[sdiy] Help needed with ASM based VCO

[René Schmitz]

Hi Ian, Chris and all,

Ian Fritz wrote:

> Careful there.  You don't get something for nothing.  The J108 has much 
> higher leakage current -- a factor of 30 compared to the 2N4391.  (And 
> the sheet I have doesn't even specify the capacitance.)  So you are 
> trading off some low-end performance.  Maybe you don't care, which is 
> fine, but the original FETs (2N4859 or 2N4391) were carefully chosen to 
> give the best overall performance.

I know I keep repeating this, but anyway: There is no point in makeing 
the discharge faster than the OP can handle. This only leads to 
overshooting behaviour. There is an *optimal* Ron (actually Idss, but 
they are related to each other) for the given opamp slewrate.

(Its actually a mixture of both Ron and Idss. It depends on the 
threshold voltages, if you always keep some voltage across the cap its 
Idss (if you never enter the ohmic region), if you discharge all the way 
to GND, you enter the ohmic region at some point, with a much slower 
discharge(!).)

> But an even better solution seems to be MOSFET switching, such as in the 
> designs I put up a couple of monthe ago.
> http://home.earthlink.net/~ijfritz/xfer.htm

I've been playing with MOSFETs for discharge for quite some time, the 
earliest examples are even on my site (VCO1 and VCO2). But I never quite 
liked the behaviour at the transition. I mean those dirty oscillations 
around the discharge, like you show also on the scope shots at your 
site. I suppose this is really due to charge injection. With a careful 
balance of Idss and cap size to the slewrate of the opamp, you can get a 
much cleaner transition with a JFET. Which is beneficial IMO for 
waveshaping the sawtooth into a triangle.

Cheers,
  René


-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159