FETs on the ASM-1 VCO.

[Rene Schmitz]

At 19:21 14.12.98 -0700, you wrote:
>Magnus --
>
>We discussed this a few months ago. What you write is fine as far as it
>goes, but, as Terry Michaels pointed out, MOSFETS have a disadvantage in
>capacitance, and also in leakage. In addition, if you are using an opamp
>integrator, the rate at which the output can slew also matters.  It
>would be interesting to see some actual data, though -- details on
>spikes in the waveforms and on tuning accuracy and range.

I use MOSFETs as discharge switches, and have compared them to jFETs.
(though I'm still not sure if it acts more like a diode in my circuit, that
would lead to another interesting topic, why not use diodes as discharge
switches? A diode made from a jFET has pA leakage! This would work only for
the VCOs that have the cap to the supply.)
The MOSFETs have higher leakage, and there *are* effects from the capacitance.
I get a tiny wiggle at the reset, the waveform that I see at the output is
further 
distorted by the opamp buffer. When I use a FET here the wiggle disappears.
I use small signal FET (BS108) these don't have the low Rds(on) of the
power FETs 
(BUZ et al.)
These small signal devices have channel resistances equal in magnitude to
that of jFETs, 
so I would prefer the jFET (though I use MOSFETs... kind of schizophrenic :) 


>BTW, my updated version of Terry's VCO switches pretty fast and only
>needs a touch of HF compensation. Maybe it could be improved, but it's
>hard to see why you would need anything better -- unless you
>specifically need ultrasonic frequencies.

The fact here is that the reset pulse that is generated when the comparator
switches has the same length regardless of frequency. It has to be made
long enough to allow the cap to discharge fully. 
There were a lot of discussions about the ASMs VCO not oscillating, since
the cap (C3 18p or C4 in Ians schem) that determines the reset spike was
rather on the small side. 
I personally prefer a true schmitt-trigger where there are two tresholds, a
high one when the reset cycle begins, and when the cap is discharged to the
lower threshold the pulse ends, that way you have the fastest possible
reset spike, and if your trigger circuit is fast enough and of sufficiently
low inpedance you won't have trouble with the MOSFETs higher gate source
capacitance. Leakage is of course bigger than that of the jFET, but if you
don't need the *centi* Hz ranges this is no problem. The performance in the
ultrasonic range (20k+) is better with MOSFETS, but my ears are not that
good to hear any THD at 100Khz ;-)

Ok, besides this is sort of an apples and oranges thing, since my VCO is
completely different.
But I think the differences between MOS and j-FETs apply to all VCOs.

Bye
 René

             ,   : (uzs159 at uni-bonn.de)
|") [" |\ | ["   : (http://www.uni-bonn.de/~uzs159)
|"\ [_ | \| [_   : (http://members.xoom.com/Rene_Schmitz)