Low-distortion Audio-range Oscillator

[dirkwright]

So, we have the SG 505 phase shift oscillator, the HP 239A bridge T oscillator, and the Cordell SVO. All of them are buildable. I have no clue which to choose.

[coluke]

If you have no experience at all in low THD generators I would suggest you to begin with the simplest one, just to experiment a bit and see what kind of problems you have to face to build and * measure * it - for reference, the full schematics of my clone is in post #255.

Ciao,

L.

[dirkwright]

Originally Posted by coluke

If you have no experience at all in low THD generators I would suggest you to begin with the simplest one, just to experiment a bit and see what kind of problems you have to face to build and * measure * it - for reference, the full schematics of my clone is in post #255.

Ciao,

L.

OK, I must have missed your schematic, obviously. Thanks for sharing.

What's the 1kohm trimmer near U1 for? and the other one by U6?

Is the low input bias of a jeft opamp like the OPA2134 critical to the performance of your oscillator? If not, then there are lower noise opamps out there that could be used instead.

[jcx]

Originally Posted by dirkwright

I'm not an expert, but I would say that if an oscillator was better than 1ppm THD with noise @ -130dB or so, over the range of 10Hz to 100kHz, then that would be a pretty good oscillator.

then you're done already - just buy a DAC:


ES9018 SABRE32 Reference 32-bit 8-Channel Audio DAC
DNR (dB) 135 (mono) 133 (stereo) 129 (8ch)
THD (dB) -120

[dirkwright]

Originally Posted by jcx

then you're done already - just buy a DAC:


ES9018 SABRE32 Reference 32-bit 8-Channel Audio DAC
DNR (dB) 135 (mono) 133 (stereo) 129 (8ch)
THD (dB) -120

Well, as I said, I'm no expert. Maybe these oscillators we are considering are better than that? I don't know.

[KSTR]

Any experts on SG505's here on this truly exiting thread?

Both my colleage and I recently purchased one of these and we already tried a few things to improve the residual harmonics but to no avail. I don't exactly find the SG505 easy to understand in all details (AGC section, notably) but I don't claim to be an expert. Both units are still fully within spec (as verified on the AP2722), so it's really about tuning the "last bit out of them".

So, anybody have a hint to known SG505 tricks (except for getting the hum down which is managable)? Not that I would expect the Tek guys didn't try tweak the hell out of it at their time, but after so many years there might be a chance that someone found ways to improve it with moderate effort.

[Samuel Groner]

So you do have your own oscillator up and running.

Of course--I'd not comment at this detail if I didn't had the background.

If you have no experience at all in low THD generators I would suggest you to begin with the simplest one, just to experiment a bit and see what kind of problems you have to face to build and * measure * it.

Definitely. Building low distortion oscillators is a steep learning curve!

The schematic is only half the rent--bad layout can completely ruin performance.

What's the 1 kOhm trimmer near U1 for? And the other one by U6?

The former pre-trims the Q of the oscillator such that J1 is working with just as much voltage across it as necessary. The latter adjusts the symmetry of the rectifier to minimize ripple.

Then you're done already - just buy a DAC:

ES9018 SABRE32 Reference 32-bit 8-Channel Audio DAC
DNR (dB) 135 (mono) 133 (stereo) 129 (8ch)
THD (dB) -120

An ES9018 will not have -130 dB dynamic range at full scale. This figure is measured at -60 dBFS and degrades above ~35 dBFS. Also it would surely not give 1 ppm distortion at 100 kHz!

Maybe these oscillators we are considering are better than that? I don't know.

A reasonable and realistic design goal for a very good DIY oscillator is perhaps:

* Frequency range 10 Hz to 100 kHz, in 4 decades
* Amplitude range -60 dBu to +20 dBu
* Harmonics below -125 dB for 20 Hz-20 kHz
* Harmonics below -110 dB for 10 Hz-40 kHz
* Harmonics below -100 dB for 10 Hz-100 kHz
* THD+N (20 kHz BW) below 110 dB for 20 Hz-20 kHz
* THD+N (80 kHz BW) below 100 dB for 10 Hz-40 kHz
* THD+N (500 kHz BW) below 90 dB for 10 Hz-100 kHz
* Settling below 1 s for 100 Hz-100 kHz
* Settling below 5 s for 10 Hz-100 kHz
* Amplitude flatness below 0.1 dB for 10 Hz-20 kHz (ref. 1 kHz)
* Amplitude flatness below 0.5 dB for 10 Hz-100 kHz (ref. 1 kHz)

Samuel

[PChi]

Originally Posted by Bob Cordell

HI PChi,

As a point of reference, the oscillator I have in my THD analyzer from 1981 settles to 0.1dB at 20Hz within 1 second and to within 0.01 or so within 3 seconds. This is for either decade changes or frequency changes within a range. Total analyzer THD at 20Hz is less than 0.006%, but I think most of that is in the analyzer side and that the oscillator is much better than that.

The SV oscillator uses a simple full-wave agc detector, but uses speed-up diodes across most of the integrator input resistance. Also, the agc detector capacitors are switched for the decade range and are all precharged by each other to the nominal operating value from whatever range was previously in use. Both of these techniques speed up settling when frequency or range is changed.

The design is shown in the THD analyzer article on my website at CordellAudio.com - Home.

Cheers,
Bob

Hello Bob,
Thanks for the website. Sorry for the dely in replying. I am back in work so cash rich but time poor. I remember trying to get a copy of Audio here in the UK for your article without success.
My design goals were:-
To learn KiCAD.
Continuous tuning (in ranges)
All mounted on a PCB for consistency.

I wasn't bothered about settling time.
I used a Phase Shift Circuit to minimise gain changes with frequency adjustment using unmatched dual ganged potentiometers. But this increases distortion with the high operational amplifier common mode signal.
Sample and hold because I have had problems with loop stability in the past.
High signal level for low noise but reduced linearity.

If I designed another I would consider using an analog multiplexer with close tolerance surface mount resistors which are quite reasonably priced. Multipling DACs don't seem that linear.
Lest see some more design ideas realised with test results.

It's going to be hard to compete with AP because it's their job and they will have spent many hours which the amatuer can't do.

[coluke]

Originally Posted by Samuel Groner

Definitely. Building low distortion oscillators is a steep learning curve!
(...)
The schematic is only half the rent--bad layout can completely ruin performance.
(...)
The former pre-trims the Q of the oscillator such that J1 is working with just as much voltage across it as necessary. The latter adjusts the symmetry of the rectifier to minimize ripple.
(...)
A reasonable and realistic design goal for a very good DIY oscillator is perhaps:

* Frequency range 10 Hz to 100 kHz, in 4 decades
* Amplitude range -60 dBu to +20 dBu
* Harmonics below -125 dB for 20 Hz-20 kHz
* Harmonics below -110 dB for 10 Hz-40 kHz
* Harmonics below -100 dB for 10 Hz-100 kHz
* THD+N (20 kHz BW) below 110 dB for 20 Hz-20 kHz
* THD+N (80 kHz BW) below 100 dB for 10 Hz-40 kHz
* THD+N (500 kHz BW) below 90 dB for 10 Hz-100 kHz
* Settling below 1 s for 100 Hz-100 kHz
* Settling below 5 s for 10 Hz-100 kHz
* Amplitude flatness below 0.1 dB for 10 Hz-20 kHz (ref. 1 kHz)
* Amplitude flatness below 0.5 dB for 10 Hz-100 kHz (ref. 1 kHz)

Could't agree more

As far as the opamp is concerned: you need both high speed and high GBW to keep THD low at high frequencies, high input impedance, very low noise current density - in this particular application an OPA2134 performs definitely better (at low to mid freqs, at least) than, say, a LME49710, although this one has 1/3 the noise voltage density of the former. The original 239A employed an HA-2625, which is unfortunately classified either as obsolete or retired by Intersil - it was a bipolar uncompensated opamp featuring high speed (100 MHz GBW, > 60 dB OLG @ 100 kHz, SR 35 V/usec), high input impedance (300 Mohm), low noise current density (0.16 pA/sqrt(Hz)): these are not-so-common specs even in modern opamps, unless you are going to choose expensive beasts like the OPA637 - all in all, OPA2134 is a good starting point...

L.

[davada]

Originally Posted by 1audio

I have a similar solution that is mostly discrete if you prefer. It would be cheaper as well.

Hi Demian,

I'm interested in looking at everything. If it's not too much trouble can you post it?