[sdiy] They aren't sawtooths, they're ramps

[Magnus Danielson]

David G. Dixon wrote:
> Boy, am I ever stupid!  I designed, simulated, laid out and built a
> Triangle-Core VCO based on the 2164 Expo, and then it didn't work.  Why not?
> Because I forgot (for the second time!) that the 2164 must feed its current
> to virtual ground.  I wired it directly to the Iabc of LM13700, which rides
> at about -13.5V!
> 
> At least I've got a nice front panel, and I can reuse all the board parts
> for a Saw-Core, so I'm really only out a bit of board material and a few
> hours of wasted time, but still (urrggh!).
> 
> This all stems from the fact that there is no SPICE model for 2164 (Sam
> Hoshuyama not withstanding), but it is easy enough to put the basic I-V
> relationship for the VCA into SPICE.  This lulled me into a false sense of
> security.
> 
> I'm completely fed up with this now, so work on the Triangle-Core 2164 Expo
> VCO is officially cancelled.  The Saw-Core version will have to be good
> enough (it's actually very nice).  I'm going to go to the garage and etch a
> new Saw-Core board right now, and try to forget all about this whole sorry
> episode!

Why?

Shouln't you be able to current-mirror with a op-amp servo to bias the 
PNP current mirror to sit as a virtual ground on the 2164 output and 
feed the LM13700 with that, and then just swap the direction of the 
output current of the 2164?

Doesn't it ticle you to figure out how to circumvent that designflaw to 
make it work?

Cheers,
Magnus