matched expo source / sink idea

[Haible Juergen]

Hi Terry,

thanks for starting the discussion.

	>A couple thoughts.  The current through Q4 is the base current of
Q1, 
	>typically a very small amount of current.  The Vbe of Q1 and Q4
will
	>therefore be very different, 

Yes. But is this really a problem ? 

	>Also, I don't
	>know how much compliance you need from this to drive another
circuit, but
	>you might find that voltage variations on the collector of Q3 (or
Q2) will
	>probably cause small changes in the collector current of Q2 (or
Q3).  

That's good to keep in mind ! 
When the source output drives a 3080, this would be no problem (at least if
there's no series resistor to the the Iabc input).
The sink should preferably drive a virtual summing node (opamp integrator
...)
with the reference not being GND, but a few volts above GND.
When driving a simple capacitor, a cascode should be included. (JFET
with base to GND, and D-S channel between the sink output and the cap.

	> The scale factor will be 36 millivolts/octave, twice the usual
value.

Glad I didn't label the divider resistors (;->) (;->)


More ideas:

Omit Q4, and put Q2 in the feedback loop of an opamp.

JH.