Hard limiting a CV?

Started by Skrogh, July 24, 2011, 03:43:12 PM

Previous topic - Next topic

Skrogh

Is there any easy way to hard-limit a CV, without clipping an op-amp (as this could damage the opamp, I've heard; LTspice also shows that this greatly increases power consumption).
Diodes in the feedback loop of an inverting op-amp has a way to "soft" limiting slope for my liking.

Thanks for any help with this, probably noobish question  ::)

Hides-His-Eyes

What's the source impedance?

Skrogh

At the point where I want the limiting to happen, the signal/CV is already buffered, so pretty low I guess?

CynicalMan



Put diodes in series to change the clipping voltage (the sum of the diodes' forward voltages). A value of 10k for R1 is normally used.

Or, look at section 2-2-4-3 here for an adjustable op-amp clipper:
http://www.classle.net/sites/default/files/text/41/non_linear.pdf

Skrogh

Thanks a bunch for the article!
LTspice still shows that the op-amp consumes about 20mA when in the clipping "zone" in opposition to 1-2mA when working linearly. Furthermore the inverting and non-inverting inputs comes awfully far apart from each other, am I being overcautious when thinking, this might harm the op-amp?

the other Idea, with a diode to ground (or in the feedback loop), clips too softly, at least with a 1n4148.  Do you know of a diode (or other device) with a more harsh clipping curve

CynicalMan

#5
That should be okay. What matters more is the power dissipation. Alt-click the op amp on LTSpice to see it. Then you can look up the datasheet for the op amp, which will show the maximum power dissipation. For example, the maximum for a LM741 is 500mW.

Edit: I just tried it in LTSpice, and I'm getting input currents in the uAs. Are you sure your sim is correct? A bad op amp model maybe?

Skrogh

Tried with a different model, fixed the problem. Cool thanks.

But it would still be far more convenient, if I could just use some hard clipping diodes.  I heard that LED's clips harder than signal and germanium, any truth to this?

slacker

I was going to suggest the same active clipper Alex posted about, I've used it and it works very well.

What are you actually trying to do with this, depending on what it is there may be ways to do it without having to clip the CV?

Skrogh

The envelope part of a filter really quickly gets too "open", if the filter is opened to much it becomes unstable. But i solved the problem with some LEDs for clipping diodes.

Thanks for the help anyway :)

CynicalMan

That's weird, because LEDs don't clip as hard as silicon diodes. Still, if it works, it works.

Skrogh

Silicons clip at a lower value, if you normalize the input voltage, LEDs actually clips harder (In a simple feed-back loop)
Does that make sense?