SV: Re: [sdiy] ca3080-based tri-to-sine

[ryan williams]

replying to two of your messages,

Karl Ekdahl wrote:
> well, if i did understand all the math-talk in the
> paper i'd find the error, but it's way over my league
> right now...  what i gather from looking at the
> schematic is that the ca3080 forms a comparator where 
> i get
>
> Vin+ = +-0.01498V (Iin+ = +-68uA)
> Vin- = +-0.078V
> 
> and the outcome is
> (Vin+ - Vin-) * gm = Iout

for the CA3080, and LM13700 (without diodes). the output is 
Iout=19.21*Iabc*(Vin+ - Vin-)
which is like you said, but the equation I put above is not the whole 
truth. The gm is nonlinear and changes with Vin and temperature. this is 
just the small signal approximation of it. if you want to see why an 
overdrived OTA distorts the signal, then you can look at the nonlinear 
Iout equation:
Iout=Iabc*tanh((Vin+ - Vin-)/(2*Vt))
try plotting that, or actually, I did it for you. Take a look at this 
note that demonstrates the operation of an OTA based triangle to sine 
converter:
http://www.sdiy.org/destrukto/notes/tri-sin_converters_using_otas.pdf

note: I have never seen the circuit you were building.

> i can't see how this is going to shape anything into a
> sinewave though

the note above shows this.

> Well, thank you all for your help, but it seems my
> ca3080 is toast since it's output is basically static
> and not really any waveform whatsoever. That was the
> last ca3080 i had so now i will never build anything
> with them again, not that i'm dissapointed about this
> it's just that it seems kinda useless to try and learn
> more about synthdesign using nonavaliable parts.

normally, I would agree that it's useless to learn about nonavaliable 
parts, but in the case of the CA3080, I disagree. You may never design 
one of these into a product, but the OTA structure is very handy for 
synth designs. Many VCFs and VCAs use OTAs or circuits very similar to 
an OTA. If you can manage to properly analyze and derive the equations 
for an OTA it will deffinently help you with many other synth circuits. 
besides, as already mentioned, the LM13700 or cheaper replacement 
NJM13700 is essentially the same part (if diodes are not used). take a 
look at Achim Gratz' ota paper for a very complete analysis:
http://stromeko.synth.net/diy/OTA.pdf
I have some other, related links on my web site:
http://www.sdiy.org/destrukto/resources.html

-ryan williams