[sdiy] TB303 Slide

[Tim Stinchcombe]

> >> Does the R2R DAC in the TB 303 feeding the glide capacitor 
> directly 
> >> actually achieve constant glide time regardless of the 
> voltage span 
> >> between notes. (ie: glide time from 1V to 1.83V is the 
> same as from 
> >> 1V to 2V?)
> >
> > Yes. The output resistance of the DAC is always 100k, regardless of 
> > the voltage, and so the time constant with the 220n (C35) is 22ms - 
> > thus after about 60ms the slide is complete, no matter how large or 
> > small the voltage change was.
> 
> I guess it's "constant" glide time?
> 
> What I mean is... that it's an ordinary exponential RC 
> charging curve, having longer times for bigger steps, never 
> completing the slide? :-)
> 
> A glide from 1 V to 2 V is shorter than a glide from 0 V to 2 
> V. When gliding from 0 to 2, it goes quicker at the start and 
> by the time it passes 1 V the rest of the glide is identical 
> to a glide from 1 to 2. So the time is indeed longer... The 
> glide time is just "constant" in the sense of reaching within 
> a certain percentage (say, 90%) of the total glide distance - 
> but no such percentage can be considered a
> *complete* slide for all different distances. :-)

Agreed, each person can nit-pick what they regard as 'complete' by picking
their own favourite no. of time constants to get as close to the final
voltage as they like - however, within what I believe to be the normally
accepted convention for such a circuit, 'constant time glide' seems a
useful distinguishing description against alternatives such as 'constant
rate glide', which has markedly different characteristics.

However, I will take issue with part of this:

> A glide from 1 V to 2 V is shorter than a glide from 0 V to 2 
> V.

Mmm, well I'm not buying this argument:

> When gliding from 0 to 2, it goes quicker at the start

yes, ...but...

> and 
> by the time it passes 1 V the rest of the glide is identical 
> to a glide from 1 to 2.

...no, I think not! The 1 -> 2V curve is *always* above the 0 -> 2V curve,
with both converging toward 2V as the time gets large. Without bothering to
be rigorous about it, it looks to me that at any given time the gradient of
the 1-2 curve will always be less than that of the 0-2 curve, and so nowhere
can the 0-2 curve be 'dropped' onto the 1-2 curve!? Because the time
constant *is* the same in both cases, when the 0-2 curve passes 1V, it has
much less time to make up the remaining 1V, and so must do so at a faster
rate than the 1-2 curve! 

Tim
__________________________________________________________
Tim Stinchcombe 

Cheltenham, Glos, UK
email: tim102 at timstinchcombe.co.uk
www.timstinchcombe.co.uk