settings)

That seems about right. Pulldown for Q175 changes with the multiplier setting, changing the flow of current through CR170.

to measuring -3V when the oscillator stopped.

Whoa, I think you got something!

Looks like the oscillator is locking itself in the negative cycle it seems.

But the problem is still evasive.

Maybe check the voltage across R155 during failure, and working? A negative voltage at the collector of Q292 indicates current flowing through it. Q292 is controlled through Q230A.

Check the circuit pictured below.

Also try removing CR204. Take it out of the circuit and see what happens. Earlier versions didn't have it.

Frankly I'm out of good ideas. Maybe someone else on the forum can help?

[gcwills]

Success
After a lot of extra time spent on this challenging function generator I stumbled upon a fix for the oscillator dropout. I focused my attention on the area around Q292 and Q290 as it appeared that there were bias issues associated there.
Partly out of desperation I added some resistors from the collector of Q290 to ground and the oscillator was able to maintain oscillation a little further up the dial. I then tried the same approach to the collector of Q292 to no effect so I wondered if increasing the value of R155 (the Q292 load) would have a positive result. As it turned out it did and adding a 100 ohm resistor in series with R155 (820 ohm) allowed the generator to operate fully to the high end of the frequency dial. I then removed R155 and the 100 ohm and substituted a 1K 1/2W resistor. I was able to confirm that the generator worked fine across all of the ranges and waveform settings. The generator was then calibrated according to the manual.

So was this a bodge or a repair? I think it is a repair as the generator didn't have any faulty component that I could find and worked albeit in a limited way. It just failed to oscillate on certain settings. My guess is that there was sufficient drift in components around the oscillator causing this issue. Increasing the R155 value a small amount increased the loop gain enough to resolve the issue. Of course I am open to better brains than mine with better explanations?
Big thanks to BlownUpCapacitor for your assistance - appreciated!

[gcwills]

To add to the story of my FG502 issues, I came across this reference to a similar issue with a FG502 on the TekScopes group:
https://groups.io/g/TekScopes/message/180842?p=%2C%2C%2C20%2C0%2C0%2C0%3A%3Acreated%2C1%2CFG502+doesn%27t+start+at+some+specific+settings%2C20%2C2%2C0%2C81500580

Of interest to me was the comment "Perhaps some of the transistor gains decreased over time and now positive feedback coming back through C247 is not strong enough."

[gcwills]

The FG502 is the gift that keeps on giving! Having sorted the issue of signal dropout at the higher end of the frequency dial I calibrated the unit and was about to box it up when the generator output began drifting in frequency 
Any movement of the chassis caused significant frequency shift. Having learned that the generator frequency was determined by varying the DC voltage into the VCO, I checked the output of U135 and U170. The output DC level of U135 was stable, but that of U170 drifted significantly when the chassis was moved. Further tests showed that R170, the SYM trimpot was faulty and was replaced. The generator then worked with stable output. The covers could then be replaced and the generator put into service