SV: [sdiy] How about some SDIY?

[Aaron Bader]

On Thu, 6 Oct 2005 15:03:31 -0700 (PDT)
Harry Bissell Jr <harrybissell at prodigy.net> wrote:

> I tried a quick spice simulation of Aaron's
> idea and it 'sort of' works. For the most part
> it generates a triangle wave... and it becomes
> sawtooth only at extremes of CV setting. The change in
> frequency is really extreme, and it cannot get a very
> fast rising
> or falling ramp.

That's exactly what happened on breadboard.

I've already got some ideas on how to fix this, some more drastic of a
redesign than others. Basically, I've realized that in order for the
frequency not to interact with the duty cycle, the positive and negative
charging currents will have to vary in a nonlinear manner.

Specifically, seems like it is going to be two exponential curves will
be required in inverse of each other. (seems like synth circuits are all
just exponents and voltage-controlled gain cells.)

My next guess is an exponential amplifier that can generate postive and
negative voltages in between the schmitt trigger and the OTA, with the
offset voltage fed into that. Also gets around the fact that OTA's don't
like huge differential voltages.

Limitations I can see already: First, the sweep voltage must be small.
Specifically, no larger than the size of the square wave output by the
schmitt trigger. This is no big deal if you drop in an opamp summing
stage with a gain less than unity.

Second, you are right in that it will be pretty difficult to get a
'true' sawtooth with a lighting fast reset time. I consider this to be a
minor limitation as sawtooth VCO's are not going away any time soon even
if I do pull this off, and it'll still make a good 'sweep' sound. Some
pretty anal calibration and nulling of offsets would help as well.


And lastly, one thing happening on breadboard is that towards the upper
frequency range of this OTA oscillator design in general is that the
amplitude starts to become frequency dependant. I'm not sure exactly how
badly as my 'scope isn't calibrated and I have never found a manual to
do so (Vu-Data 27-something, not in front of me at the moment... I
should fix that Tek T932 with broken knobs I have.)

I think it's caused by finite slew rate of the comparator, meaning the
wave ramps up or down faster and allowing it to overshoot a bit more
while the comparator (TL072 in this case) is slewing. Urgh, that's going
to effect the linearity as well. Doesn't seem to be a huge problem until
the high end of the frequency range, but still something that would have
to be delt with if accuracy is desired.

-- 

"Not a bird not a leaf not a sound/And well after the close of time" 
	--Front 242

Public Key: http://sparked.zadzmo.org/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20051006/d0869470/attachment.sig>