[sdiy] new schematic posted

[harrybissell]

One other thing... the time constant of the lag changes when you change the
shape...
the linear is MUCH faster than the log for the same attack/decay settings.
Maybe
the next mod is to take the times constant as well ???   :^P

still its simple and effective.

H^) harry

harrybissell wrote:

> Hi Ingo...  (inline)
>
> > Nice circuit. I must confess, it took me a while to find out why the
> > output of AR3 is tapped off *after* the diodes and not directly from
> > the opamp's output. This is to compensate for the diode voltage drop,
> > right?
>
> right.
>
> >
> > But doesn't this cause some interaction between the attack and the
> > decay pot if the shape pot is not in one of its end positions? As I
> > understand it, one one of the diodes conducts at a time. Say, the
> > attack diode conducts and the attack pot is set to some high
> > resistance. Isn't there some voltage fed into the integrator via R9 and
> > R10 and the decay pot? Unless the shape pot is in an end position, in
> > this case this path would be shorted to ground or virtual ground. But
> > if the shape pot is in center position, R9 and the two halves of this
> > pot form a voltage divider. So the decay pot would have an influence on
> > attack too.
>
> Hmmm. I have not noticed any interaction.  Both sides of the shape pot are
> at ground... The worst case would have to be with the pot at center
> rotation...
> each side would be 5K... in parallel that is 2.5K to ground.  The 20K /
> 2.5K
> divider would limit the effect of feeding back through one 20K... past the
> 2.5K to ground, through the other 20K... and through whatever remains of
> the decay pot.
>
> You could bring the resistor levels up (proportionally) and the effect
> would be still
> less... but it may be that they do interact a little. I'll try and look for
> it.
>
> > Another question, as I see it, in linear mode (wiper of shape pot at
> > ground) AR3 is overdriven (no local feedback). Can't it latch up?
>
> I suppose it could... but there is always the feedback from the integrator
> as well.  I've built a log of linear lag circuits with this technique...
> and have not seen
> a latch up.  If you were going to have one, it would probably be at some
> negative
> input voltage (imho).
>
> Choosing a different opamp could cure that if it was found to be a
> problem...
>
> Thanks for the comments !
>
> H^) harry