Oscillator Distortion

Sorry, the trace of the 'ripple' should have been this link:

Looks like an AC-coupled square wave with a too-short time constant, which probably means that either the oscillator gain is running wide open, or the second stage gain is too high, due e.g. to its feedback loop being open.

Is the output amplitude close to the knob setting, or is it way off?

And can you get your hands on a schematic?

Cheers

Phil Hobbs

I'll see what I can do, Phil...

My guess was that the light bulb filament broke from the shock.

It's probably the founding HP 200A Wien Bridge audio oscillator circuit from 1939.

.

Joe Gwinn

That's a bit tricky to ascertain at the moment. I might be able to verify it later. However, I have managed to get the relevant section of the schematic. The supply is just under 12V. The nomenclature is a bit odd. The core of the oscillator is comprised of VT10, VT11 and VT12. All the other stuff beyond those to the right is for shaping and whatnot.

Somewhat unusually for me, this is not an HP instrument. It was made by some company called Venner in London. It did come with a very helpful manual which gives expected waveforms at various key points in the circuit. As you might expect, it shows sine waves for the TPs in the osc section.

HP's patent ran out a very long time ago. The bit about tungsten filaments was the tell. Read the HP patent - it's short.

Joe Gwinn

Maybe the light bulb broke.

It would have to have failed short-circuit for the gain to be too high, and that’s not the usual failure mode.

Cheers

Phil Hobbs

Anyway, it looks like the gain's too high, then. I don't have access to the generator again for a day or two, but I will focus on that aspect and hope to identify the precise issue through that. Thanks, Phil; you're a pal. ;-)

CD

Is there a chance that the impact broke one of those fine tungsten filaments? Do they read low-Z when cold, as they should?

The time reponse of the filaments, as compared to the Q of the basic oscillator, must be interesting. There are no doubt papers somewhere.

Venner were famous for their time switches made since forever and used in street lights before photocells.

It's one thing I need to look at, if only for the sake of completeness, next time I have access to it. Since these are so hard to replace, I'd really rather deal with my original suspicion that something got shorted out in the fall.

Yes, the TH2 thermistor on the schematic has gone open-circuit. Now I do have some spare oscillator thermistors which I rifled from old test gear I broke up for spares and always made sure I got the thermistors out of wein-bridge circuits. Unfortunately, the ones I have are unlikely to have the same characteristics as the failed one. But I'm reasonably optimistic I can correct for that by going through the re-alignment procedure in the manual.

This one's the TSA628. I have the service manual for it, but it only quotes Venner part numbers for all devices listed and "R23" for the type/ value of this thermistor. I don't think either of those numbers would be of much use today. TH1 is "Type R52" so while they appear identical, it seems they aren't. TH1 and TH2 both *appear* identical: glass encapsulations about an inch long by 3/8 wide at a guess. They're juxtaposed together on the board. I'm guessing - it is only a guess - that TH2 (the failed one) functioned as some sort of AGC to stabilise the amplitude of the oscillator and the TH1 was the actual Wein-Bridge element as is commonly understood in this type of oscillator. That would account for why - as Phil Hobbs observed - the gain has gone up enough to run the output into the supply rails and give rise to the distortion I'm seeing.

I've got an R54 somewhere. I think they are still available as NOS.

I'm just wondering if a modern bead thermistor would work as well. The ones in this 56 year old piece of kit are large, glass-encapsulated types, but maybe that was just the way them made them back then when everything was bigger. I might try a few bead thermistors in place of the busted one once I've removed it just out of curiosity.

The advantage of the vacuum encapsulated ones is that it requires vastly less power to heat them, and thus I doubt an open bead one would easily work at the same signal level.