[sdiy] Expo VCOs: some thoughts

[Tim Ressel]

Happy Sunday! And by Sunday I mean the sun is out today...

I've been thinking about volt/octave VCO performance. My customer wants 
a triangle core VCO with excellent tracking. And of course he wants it 
for $0.25, but that's another story. I started looking at expo VCO 
design from the standpoint of tracking. Typically there are two pieces 
to an expo VCO: the expo converter and the VCO core, and they can be 
looked at separately. I'm going to tackle the core side because that is 
where the problems are with triangle cores.

I looked at two designs: LM13700 and 4046. The focus is on linearity 
between the control current and the resulting frequency. There are other 
issues, but I feel linearity is the biggest. In an ideal core the 
linearity would be perfect, but as I looked at the datasheets it became 
clear these were not ideal. Both the 4046 and the 13700 are not so great 
for linearity. Don't get me wrong; they are okay devices. It just that 
our performance needs are beyond their capabilities. It's not their fault.

So now I am looking at a switched current mirror type design similar to 
that found in the venerable 566. I believe the non-linearities will be 
reduced to opamp, capacitor, and comparator issues. That's all, just 
that ;-)

My customer also needs a small footprint as we are trying to cram 4 of 
these on one board. As a great boss once told me: "Size, price, 
performance: Pick any two."


-- 
--Tim Ressel
Circuit Abbey
timr at circuitabbey.com