Climbing the Moog ladder

[Martin Czech]

Snip!
> Now, I think my CV problem could be avoided by using matched transistor
> pairs up the ladder. At the moment I'm using CA3046 for the top and
> bottom pairs. But I might try matching the others as well. There is a
> 5mV differential at the top of the chain when modulating the filters
> cut-off from 100Hz to about 20KHz. This can only be down to mis-matches
> in the chain. Oddly enough it is worse at low Q settings.
> 
> So who makes the cheapest matched transistor? 
> 
You do ;->!

No, really, MAYBE it is possible to achieve quite good matching if you
match trannys out of the box by measuring Ube @ some Ic.  The moog
manual explains how to do this. Some people recommend measurements at 2
points i.e. 2 Ic levels (maximum und minimum bias).  For this purpose I
got a box of 5000 BC549C.  I'm thinking about a "matchbox" where 12 pnp
or npn (sockets) can be maesured with a 12-1 switch with constant
current. I hope i'll be able to maesure some dozent in 15 minutes or
so, after thermal equilibrium.  (But thermal matching will be difficult
in this discrete case. Maybe a plastic box arround the trannys will prevent
large temp gradients and slow down heat flow.)
But not now, if the weather is like this, I refuse to stay inside the
house. If the winter comes ....

The CA3046 is quite "cheap" and has only ONE matched pair, more or less.
I think you won't find something cheaper. Next is LM394 (right number ?).

ARP uses additional transistors in the ladder (all 3046). I don't
know, maybe the are for supression of such dc shift.

By the way: it could be interesting to add a highpass RC-Filter to the feedback 
loop. This way, it should be possible to avoid the level drop at low 
frequencies at high q settings.

m.c.