SV Oscillator PCB group buy

[dirkwright]

Originally Posted by AKN

Yes, but if the primaries differ slightly (in real life). Can you get the same performance by taking feedback from just one winding? Me thinks that both windings should be included.

Performance is actually not much better with the Jung circuit shown in the document I cited earlier. This circuit has the 2nd harmonic at only about -97dB down @ 20Hz, 1.4 Vrms output. Transformer has permalloy core, 300 turns, DC resistance of 18 ohms per winding, etc.

[AKN]

Have you tried to adjust gain in lower OP-amp as in circuit by Ljung?

[dirkwright]

Originally Posted by AKN

Have you tried to adjust gain in lower OP-amp as in circuit by Ljung?

Yes, I varied R5 by +/- 100 ohms with much worse THD each way.

While these are probably the best that can be had from a line level output transformer, we're still talking about a second harmonic down about 90dB or so. Since I don't have all of the particulars of the transformer, I have to guess at some of the parameters. The Jung circuit is slightly better than the one I originally used. I'm still twiddling with it.

[AKN]

Originally Posted by dirkwright

Yes, I varied R5 by +/- 100 ohms with much worse THD each way.

While these are probably the best that can be had from a line level output transformer, we're still talking about a second harmonic down about 90dB or so. Since I don't have all of the particulars of the transformer, I have to guess at some of the parameters. The Jung circuit is slightly better than the one I originally used. I'm still twiddling with it.

Ok, I hope the distortion creeps down further and further...
.

[dirkwright]

OK, now we are getting somewhere. I guess the devil is in the details. I have the transformer windings in parallel, BTW, just as before. THD is 5x10^-5%, 2nd harmonic is down -130dB from the fundamental @ 20Hz. Performance is just as good @ 1kHz. Adding buffers doesn't really help. So, thanks for the suggestions. Noise is extremely low as well.This one is worth building.

[AKN]

Great!
How do different secondary loading influence distortion, any need to reset null?

[dirkwright]

Originally Posted by AKN

Great!
How do different secondary loading influence distortion, any need to reset null?

I don't know. I doubt it matters.

[AKN]

Maybe a misunderstanding, i.e. I wasn't clear? I thought if you had your simulation up and running it would be easy to change transformer secondary load (R6).

I'll be implementing transformer output as well. Really nice to have galvanic separation from the noisy PC ground that jerks around with high frequency.

[dirkwright]

Originally Posted by AKN

Maybe a misunderstanding, i.e. I wasn't clear? I thought if you had your simulation up and running it would be easy to change transformer secondary load (R6).

I'll be implementing transformer output as well. Really nice to have galvanic separation from the noisy PC ground that jerks around with high frequency.

Oh, I understand you mean the load resistor. It shouldn't matter, but I don't really know.

I recommend a transformer with a high nickel core (80%). They simulate (and in real life perform) much better than anything else.

[AKN]

Hi,

Doh, I get it now. Needed some time (sleep) to understand.