-hp- 3311A function generator distortion

[TimFox]

Looking for someone with experience on adjusting the sine-wave shaper for lowest THD.
I recently obtained a good deal (eBay) on a 3311A function generator for about $50 "as-is".
It was in good physical shape, originally from a university laboratory.
At the low price, I was not surprised that I needed to repair it.
I was surprised that both 7818 (TO-3) regulators were blown, but I found some on eBay and the unit functioned well after replacing both 7818s and spraying all the switches (no useful output before cleaning the switches).
The adjustments went well (hard to read PDF from the manufacturer's site), but I have a minor problem with the distortion adjustment.
Before THD adjustment, following the manual, I adjusted the triangle amp offset, square wave symmetry, and tuning calibration (in that order) to reach specification.
I have an -hp- 339A THD analyzer, that reports the percent distortion directly, independent of the input level.  Before adjustment, the maximum output at 10 kHz (10 x 1 kHz) was 3.6 V at -25 dB THD (4.25%).  The manual warns that adjusting R149 can result in low output, and I found that to be true.  I was able to achieve -30 dB THD at 3.5 V rms output (into 600 ohm load); lower distortion was possible, but the output fell below the 10 V p-p (3.53 V rms) level.  The strange thing is the other two (positive-peak and negative-peak) pots:  on both pots, minimum distortion was achieved at full CCW setting of the pot, applying zero bias to the diodes.
Should I have started with the peak-adjust pots?  That seemed only to increase the distortion.
The manual only directs to adjust the three pots for minimum distortion, being careful with R149 not to fall below 3.53 V rms.

[Kleinstein]

The right adjustment should not have the pots all the way at the end. I would start with R149 at about center. Than adjust the two other pots in parallel to get minimal distortion. This should not happen at zero setting. Otherwise something seems to be broken. If all goes right R149 could be adjusted for lowest distortion. If a large change needed repeat with the two other pots.

With so much distortion the distortion might be even visible on the scope. So one could see what type of distortion is present. It's possible that some of the diodes or the following amplifier is broken.

[edavid]

Since he has the HP 339A, he can view the residual directly.

[TimFox]

Thanks for the useful advice.  I'll get back to this repair next week and report.  Except for the usual non-continuous derivative at the peak, the appearance of the 3% distorted sine on a CRO is reasonable.

[TimFox]

I found a problem with my original THD test set-up:  following the 3311A manual, I did the adjustment at 10 kHz (10 x 1 kHz), but I had accidentally left the LP filter on the 339A set to 30 kHz, which almost meant that I was measuring only 2nd-order distortion, which would be dominated by asymmetry between positive and negative peaks.
Anyway, starting over I noticed that the two "peak" adjustments (R163 and R165) could get nice-looking peaks (reduced "corner" at extrema), but the indicated THD wasn't great.  I still noticed that adjusting R149 could obtain low THD at the expense of reduced output.  Therefore, I set R149 to get the required 3.5 V rms (into 600 ohms) and tweaked the other two.  In fact, the best THD was obtained with R163 CCW, but R165 slightly higher.  Adjusting either or both away from CCW gave a good looking scope trace, but higher THD ("fat" sine waves to reduce the corners).
The 3311A spec is THD < 3% at 3.5 V rms, and I obtained about 2.5% at 3.5 V rms, measuring at 1, 10, and 100 kHz with appropriate filters on the 339A.