Show me this same circuit in the literature and I’ll renounce my claim on it; it’s a real simple concept with Birt origins, and when it happened I was real surprised i hadn’t seen it before. Happy to share this and I hope people use it, please attribute if it is indeed novel and goes anywhere.
Should be a healthy bump below -115dB with sub-3nV/rtHz amps, but not too shabby with the usual suspects. Certainly great in through-hole with LM4562s. Beats the unity gain Baxandall, as the diff amp has only a percentage noise authority at full clockwise, and falls rapidly from there.
The fader driver level falls as the fader lowers, i.e. U1A is -6dB when U2A is -12dB at center in my example.
Still the ol’ s-curve here, but as you know it’s pretty consistent across the usable range until you start “slugging” to -14 or -15dB at center. I like -12 because of the round numbers at quarter rotations, and with a 21-detent pot it’s basically 1.2dB/click until you’re down past 10 o’clock. Can’t do that with a slugged 1K or 2K pot unless you’re getting into parallel amps or are particularly keen on the BUF634A. Can’t easily channel-match for stereo with a slug (or the Bax for that matter), either. But now it’s easy!
Need for servo is questionable with modern ICs. I’m curious how many microvolts is too many microvolts on one end of a fader whose wiper is AC coupled.
I’m also curious what the equivalent impedance is at each side of R1+R2 as the fader lowers. I suppose one could calc current and extrapolate from there.
Have at it, LMK any thoughts.
[Edit: Below is a version of the origin story, the center attenuation level is not as easily trimmable, but with good R1/R1 and R2/R2 matching the only wild card across two channels should be the accuracy of the dual gang wiper positions. TT P260 or Bourns 51 should do it, no need for a Blue Velvet or TKD or whatever spendy stereo attenuator hifi people like now. /end rant …U3A has greatest noise contribution at CCW, still less than U2 through mid-travel, and is simply a bias source at CW.]
Should be a healthy bump below -115dB with sub-3nV/rtHz amps, but not too shabby with the usual suspects. Certainly great in through-hole with LM4562s. Beats the unity gain Baxandall, as the diff amp has only a percentage noise authority at full clockwise, and falls rapidly from there.
The fader driver level falls as the fader lowers, i.e. U1A is -6dB when U2A is -12dB at center in my example.
Still the ol’ s-curve here, but as you know it’s pretty consistent across the usable range until you start “slugging” to -14 or -15dB at center. I like -12 because of the round numbers at quarter rotations, and with a 21-detent pot it’s basically 1.2dB/click until you’re down past 10 o’clock. Can’t do that with a slugged 1K or 2K pot unless you’re getting into parallel amps or are particularly keen on the BUF634A. Can’t easily channel-match for stereo with a slug (or the Bax for that matter), either. But now it’s easy!
Need for servo is questionable with modern ICs. I’m curious how many microvolts is too many microvolts on one end of a fader whose wiper is AC coupled.
I’m also curious what the equivalent impedance is at each side of R1+R2 as the fader lowers. I suppose one could calc current and extrapolate from there.
Have at it, LMK any thoughts.
[Edit: Below is a version of the origin story, the center attenuation level is not as easily trimmable, but with good R1/R1 and R2/R2 matching the only wild card across two channels should be the accuracy of the dual gang wiper positions. TT P260 or Bourns 51 should do it, no need for a Blue Velvet or TKD or whatever spendy stereo attenuator hifi people like now. /end rant …U3A has greatest noise contribution at CCW, still less than U2 through mid-travel, and is simply a bias source at CW.]
Last edited:
…it can be three, though, depending…and just two more than a regular fader buffer.
