Anyone installed a variable NFB control on an ST70?

[Brinkman]

I asked a similar question on Dynaco Doctor but feel like I need more input...

I'm using the VTA driver board and no longer require the stock 10K pots on the ST70 for biasing purposes. In their place, I would like to wire up some high-quality pots (or stepped attenuators) for use in my negative feedback network.

My thinking is as follows, and needs double-checking:
Each channel would have a pot in series with a resistor of the recommended lowest resistance (in other words, the maximum NFB before the amp became unstable). Each pot would be of such a value that in the middle position each NFB channel would have a series resistance of 3.8Kohm (my fixed resistance now, running from the 4-ohm tap). For example, a 2.7Kohm resistor in series with a 2Kohm pot. This would yield a variable range of +/-1Kohm.

I've seen amps with variable feedback controls so you can tune NFB to suit your listening tastes, but have little idea how complicated a procedure it would be to put similar controls in an ST70. I could use some feedback, if you pardon the pun.

Best,
Brinkman

[Shannon Parks]

Why not just use a 100k linear pot and run the output connect to pin one, and run a wire from pins two and three to the negative feedback connection on the board?

[Brinkman]

Why not just use a 100k linear pot and run the output connect to pin one, and run a wire from pins two and three to the negative feedback connection on the board?

I was under the impression 100Kohms of resistance would be too drastic a sweep; 100Kohms resistance on one end almost none on the other... then again I'm still learning the basics, so please correct me.

I saw a chart in an old audio engineering text I have (Engineering Electronics, by John D. Ryder. Pg. 187) that compares the frequency response of an RC-coupled amplifier as a function of NFB, in %. There seemed to be a Goldilocks & the Three Bears-style relationship between NFB & freq. resp., so I wanted a range that would yield very little to no feedback on one extreme and say, 50% on the other. I suppose 100Kohms may be the correct range for my driver... I see what Roy thinks.

Best,
Benjamin

[mesherm]

The only thing close to variable NFB I have done was to use the front switch position normally used for Stereo/Mono operation to switch between the standard VTA driver NFB resistor (after correcting for 8 ohm tap) and a resistor approximately 40% higher. I did that because I was running the output tubes on my ST70 clone permanently in "triode" mode and wanted the option of normal or decreased NFB. Naturally there is more gain with the decreased NFB but I usually end up keeping it in the "normal" position.
I would do some testing using a pot to vary the NFB to find the operational limits and then wire in 4 position switch in one pot location and put a dual variable pot in the other location so you could have your choice of either variable NFB with one switch position or 3 fixed NFB values in the other switch positions.

[Shannon Parks]

Benjamin, I would consider Roy's value the minimum value resistor for feedback, so I would leave that in place. The 100k pot would just add 0 to 100k of resistance to that, to allow you to lower the feedback and raise the gain. With this method there won't be any stability problems.

[Quad]

Mike, for the VTA board, what is an appropriate resistor value
if using the 8 ohm tap for NFB rather than 16 ohm? Any change
to the 220pf cap (C11/12)?

[mesherm]

The "normal" method is to multiply the resistor value by 0.707 and the cap value by 1.414.