Re(2): 2nd generation VCFs (was: Xpander VCF)

[Haible Juergen]

	>The patent
	>calls for n,p,n,p current mirrors. n,n,n,n also works but has about
	>1/2 the gain. 

How could that be ? Current mirror means reverting the direction, so
how could it possibly work with npn only ??

	> It's very much like it! I'm surprised they got a patent. I can't
argue that
	> it's not a improved version but it's very close to the original.

I like it better the longer I look at it.

In my previous mail, I had already typed a paragraph to say that I can't see
where the improovement in terms of CV feedthru rejection would be, but
then I saw my error and deleted that remark.
In the Moog ladder, there would be CV feedthru from different Betas
in one row (subtracting a tiny amount of current from each side of
the ladder, and with beta mismatch subtracting more from one side
than the other), and with the 3-transistor mirrors that effect is smaller !
(At least that's how I understood it.)

Which makes it a real improovement over the Moog cascade, not just
a workaround.

Back to the Moog ladder, if anybody wants it ultra low CV feedthru,
one should match the transistors (and all of them in pairs) for
Beta as well ! (And the top and bottom ones for Vbe, of course.)

But it's not really necessary for normal synth VCFs. Remember the
SSM2044 applications comprise the use of a VCF *after* the VCA,
and more general audio applications, where thumping would be more
critical.

JH.