[sdiy] Help, I'm Desperate! (Charge Injection with DG408)

David G Dixon dixon at mail.ubc.ca
Fri Dec 14 21:09:41 CET 2018

I never have problems with analog electronics.  In the case of this scanner,
for example, I designed the analog part so that there are plateaus in the
control voltage when switching occurs, which gives me about 200mV of slop in
the CV for switching.  This circuit worked the very first time (even though
I've rebuilt it three times).  The problem here was all digital.
Later today, I'm ordering some DG444 and DG445, and also some single SPST
switches (don't know the DG number yet) and I'm going to rebuild this thing
Roman's way (with a few minor modifiations).  The analog part is perfect and
won't have to change at all.
On the digital side, I'm worried about the fact that the LM3914 comparators
do not have built-in hysteresis, and there is no way to install it, so I'm
worried that the comparators will oscillate.  I'm going to stick with LM339,
fully rigged for hysteresis (about 10mV worth) as this worked well.  Does
anyone here have any experience with oscillating LM3914?


From: Rutger Vlek [mailto:rutgervlek at gmail.com] 
Sent: Friday, December 14, 2018 5:36 AM
To: David G Dixon; SDIY List
Subject: Re: [sdiy] Help, I'm Desperate! (Charge Injection with DG408)

| It just shouldn't be this hard.   

Ha, I totally agree! I once explored an approach to ring-modulation using a
single VCA and an inverting switch. It seemed like a nice, cost-effective,
approach (especially since I wanted to use a discrete VCA that would be too
bulky for another approach). After two prototypes of this circuit, I simply
gave up. In the analog domain it was simply too difficult to trigger the
switch exactly at the point where the VCA was fully closed, due to circuit
accuracy and noise. It was a hard lesson in the reality of analog
electronics. I went down the route of simply duplicating the discrete
circuit for the negative quadrants, and it works extremely well albeit not
super-linear, but that was a deliberate choice.



Op wo 12 dec. 2018 om 03:56 schreef David G Dixon <dixon at mail.ubc.ca>:

It just shouldn't be this hard.


From: Synth-diy [mailto:synth-diy-bounces at synth-diy.org] On Behalf Of Roman
Sent: Tuesday, December 11, 2018 1:20 PM
To: synth-diy at synth-diy.org
Subject: Re: [sdiy] Help, I'm Desperate! (Charge Injection with DG408)

I know it's too late, as the circuit at this stage has already gone slightly
different path, and I already gave up, but couldn't help drawing this little
schemo that explains my idea of using single switches with that obsoleted
chip. Switches should be DG444 or similar (0=closed, 1=open), or replace
diodes with NAND gates and use any regular dirt cheap switch like 4066.

There's no switching happening while VCA is open, provided that VCA control
triangles match the edges of LM3914 stages. The only switched channels are
the ones routed to muted VCA, so in theory there should be no clicking

Basicaly this is just the circuit that makes this input switching sequence:



by creating one step overlap between stages. Only 6 inputs in this example,
but can be extended to any other number.


IMHO switching active channel will always produce some clicking for many
reasons: not matched channels, difference between transistion times H->L and
L->H as Ingo said, break-before-make feature of DG407, slow response of
4532, and who knows if not also charge injection.


Dnia 11 grudnia 2018 18:24 Ingo Debus <igg.debus at gmail.com> napisał(a):

Am 11.12.2018 um 08:10 schrieb David G Dixon <dixon at mail.ubc.ca>:

There is now

only one tiny problem: There is still an almost imperceptible click when the

CV crosses 0V (and the logic control voltage crosses 2.5V) in the positive

direction.  This is when the logic switches from 011 to 100 (i.e., all three

bits change).  Interestingly, I don't hear the click at all when the CV goes

in the other direction (100 to 011).

Couldn't this still be caused by a very brief "forbidden state" during the
transition from 011 to 100? In your case, the forbidden state would be 000
(MSBit changes slower that the other two) or 111 (MSBit changes faster).
Probably just a difference between rising and falling slope. Can you check
with a scope?



Synth-diy mailing list

Synth-diy at synth-diy.org


Synth-diy mailing list
Synth-diy at synth-diy.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20181214/7bcd9cc5/attachment.htm>

More information about the Synth-diy mailing list