LM13600

[Haible Juergen]

	>The LM13600 has lower buffer (darlington) bias current at low Gm,
and
	>higher current at higher Gm... at the possible expense of feeding
some
	>"thump" into the darlington as the Gm bias is varied (like VCA or
VCF with
	>fast attack...)
	>The LM13700 has fixed bias for the buffer, alas... at too high of a
value
	>to be really effective witlow Gm inputs...

Do you mean they do load the first transistor of the darlington with a fixed
current sink ? There is nothing like that shown in the data sheet, but it
*does* make sense, and might be the "missing link" to explain why the
13700 is bad for VCF's that use the buffers.

	>Tom G has a VCO which doesn't work with the LM13700... but says the
LM13700
	>is OK for the MS-20 VCF design...

And I say the 13700's I've tried do *not* work in the (unchanged) Korg MS-20
VCF design.
I have tried, and the 13600's are good, but 13700's no good. The filter
still works, granted,
but you get excessive thumping with the 13700.

I never really understood *why* this is so, however. A, while finite,
constantly high Z
buffer ('700) should cause less trouble than a variable Z buffer ('600),
given that
the max Z of the '600 is not higher than the constant Z of the '700.
But Harry's suggestion could be an explanation. Then the constant Z of the
700
might be lower than the maximum Z of the 600.

JH.