super nice transconductance amplifiers

[CGS2510.SEP.STUDENT.UCA at CUB.UCA.EDU]

> > Since when was $6.20 low cost.  You can make a very fine ring mod 
> > from two LM13700's and a couple dual op-amps and
> > $.73 + 2X $.35 + resistors < $6.20
> 
> Really? Will 13600's work as well? 
> 
> Do you, by any chance, have a schematic for that?


The 13600's and the 13700's are identicle with the slight exception 
that the 13600's use a mirror of the gain control bias current for 
the drive of the output buffers.  I never use the buffers and fell 
happier using the 13700's since it would seem that the additional 
junctions add a bit more noise to the bias mirror.   I guess it is 
conceivable that I could scan a drawing and send it, but I think I 
can explain well enough.

Take two 13600's and tie the outputs together and run that into the 
inverting input of an op amp that has a resistive feedback.  One 
signal source goes to the inverting input of one transconductance 
amplifier and the non inverting input of the other.  From here, 
construct a circuit that depending on the polarity of the second 
signal source routes the magnitude (as a current) into the bias 
current pins of the 13600.  This can be accomplished by using two 
dual op amp current sources.  They automatically do the current 
selection since the bias current is refered to the negative power 
supply rail on the 13600.  Use two, one plain, the other inverted, 
one to each amplifier and it should work great.  I had troubles 
getting better than 75 or 80 dB SNR out of this setup and it temds to 
have severe harmonic distortion for low bias currents.  You must play 
with the input gains of the signal sources to get a desirable result.