Modular Synth Panels

On 2/19/08, James Elliott put forth:
>I have a couple questions regarding the construction of the CGS65
>(Tube VCA/Timbrel Gate) module as it relates to the Bridechamber
>panel.

Since you mentioned Bridechamber, and a bias switch, I'm going to
assume this is MOTM format.  Adding a bias switch to half of
Richard's panel was my request, so Scott added it, because that's the
kind of happening guy he is :)

Now, ironically, I haven't built my cgs65 yet. To save energy, I have
the heat turned off in the room off the garage where I do my
soldering.  Not because I would want to prevent global warming, it's
20 degrees this morning, but to avoid a heating bill the size of a
CS-80.  I also have been spending the little "gear time" I've had
doing other things.

Anyway, the first thing I would do is get some pot chiclets.  This
will make the front panel wiring much easier.

http://www.wavemakers-synth.com/motm/chiclets.html

>1) Do I wire Input 2 to the "Audio SN" on the PCB? If not where does
>Input2 go?

I have no idea what "Audio SN" means, but I'm not planning on using it.

Each pot acts as a voltage divider.  I always forget which pin is
which.  You can use a an ohmmeter to determine that the wiper is at ground
when the knob is fully CCW.

Wire ground to the correct end of the two input pots, wire the input
1 jack to other end of the input 1 pot, wire the output jack to switched
lug on the input 2 jack, wire the input 2 jack to other end of the
input 2 pot, wire the wiper from each input pot to the audio inputs
on the PCB so that each signal goes through its own 100K resistor of
the summing amp.

Wiring the switched jack on input 2 allows you to feed the signal
back into the tube when a second input is not being used.

>2) For the bias switch, am I supposed to use a DPDT to create a
>circuit between the two optional pairs on the board?

I don't think you are "supposed" to do anything.  How you want to
bias the grid is up to you.  This is what it says on the CGS site:

"On the first (yellow) run of PCBs, the grid bias resistor (330k)
goes to the negative rail. In this position, it causes some
distortion of the signal. This can be a good thing. I prefer the
resistor in this position when using the unit as a timbral gate. This
is shown in BLUE on the schematic. An alternative is to bias the grid
to around 1.2V with respect to the negative rail - in other words, to
the other end of the heater/cathode."

If you were to do that, you would use a SPDT switch -- wire the
actuator to the 330K resistor going to the grid, and one end of the
switch to the ground side of the cathode, and the other end to the
negative rail.  A center-off switch would add another setting.

You might want to look at a few websites that discuss grid biasing.
What makes the cgs65 somewhat unique, is that the cathode is
connected to a negative voltage (rather than using a resistor and cap
to make the cathode positive in relation to the grid).  Generally,
the grid is biased negative in relation to the cathode for at least
two reasons, to prevent excessive grid current, and to offset the
input to reduce distortion.

If the input signal is too negative there is plate-current cutoff,
if it is too positive there is plate-current saturation, resulting in
distortion.  Biasing the grid between these two allows for the most
gain and the least distortion.

I'm thinking since the input cap is connected to an op-amp
(something you generally don't see in tube circuits), that the 330K
resistor and input cap aren't going to behave like a typical shunt
grid-leak circuit.  If grid current charges the cap, the op-amp will
swing around to compensate.  I'm also thinking since the cathode is
negative, the input wouldn't need as much of an offset to produce a
symmetrical output.  So I'm going to try to a few different things
and see what happens.

Anyway, asymmetrical distortion is one of the things that make tubes
sound like tubes.  So by adjusting the grid bias, the characteristic
of that distortion can be adjusted.

My plan is to omit the potentially noisy diodes, and use a current
limiting resistor instead.  Then by running a sine wave through the
VCA, I'm going to figure out what bias voltage results in the least
distortion, then figure out another setting that produces more
asymmetrical tube distortion.

Also, others have reported a high-pass effect, so I'm going to try
increasing the input and output caps, and a cathode capacitor, and
see what happens.