spikes (was:AW: FETs on the ASM-1 VCO.)

[terry michaels]

Message text written by Haible Juergen
>With all this talk about MOSFETs, JFETs and even diodes:

Is there anything wrong doing it with BJTs ?
I've used bipolar transistors for discharging in my VCOs all 
the time. I haven't done much research myself, just copied what
I saw in many commercial circuits. It works ok, though the spike
in the (converted) triangle could be smaller ...
Is this a general problem with BJTs, I mean do I get much shorter
reset times with FETs ?

IMO, the only reason for ultra short reset time is triangle
conversion. And the problem is worst in LFO applications.
Sure, I know, first rule in Building VCOs: Start with triangle,
not with saw. But then I absolutely love the variable shape
of Osc 3 of the VCS3 for modulation - it's a very unique
effect - and you can only get that when your oscillator core 
is producing a saw wave. So here is an application where
short discharge times are important. (I don't speak of
nanoseconds here, but not of 1nF capacitors either. It's
a 1u electrolytic, if memory serves.) I could improove things
a little when I used a darlington.

JH.

<

Hi Juergen

I admit I've never tried a bipolar transistor for VCO reset, but I think
its a less than optimum choice.  The reset point cannot go to zero volts,
because of the saturation voltage of the BJT.  The offset voltage so
created might be temperature dependent.  The saturation voltage will vary
with collector current, which changes as you change frequency.  Charge
storage might cause problems in trying to turning the BJT off quickly. The
beta of the transistor, and so the discharge current, approaches zero as
the collector to emitter voltage approches zero.

I think a darlington would be a bad choice, as you then have a saturation
voltage of at least one diode drop, and it definately varies with
temperature.  Getting a darlington to turn off quickly at the end of the
reset can be tricky, they are much slower than a single BJT.

TM