mic pre gain control for CFA (bloak, etc) using Linear pot

I was trying to think of a way to use an ordinary linear taper pot to control the gain of a current feedback amplifer for the application of mic preamp and came up with this:

edit: CONTENT REMOVED

so you are trying to actually change the gain of the opamp itself by modifying the feedback?

Yes it would work but you would also travel through a range of changing impedence and noise wouldn't you?

I think it would make more sense to stick with a set gain and have a voltage divider before it in order to change the audio level into the amp stage right? At least then you can tailor your impedences according to the amp stage a lot easier IMO..

But I could be wrong..

[quote author="Svart"]so you are trying to actually change the gain of the opamp itself by modifying the feedback?
[/quote]

? uh, yes. and this is quite common. usually the series leg is held constant (especially in a CFA circuit as it sets the bandwidth), and the shunt is a reverse-log variable pot in series with a big cap to ground (to keep DC gain at unity). did you try to analyze my circuit or are you down on it simply because it looks strange?

my objective was to use a linear taper pot for cost reduction reasons. this is not an all-out hifi atempt. but the performance ended up very reasonable. the bandwidth does change a little as the pot is swept, but it is better than you would think. very accectable and stable at all settings. the ergonimics are also surprisingly good. the gain curve "feels" smoother than most rev-log pots.

has anyone tried it?

mike

> it works as expected

Which is what I'd expect. You promise 15dB to 18dB range of gain. If it were exact log taper, mid-rotation would be -7.5dB or -9dB. A true linear taper would be -6dB at mid-rotation. It is fairly easy to skew the mid-point up or down a few dB.

If 15-18dB of range is adequate, this is a sweet solution.

For more general use, 20dB or even 40dB of reasonably equal-degree-per-dB action is desired, and that gets tough with linear pots.

It can be approximated usefully well with two amps, and the linear pot in between shifting the gains of two stages. Steve Dove published one implementation using linear for around 35dB range and no severely cramped ranges.

[quote author="PRR"]
If 15-18dB of range is adequate, this is a sweet solution.
[/quote]

actually there is a range switch that shifts the family of curves up making it just about 40dB total in the final version.

this was designed as a mod for a shure M67 mixer. I used one of the shure input transformers and PSU. there is a switch that I reclaimed right above the pot for the gain range switch (used to be low-cut filter). the whole circuit lives on a tiny PCB that is supported by the pot leads. damn, I should have taken a picture before I gave it back to the owner.

(like the one on top)


mike p

did you try to analyze my circuit or are you down on it simply because it looks strange?

Click to expand...

Well I didn't realize my comments sounded derogatory.

the bandwidth does change a little as the pot is swept, but it is better than you would think. very accectable and stable at all settings.

Click to expand...

That is good then, I only briefly looked at it and thought that it would change more than you have come to measure.

and this is quite common. usually the series leg is held constant (especially in a CFA circuit as it sets the bandwidth), and the shunt is a reverse-log variable pot in series with a big cap to ground (to keep DC gain at unity).

Click to expand...

I have seen a lot of trim circuits with the pot in the feedback, but not those working through a full gain range. I was always of the mind that this was because when you change impedences in the feedback circuitry you start creating other problems. again, I have not analyzed your circuit. as long as you have a handle on the situation then anything that works is great, just beware of the high impedence vs. low impedence you could see at a CFB opamp's inputs since the INV input is notoriously low impedence.


Big cap to ground for DC is a very common solution though.


Glad you are happy with it because that is what counts.

:thumb:

Would you mind sharing your M67 mod with me? I have one that I've wanted to turn into a dual preamp. Currently, it's a practically useless hiss generator, but I figure it has transformers, jacks and a ps. Many thanks, Keith Cary

[quote author="lohi"]Would you mind sharing your M67 mod with me? I have one that I've wanted to turn into a dual preamp. Currently, it's a practically useless hiss generator, but I figure it has transformers, jacks and a ps. Many thanks, Keith Cary[/quote]

This is different than what I posted before. Try an uncompensated 5534 for U1.


R4 is the gain control and should probably be a rotary switch with resistors. or 10k-20k reverse log pot would probably work ok. for the FET use 2sk170, not what I have shown.