Lightspeed Attenuator a new passive preamp

[Nelson Pass]

Here the graphic for NSL32SR2S:

[Nelson Pass]

And here's the closeup:

[Nelson Pass]

Here's a distortion curve for the NSL32SR3:

[Nelson Pass]

And here's a distortion curve for the NSL32SR2S.

These curves were created with a 600 ohm source impedance,
and the current through the diode was set to produce attenuation
of -0.1 dB, -0.5 dB, -1 dB, -2 dB, -3 dB, -6 dB, and -10 dB
which corresponds to

50 Kohm, 10k, 5k, 2.3k, 1.5k, 600, and 277 ohms.

[Nelson Pass]

And last post of the evening, here's a scope shot of the waveform
and distortion, showing the 3rd harmonic content.

This shows that the element's resistance decreases in response
to waveform voltage. Used as a shunt to ground, this is a
compressive characteristic. Used in series, it has an expansive
characteristic.

[Panelhead]

I built my first unit from R3S units. It worked fine, even tried it for a while with linear pots to take advantage of the non-linear response. This was entirely because I ordered the wrong parts from Allied.
The next couple were all built using R2. And then swapped to log pots for voltage control. The only difference I could find was more better control at high attenuation.
The R3 grade, and linear pots ramped up quickly. Normal listening was a little hair trigger. Highest attenuation was not enough sometimes with Lowthers. Swapping to more normal efficiency speakers helped with this a bunch.
Sonically, the R2 and R3 sounded identical to me. There appears to be some difference is distortion, but my speakers most likely swap this.

George

[Nelson Pass]

I chose the NSL32SR3 for its better matching, linearity, and range,
and created attached circuit.

The values of R1 and R2 were chosen to make the summation of
the resistors of LDR1 and LDR2 equal to 10K ohm, which then sets
the minimum input impedance of the circuit with the rail voltages
you have (in this case the commonly found +/-15V).

Then you attach the potentiometer with the values shown with the
bottom end going to the -15V rail and the top to a variable regulator.
The nominal value of this regulator is 7.5V, but it needs to be
adjustable so that you get the range you want.

The range that this was adjusted for was -1 dB at the top and
-40 dB at the bottom. More range is possible, but at the expense
of the taper characteristic.

The buffer section consists of matched 2SK170 and 2SJ74, matched
for Idss. If you do not have matched parts and DC offset is an
issue, then consider using an output coupling cap.

The bandwidth of the active buffer is extremely wide, and for that
reason 1K ohm input and output resistors are appropriate.

The distortion of the active buffer is about .003% on a clear day
at 1 V, which is less than you will be experiencing with the LDR
portion of the circuit.

Of course you can tart this up as you like - I consider this the
minimal circuit.

[Nelson Pass]

Here is the attenuation curve I got, showing dB of attenuation
versus the clock position of the volume control.

[Nelson Pass]

Lastly, here are the distortion figures I got with a constant 1 volt
input to the attenuator. More input voltage naturally results in
greater distortion, except at the -6 dB value, where the 3rd harmonic
distortion of the two LDRs cancels. From -1 to -6 dB, the distortion
is compressive, and from -6 to -40 dB the distortion is expansive.

Any questions?

[Nelson Pass]

I forgot to mention that because I have not followed the entire
thread, it is possible that I have duplicated something without
knowing it. If that's the case then

Excuuuuuuuse Meeeeee!