Super Low Distortion Ultra Pure Audio Oscillators Revisited

[JR.]

I really liked polystyrene.... somewhere I had some 0.1uF polystyrenes that as fat as my thumb.

JR

[mediatechnology]

I have your 8200 pF polystyrenes...

I need to keep both J1 and J2.
Though they may sim the same with 1 or 2 links I just noticed that for a balanced-only configuration, where there would only be a single 10 nF at C4 and a single VR, there would no longer be symmetry.
So the J1/J2 shunts need to remain split.

[terkio]

I do not understand about J1 J2.
I see a balanced-only configuration as a underpopulated board, no J1 no J2 no VR2 but a strap, four caps become one.

[mediatechnology]

If the two hot sides of J1 and J2 get connected together because one J is eliminated, C4 is populated and C6 becomes a link, then the bottom side of VR2 connects directly to the node of R9/R10. (C3/C5 not populated and J open.)

J2's left-hand connection should instead go to the node formed at the center of C3/4/5/6.
J2 is only installed when Nacho is switchable.
J1/J3 then switch balanced/unbalanced allowing a DPST switch to be used.

If the board is built balanced only then neither J1/J2/J3 are installed and the bottom of VR2 doesn't get connected to R9/R10 through a linked C6 because J2 is open.

Make it like this:

[terkio]

For the balanced only version, I would manage going from 4 caps to only one, by means of footprints to allow the cap to jump over the J1 node.

[mediatechnology]

I could do a C4A or C4B with C4B jumping across the jumper to the R9/R10 node as you suggest.

[fredbloggstwo]

Hi folks,
Just been through this thread and it looks quite exciting.

I have a couple of questions though:

1 what would be involved in turning the oscillator into a switched full range device (20 to 20k)

2 if you are trying to get the most from a 5532/5534, have you seen the trick about using an external diff pair connected to the compensation pins on a 5534. It is in the dim and distant past of my memory, and I can't find my reference for it, but bypassing the 5534s front end allows one to drop the noise down quite a bit and get around those pesky input diodes if needed.

Cheers

Mike

[JR.]

Hi folks,
Just been through this thread and it looks quite exciting.

I have a couple of questions though:

1 what would be involved in turning the oscillator into a switched full range device (20 to 20k)

switches, more components, and probably a few tweaks. But Wayne is the man for this one.

2 if you are trying to get the most from a 5532/5534, have you seen the trick about using an external diff pair connected to the compensation pins on a 5534. It is in the dim and distant past of my memory, and I can't find my reference for it, but bypassing the 5534s front end allows one to drop the noise down quite a bit and get around those pesky input diodes if needed.

you could probably find it and other similar circuits in a search right here.....

Cheers

Mike

based on the results Wayne is getting why bother?

JR

[mediatechnology]

Hi folks,
Just been through this thread and it looks quite exciting.

I have a couple of questions though:

1 what would be involved in turning the oscillator into a switched full range device (20 to 20k)

2 if you are trying to get the most from a 5532/5534, have you seen the trick about using an external diff pair connected to the compensation pins on a 5534. It is in the dim and distant past of my memory, and I can't find my reference for it, but bypassing the 5534s front end allows one to drop the noise down quite a bit and get around those pesky input diodes if needed.

Cheers

Mike

Hi MIke!
Thanks for the post.

You'll find the info about the external diff pair to the compensation pins here: viewtopic.php?f=12&t=1135
The OP is for the FET example but there are BJT examples later in the thread.

My thought was to experiment with an optimized 100 Hz version and a 19/20 kHz resistively-summed pair and just switch the outputs.
The 19/20 kHz combo won't need a notch filter or ultra low THD since I will use it to look for IM at 1 kHz.