[sdiy] Odp: Re: Running a Bipolar Comparator..

[Ben Bradley]

I read through the thread and I'm still not sure what you want to do
as in what range of output voltage you want, and especially if it's in
reference to the ground (midway between the bipolar supplies) or to
the -V supply, but it occurs to me that the LM319 may be the
comparator you want. It will run on a bipolar supply and also has a
"ground" input that connects to the emitter of the open-collector
output transistor. I've used it for driving the logic levels on the DG
series analog switches.

On Sat, Mar 25, 2017 at 3:00 PM, Roman <modular at go2.pl> wrote:
> If it's locked at "off", then maybe the gate input is below 0. Do you supply
> -12V to the 339 only, or the input pot or that 100k resistor too? Measure
> voltage at comparator inputs.
>
> And I think I finally know what you meant by saying "release gets shorted at
> extreme setting" which was totally confusing at first. If it was supplied
> from -12V, LM339 at "release" state was forcing the cap via diode and
> 100ohms down to the negative voltage. And if it was tantalum, it started
> conducting. This is probably why you didn't get the envelope go down to
> -12V, but maybe around -1V.
>
> Replace output buffer with non-inverting amplifier of rather small gain, say
> 1.1 (10k feedback, 100k to GND) or so, and add another resistor from +12V to
> [-] summing node to compensate for diode drop. You say it goes down to
> 0.65V. So in this case about 180k form +12V to opamp's [-].
> It will slowly go down after that, when the voltage across diode drops
> further. You can prevent that to some extent by adding another diode:
> cathode to the "release" diode's anode, anode to +12V via 100k. So the
> envelope base will be held at 0.65V, and ouptut amplifier will compensate
> that to 0V.
> Hope that's clear. I can draw a schemo if you like.
>
> Roman
>
> Dnia 25 marca 2017 18:51 Busby Bergson <busby.bergson at gmail.com> napisał(a):
>
> "no longer seems to work" in what way exactly?
>
> The threshold adjustment didn't seem to alter the gate at all. It was
> locked at "off." I could take a closer look.
>
> What this looks like is that the release curve is doing fine down to
> 0.65 volts, and then only drops very slowly after that.
>
> I did notice this... is there a diode that people use for this
> circuit, normally? In fact, maybe a better question at this point, is
> whether there is an general open-source version of this circuit
> function - it seems like I've hit a number of walls with it!
>
>
>
>
> On Sat, Mar 25, 2017 at 11:18 AM, Tom Wiltshire <tom at electricdruid.net>
> wrote:
>
>
> On 25 Mar 2017, at 14:26, Busby Bergson <busby.bergson at gmail.com> wrote:
>
> Thanks for all of the great information..this helps a lot.
>
> Been testing all morning, and for the time being, the 1N4148 diodes
> seem to have solved the issue.
> However, when I run the comparator from ground, instead of -12v, the
> circuit no longer seems to work. Any idea why?
>
>
> "no longer seems to work" in what way exactly? Doesn't do anything? Doesn't
> do what you expect? Does what?
>
> How close can the comparator operate to its supply rails? lowering the
> negative supply to -12V rather than ground effectively gives you a lot of
> "headroom" below zero. If the comparator struggles when close to the
> negative rail, that might be what you're seeing.
>
> Another issue you might hit (and this is presumably why there were germanium
> diodes in the circuit originally) is that as the voltage decays away, it
> will drop below the forward voltage of the diodes (about 0.65V) and then
> they'll stop conducting. What this looks like is that the release curve is
> doing fine down to 0.65 volts, and then only drops very slowly after that.
> The diodes will leak even if they don't conduct properly, so it carries on
> decaying, but it's slow.
>
> Germanium diodes will mitigate this problem since they have a lower voltage
> drop, so it won't kick in until later in the curve, by which point you maybe
> don't care so much.
>
> Presumably, if I replace the electrolytics, +/-12 will ultimately be fine,
> right? Also,
> electrolytics should be able to handle a small-ish reverse voltage,
> shouldn't they? That might just be wishful thinking…
>
>
> They might be able to handle it (for a while), but exposing them to it isn't
> good design. Work out a way that doesn't involve toasting things!
>
> HTH,
> Tom
>
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
>
>
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
>