behavior when CV out of range

[Magnus Danielson]

From: Buck Buchanan <voltagecontrolled at home.com>
Subject: behavior when CV out of range
Date: Tue, 19 Oct 1999 15:10:51 -0700

Hi Buck!

> Hi all, got a question for the group on basic design philosophy:
> 
> What's the consensus on how a given module should handle CVs that go
> "out of range."
> 
> One of many possible examples: Some VCO is set up to handle a CV range
> of say -5V to +5V (system CV range).  And lets just say that when the CV
> goes over +5V, the sweep tracking goes out the window, the freq. goes a
> little further and stops - no harm done.  And suppose on the "bottom"
> that the module stops oscillating as the CV drops below -5V.  There
> could be other symptoms (including device failure I suppose) and this
> type of thing could apply to any voltage controlled module/parameter. 
> I'm purposely being non-specific.
> 
> So what's the feeling out there?  Do you like a module that exhibits
> poor,  non calibrated, or even bizarre behavior at the end of it's
> ranges?  Or would you rather have it setup so it's not possible to go
> out of range (either with clamps, clippers, etc.)?
> 
> Personally, I've got some modules with precision clippers in the CV path
> so the range stops cold exactly where I want it too.  But maybe that's
> not the best thing to do especially in cases where nothing "bad" happens
> as CV goes out of range. 
> 
> I know one option is to set it up such that a CV *shouldn't* go out of
> range:  say when manual control (CV offset) is at "max" and input CV is
> at "max" then the total CV will be just at "maximum".  The drawback to
> this seems that without an external CV applied, the module can't be
> tuned to it's maximum range.  I believe some Serge modules fall into
> this category.
> 
> Part of the reason I ask is that I tackle this issue almost every time I
> start a new module and I notice that a lot of the designs out there
> don't *appear* to address this issue.
> 
> Any enlightenment is much appreciated!

For what it is worth, here is a few rule of thumb that I would try to stick to:

1) For voltages that may cause overstress of some sort - rail clamp it!
   A couple of 1N914/1N4148 are cheap and should not interact with normal
   signal. It is cheap insureance.

2) For voltages that does not cause oversress, you migth in some cases cause
   upset, such that the curcuit locks until you actually have to hit the big
   powerswitch. Try to avoid that this is allowed, unless it compromises your
   intended function/tweakability.

3) You can also end up in a situation where your curcuit saturate, so you need
   to back off before it may recover. This can happend in the Formant 24 dB
   filters for instance. It MAY be a cool effect, it may also be a very
   annoying one. If one finds that it is annoying, then trying to avoid it in
   normal use could be done at most times.

4) If you specify that the curcuit has some properties within some range, this
   is really what to expect. When you go outside that range you are out on the
   unknown. If you don't hit into the clamp or something, getting increasingly
   out of tune should be OK, as long as you track well enougth inside the
   range. Most of the time you start tracking worse and worse, so it is a
   matter on how bad you may go before you leave the range.

I think this is some of the properties that we expect of a sound design.

Oh, yeah... we don't mind if it actually work ;)

Cheers,
Magnus