AW: LM13700 VCF (update)

[Haible Juergen]

	>
	>Well I've built a two pole VCF using the LM13700 and I've found
a
	>disadvantage with the internal buffers provided on the chip.
With very
	>small values of controlling current, the darlington buffers do
not
	>follow very well, the output voltage is a lot lower than
expected. Far
	>more than the 1.2volts difference in input to output. In fact
it is
	>possible for the buffer outputs to drop to the -ve rail. The
OTA does
	>not produce sufficient output current to drive the base at low
levels
	>of control current. The effect of this is that the control
range of the
	>filter is reduced, in particular the bottom end. The VCF on the
Fatman
	>will suffer from this, but it is not probably too bad if you
don't want
	>your filter to cover the whole of the audio band. 
	>
	>So call up the TL072, use this as a voltage follower instead of
the
	>darlingtons and you are away. Loads of range.

Very true - I've learned about this the hard way, too (lots of
scratching and component mounting on the copper side of
otherwise well-layed-out pcb's ).

Opamp buffers have their disadvantages as well, btw.:
They are slow, i.e. they add unnecessary delays into your
filter circuits, which will affect the frequency tracking of
resonance for example.

A better way to use the 13700 is adding a jfef stage between
the ota output and the buffer inputs. Usually I take a cheap
BF245, gate to ota output, drain to +15V, source to buffer
input, 33k resistor from source to -15V. This works excellent
in filter or envelope applications, is faster than opamps, easier
to layout, and much easier to retrofit in existing circuits.

BTW, speaking of delays from opamps: guess why Tom Oberheim
used fets + opamps as buffer stages in his famous 2pole filter,
but only included the fets in the filter loop !!

	JH.