[sdiy] VCO tuning philosophy re-visited

[Ian Fritz]

At 04:32 PM 2/7/2011, Tom Wiltshire wrote:
>I can see the attraction of a simple counter run from the LM311 pulses as 
>the base of a wavetable oscillator. But the LM311 has a linear response, 
>so you're going to need a exponential convertor in front of it, in which 
>case, you're back to the usual problems of VCO tuning.
>
>So what's the advantage of the LM311? Why not run the counter from the 
>comparator reset pulses of a standard ramp-core VCO, but tuned higher than 
>typically? Is *that* the advantage of the LM311 - it can cope with higher 
>frequencies?

As I explained in a previous post today, the LM331 -- that's three three 
one -- has no reset error like other cores do.  Yes, you still have to have 
an expo converter and to compensate and correct it.  But at ultrasonic 
frequencies the reset error of a standard core is quite large.  The 
LM331  tracks to within 0.1% up to 120 kHz or so.

Ian