AW: AW: Lin/exp control

[Haible Juergen]

	>Yes, linearity is affected since the current gain of the
transistors
	>starts to fall at those higher currents, i.e. an increasing portion
	>of the control current is drawn through the base instead of the
	>collector. Maybe some sort of darlington configuration would work
	>better. Adds to the complexity, though...

The MS-20 circuit should avoid this, at least if the beta's of the two
transistors are matched.

	>Yes, I didn't notice any problem there, at least with the
particular
	>3046 array I used. 

Similar with the MS-20. If beta's are not matched, errors at the high end
occur. Beta seems to be more important than vbe matching here.
(which is a good thing, as my multimeter has beta test built in)

	>What I did notice was that the 3086 array doesn't
	>work in this application, matching does matter.

Aren't there self heating effects involved as well ? (With rather high
currents)

Your circuit should get rid of the beta problems with a FET instead of
a transistor, but the vibrato circuit probably wouldn't be that easy 
anymore.

JH.