HP 3310A Repair Advice

[wriely]

I recently acquired a HP 3310A for repair.  The repair started out pretty straightforward, but I've now run into some difficulties.

When I first powered it on, I noted that all of the power supply rails were reading way off (low).  I traced the problem to a bad 9V regulator diode (CR26) and a bad pass transistor (Q31).  I replaced both of these parts and now the power rails are in specification (DC and AC).

Next, I noted that there was no output (other than a DC level) for all ranges of 10x and higher.  The following descriptions are with the output amplitude set to the 9 o'clock position.  They scale upwards as I increase the amplitude control.  On 1x I get an output in sine, triangle, and pulse settings but the shape is completely wrong.  On sine it is essentially +2V with some negative pulses on it.  On triangle, it is -2V with a positive edge and a negative ramp followed by an equal duration of -2V before the cycle repeats.  Dropping to 0.1x and lower the waveform expands to cover +/- 2V but the waveform shapes are still completely wrong.

Examining the output of the triangle amplifier at TP4 reveals that this is the source of the problem.  On 1x, the output waveform only goes from 2.5-6.5V and only ramps up.  On slower settings, it goes from -4 to 6.5V but again only ramps up.

I noted that Q18, Q19, and Q20 had been replaced in an earlier repair (NTE123, NTE128, and NTE129 respectively).  A junction test of all of the transistors in the triangle amplifier circuit showed no problems.  Same thing for all of the diodes in the circuit.  I extended my testing to the +/-5V level detector.  All of the transistor junctions appeared to be fine.  However, when I tested the tunnel diodes, they tested as shorts in both directions (out of circuit).  I know this is wrong so I eventually managed to get my hands on some replacements.  The replacements measure around 0.5V drop in the forward direction as expected.  I carefully placed them back into the sockets and confirmed the orientation matched that of the schematic.  I powered the circuit back up to resume testing it but unfortunately the behavior remained the same.

The troubleshooting tree diagrams in the service manual were of no help, so I decided I would check DC voltage levels in the triangle amplifier and surrounding circuits to see if I could find the issue.  Unfortunately, there appears to be an elaborate feedback loop between the current sources, triangle amplifier, +/-5V level detector, and low frequency feedback amplifier so I'm not sure how useful a bad voltage reading will be.

I found the voltages in the tuning amplifier and inverting amplifier were approximately correct.  However, once I got into the bias network I started seeing some big anomalies.  The emitter of A1Q5 was at 6.7V when it should have been +4V.  A1Q5's collector was +6.6V instead of -1.5V.  A1Q6's collector was 6.2V when it should have been -1.0V.  However, A1Q6's emitter was -4.0V just as was indicated in the schematic.  TP5 was 6.2V rather than the 1.2V it should have been.  There were more discrepancies in the +/-5V level detector, but without breaking the feedback loop there was really no way for me to be sure where the root cause of the problem is.

I spent some time confirming junction voltages on all of the transistors, JFETs, and diodes in the bias network, +/-5V level detector, and low frequency amplifier circuits but found nothing else obviously bad.  While testing CR1 and CR2 in the tuning amplifier and inverting amplifier with power on, I did observe that CR2 only had 4.4V across it rather than the 6.2V that CR1 had (they are both supposed to be 6.2V zeners).  My schematic indicates that the output of A1IC2 should be -16.5V, but I believe that is an error and it should be +16.5.  I was reading +19V here (as compared to -16.8V for the output of A1IC1).  However, the voltages at TP2 and TP3 were nearly equal and opposite so I'm not sure this means much.  CR2 may have been replaced in a past repair since it was not of the same type as CR1.

I'm looking for advice or suggestions for how to further troubleshoot this circuit.  It looks like it should be possible to at least partially break the feedback from the current source and sink by disconnecting the leads from the junction switch to the A1 PCB.  However, I'm fairly confident the current source and sink are operating correctly already so I'm not sure that buys me anything.

--
Walt

[m k]

You need a scope.

If you need a manual pick the last one.
https://www.opweb.de/en/manufacturer/hewlett-packard/3310a

[wriely]

Hi m k:

The scope views are pretty boring when the instrument is set to the conditions specified in the Service Manual.  Namely a range of 10 and the function switch set to sine.  The output waveforms and most of the test points including A1TP4, A1TP5, A1TP6, A1TP7, A1TP8, and A1TP9 are all at DC levels -- many of the values are specified in my initial post.  In these conditions, the +/-5V level detector should be flipping the waveform at TP4 when it reaches +/-5V but this is not happening.  I've replaced the tunnel diodes A1CR21 and A1CR22 with new units and performed in-circuit tests of all of the transistors in the triangle amplifier and in the +/-5V level detector and found no issues.

If I drop the range to 1 or lower, the generator switches from using the +/-5V level detector to the low frequency feedback detector and I start to see some non-DC waveforms.  Here are scope captures of TP4.  The first image is with the range set to x1 HP3310A_A1TP4_x1.png (37.63 kB. 1024x600 - viewed 20 times.), and the second with the range set to x.1.





Curious to hear your thoughts.

[m k]

I'm reading the manual 03310-90002
Serials 1048A03650 and below
June 1971

Can you verify that all selectors are somewhat operational?

Manual pages 65-66/97, I can't visualize the range selector to triangle amplifier input.
The one with three shorted and one not in between.

Left Right of manual page 67/97 is a waveform example for A1TP11 and other side of A1CR3.
Seems that the emitter side of A1Q6 is not doing much.

[wriely]

Hi m k,

> Can you verify that all selectors are somewhat operational?

Here's what I've done so far:  I visually checked all of the switch wiring to confirm it matches what is specified in the Service Manual.  Next I rotated the wafer switches throughout their range and verified I could see deflection for all visible contact points.  Finally, I wiped all contacts I could get to with a swab soaked in DeOxit.  I have not yet performed conductivity tests on every wafer switch, just a couple that were in the path of signals I was following with no problems seen.

> Manual pages 65-66/97, I can't visualize the range selector to triangle amplifier input.
> The one with three shorted and one not in between.

Are you talking about S2E(R)?  This appears to add selected amounts of capacitance to the output from the current source.  Starting with 1uF for the lower frequency ranges, then decreasing first to 0.1uF at Range=100, then .01uF at Range=1k (with 2.7 ohms series resistance), etc.

> Left Right of manual page 67/97 is a waveform example for A1TP11 and other side of A1CR3.

Yes.  These are stuck at DC levels (6.5V for the collector of A1Q5 and 6.2V for the collector of A1Q6) with the Range switch set to 10 (or higher).  The following two images are of the collectors of A1Q5 and A1Q6 (TP11) respectively with the Range switch set to 1 and the frequency dial set to 20):





These don't look anything like the sample waveforms on the right margin of p67 of the SM.  It is also interesting to note, that on the R=1 setting, I only get a DC level with the frequency dial set to 25 or higher.  As I go up in frequency, the lower peak of the waveforms moves closer and closer to the upper peak until it becomes DC at approximately the 25 Hz dial setting (360 Hz on the scope).  Conversely, the amplitude grows (lower peak drops) as the frequency dial is lowered.

> Seems that the emitter side of A1Q6 is not doing much.

FWIW, the emitter of A1Q6 has the correct DC operating point.  A1TP5 (cathode of A1CR4) is 6.2V which is 5V higher than the SM specified value.  This suggests to me that the +/-5V Level Detector circuit (p71 of the SM -- I think we're using the same one) is stuck although this likely depends on the operating mode since it is not actually used for Range=1 and below.

Edit:  All DC measurements in this and in my initial post were taken with the frequency dial set to 50 as specified in the SM.  Lower settings reduce the current supplied by the current source and sink and will certainly change some of the measurements.

[m k]

I'd say that keep 0.1 range until you can get that 2nd example waveform happening.
Now your Xp1 picture is 1st example rotated around, X1 also, but top is not so clean.

Theory of operation says that positive side act first, so inverting amp and it's clearly present.
Then negative side start discharging the integrating capacitor.

Your negative side of Xp1 is very comparator like.
Does it change if you pull CR4 up?

Maybe you've selected a ramp and a pulse.