[sdiy] Expo-Con

[Ian Fritz]

At 07:07 AM 2/3/2015, Rutger Vlek wrote:
>The last difficulty I had to overcome was how to choose an appropriate 
>reference current. From René's site I got all the equations, and Roman 
>Sowa helped me understand the importance of balancing delta-Vbe (the 
>expo's voltage input basically...) around 0, such that the transistors are 
>always working in a range they're most stable in. However, I manually 
>derived the equation that gave me the actual reference current for my VCF, 
>based on the idea that I wanted optimal stability half-way the exponential 
>control scale.

I take a slightly different view of this question.  I say accuracy is more 
important at high frequencies  than at low frequencies.  This is because, 
relatively speaking, inaccuracies at high frequencies produce faster beat 
rates relative to a perfect oscillator.  IOW, the frequency in Hz matters, 
not just the octave-to-octave tracking.  For this reason I set the zero of 
delta Vbe to correspond to something like 2 kHz.  Higher than this is not 
often used for fundamental frequency, and at low frequencies beats become 
too slow to be objectionable. I know this isn't the herd consensus, but it 
has worked well for me.

For VCAs stability is not a big concern, since we are not very sensitive to 
small amplitude changes.  Here S/N is a more important consideration, so I 
say hit your converter with as much maximum current as you feel comfortable 
with.  So for and OTA (CA3080, LM13700) maybe something like 1 mA max 
current, for either log or linear response.  For a discrete design maybe 
more depending on the transistor.

Ian