Repairing Tektronics CFG250 function generator

[jadetortoise]

I recently picked up a Tektronics CFG250 function generator on ebay. It powers on but the output wasn't working. I found the schematic online and have verified that all the voltage rails in the power supply work (-30, -15, 5, 15, 30V, though both 30V rails were closer to 25V). I also verified that the triangle-wave, square-wave, and sine-wave generators work all the way up to where they enter the output amplifier.

That's where I get lost.  I'm trying to understand how the output amplifier works, but the first pair of transistors gets me stumped. It kind of looks like some sort of current mirror, but has the emitters connected rather than the bases. I'm guessing that this configuration lets the DC offset be applied to the output of 2nd base, with the collector carrying the AC signal to the rest of the amplifier. Is this supposed to be the case? While the 3rd pair of transistors looks to be a push/pull configuration, I'm confused by the middle pair, which seem to be driven only by the 15V rails.

Currently the output is stuck at ~-25V and Q13 gets very hot. I'm probably going to replace Q13, but I can't trace the AC signal that far into the circuit and was wondering if I should be looking to replace the other transistors. I've already verified that all the electrolytic caps in this circuit are fine.

[duak]

Function generators often use some slightly unusual configurations to increase the speed and bandwidth.  Q11 is a common base voltage amplifier that has less voltage gain than the standard common emitter, but it has an extended bandwidth.

Q14 and/or Q12 are probably shorted between collector and emitter.

 

[Tomorokoshi]

Mine here at work had similar symptoms. I fixed it by replacing one of the PNP transistors, either Q11 or Q14 from what I can tell from the schematic. I used this part:
https://www.digikey.com/products/en?keywords=2N2905ACS-ND

I also bought the NPN complement:
https://www.digikey.com/products/en?keywords=2N2219ACS-ND

I bought two each of the PNP and NPN, but I only used one PNP. I can't find my notes on the repair so I don't know which location I replaced.

[jadetortoise]

Good call on the short being on Q14. Q14 is responsible for the final negative side output and was dragging the output to the negative rail. I removed it and the positive portion of the output waveform reappeared. I guess Q13 was getting hot because it was effectively conducting to the negative rail. Time to buy a couple replacements like you did. Probably only need the Q14 though. I'm glad it's not Q11/Q12 because those part numbers are obsolete now.

[magic]

FYI, Q9/Q10 is a long tail pair, an absolutely classic circuit which makes the input stage of almost all DC coupled amplifiers.
Q11 configuration is known as folded cascode and Q12 is its active load (a constant current sink).

If the waveform reappeared, Q9 to Q12 are good.

[jadetortoise]

Thanks for the help. It so much easier to search for information about a configuration when you have a name to attach to it, rather than just scanning images on google for something similar.

[David Hess]

I do not remember seeing that configuration before but I am sure I have.  Q11 and Q12 are configured as a current source and current sink with an idle current set by Q12 at about 5.4 milliamps.  Q11 would supply double that current from a halved emitter resistor however the differential pair sinks the excess current from Q11's emitter to control the output voltage. 

The complementary pair of of Q10 driving Q11's emitter is called a folded cascode and is common to prevent stacking of each stage's shift in voltage level.  Since the emitter of Q11 holds the collector of Q10 at a constant voltage just like a normal cascode, Miller feedback is reduced increasing gain-bandwidth product.  Q11's sees considerable Miller feedback however its base has a very low AC impedance, unlike Q10, so the effect is minimized.