[sdiy] Critique my gate output buffer?

Paul Glass-Steel pfglasteel at gmail.com
Sun Dec 8 17:22:53 CET 2024


Roman,

I got sidetracked for a bit with holiday guests etc., but I revisited this
and I believe all the issues I was having were due to my poor
implementation of a single-transistor version. Your version as originally
posted seems fine with being shorted to any rail, wire-ORs beautifully, and
sags very little under even unrealistically demanding loads. The only
downside I've found so far (as I'm fine with the parts count/cost) is that
it seems to max out around 12 or 13 KHz (I hope to re-measure tomorrow). If
there's a way to increase the speed, that would be great; if not, that may
just be the compromise for offering the wire-OR functionality.

Thanks,
-Paul

On Thu, Nov 28, 2024 at 4:31 AM Roman Sowa <modular at go2.pl> wrote:

> Yes, there is a good way to limit the current, now that I know you use
> 7805 especially for this. Just insert a resistor before 7805. If it's
> powered from +12, then 200 ohm resistor from +12 to 7805 input will
> provide about 30-40mA current limit.
>
> BTW, shorting the output to either GND or -12V should not make any
> difference in dissipated heat for 5V regulator. If that happens,
> probably your base resistor is too low. It's the output transistor that
> should take the heat on itself in this circuit.
> You didn't even mention accidental short to +12V. Your poor 78L05 would
> not approve.
>
> PNP was used there to make gate output go very closely to 5V rail and
> also to provide good ORing capability. You could use NPN working as
> follower too, but then you loose 0.7V and walk on a thin ice having
> outputs in off state with 5V reversed voltage (from other ORed outputs)
> on the emiter.
>
> Roman
>
> W dniu 2024-11-26 o 20:34, Paul Glass-Steel pisze:
> > Roman,
> >
> > Thanks for this - I breadboarded it this morning and it does work well.
> > I did note that, though the circuit itself doesn't mind being shorted to
> > ground or even to the -12V rail, that doing so does heat up the 78L05 I
> > was using. Is there a good way to limit the current in fault conditions,
> > without compromising the voltage under load too much? Also, I'd be
> > interested in the technical reasons behind using PNP here, though the
> > inverted logic isn't a problem for me.
> >
> > Thanks again,
> > -Paul
> >
> > On Tue, Nov 26, 2024 at 4:29 AM Roman Sowa <modular at go2.pl
> > <mailto:modular at go2.pl>> wrote:
> >
> >     I can understand that you want the gate stay at constant level even
> >     with
> >     30 such outputs OR-ed, hence the 1k inside feedback loop. But if you
> >     don't need well defined current limit as in this case, why not just
> use
> >     PNP transistor? Short circuit current will be limited then by base
> >     current times beta. Just enough not to blow the transistor up in
> >     case of
> >     actual shorting outout to GND
> >
> >     See the attachment
> >
> >     And if your circuit offers reversed gate output somewhere anyway,
> then
> >     you can even omit T2.
> >     I know opamps may be cheaper now than transistors, but let's not
> >     overkill things.
> >
> >     Roman
> >
> >     W dniu 2024-11-25 o 17:42, Chris McDowell via Synth-diy pisze:
> >      > K one more.
> >      >
> >      > What happens here is the more outputs you OR together in this
> >     way, the
> >      > lower that 22k looks and the more voltage develops across your 1k
> >      > resistor, which basically wont ever really matter. I still think
> >      > tempting your op amp to overshoot or whatever funky thing it
> >     wants to do
> >      > with your diode and cable capacitance is unnecessary given that a
> >     220R
> >      > resistor in series with the diode would handle it. or put the
> >     parallel R
> >      > and C from output to minus input like you know you're supposed to
> :P
> >      >
> >      > in my personal system, everything has some manner of 100k to
> >     ground on
> >      > the inputs and I would leave the 22k out entirely. I guess what I
> >     would
> >      > /actually/ do is build an explicit OR module.
> >      >
> >      > Cheers,
> >      > Chris
> >      >
> >      >> On Nov 25, 2024, at 11:20 AM, Chris McDowell
> >     <declareupdate at gmail.com <mailto:declareupdate at gmail.com>>
> >      >> wrote:
> >      >>
> >      >> This is like swapping your output resistor with a diode, which
> >     means
> >      >> your capacitive load is only, yknow, capacitively loading when
> the
> >      >> diode is conducting. I don't think there is a lot of value in
> >     putting
> >      >> the output resistor in the feedback loop for gate outputs, as
> we'd
> >      >> usually do that to increase DC accuracy at the receiving
> >     equipment. I
> >      >> would move your output resistor out of the feedback loop to take
> >      >> stability out of the conversation. You then have something pretty
> >      >> normal that will OR with similar outputs, up to some limit where
> >     all
> >      >> your 100ks (and whatever all the various input resistances) in
> >      >> parallel are too low a resistance to get your gates across.
> >     that's a
> >      >> lot, though.
> >      >>
> >      >> Cheers,
> >      >> Chris
> >      >>
> >      >>> On Nov 24, 2024, at 12:42 PM, Paul Glass-Steel via Synth-diy
> >      >>> <synth-diy at synth-diy.org <mailto:synth-diy at synth-diy.org>>
> wrote:
> >      >>>
> >      >>> 
> >      >>> I'm hoping to improve on the typical transistor
> emitter-follower,
> >      >>> wire-OR capable gate output buffer that many of my modules
> >      >>> have inherited from CGS and NLC. I've come up with this:
> >      >>> https://tinyurl.com/259pt83t <https://tinyurl.com/259pt83t>
> >     <https://tinyurl.com/259pt83t <https://tinyurl.com/259pt83t>> - it
> >      >>> seems happy on my bench, the voltage is stable under a lot of
> >     fan-out
> >      >>> to multiple inputs and it seems ok with long cables / capacitive
> >      >>> loads. But I thought I'd check here to see if I'm missing
> >     something.
> >      >>>
> >      >>> Thanks,
> >      >>> -Paul
> >      >>>
> >      >>>
> >      >>> ________________________________________________________
> >      >>> This is the Synth-diy mailing list
> >      >>> Submit email to: Synth-diy at synth-diy.org
> >     <mailto:Synth-diy at synth-diy.org>
> >      >>> View archive at: https://synth-diy.org/pipermail/synth-diy/
> >     <https://synth-diy.org/pipermail/synth-diy/>
> >      >>> Check your settings at:
> >     https://synth-diy.org/mailman/listinfo/synth-diy
> >     <https://synth-diy.org/mailman/listinfo/synth-diy>
> >      >>> Selling or trading? Use marketplace at synth-diy.org
> >     <mailto:marketplace at synth-diy.org>
> >      >
> >      > ________________________________________________________
> >      > This is the Synth-diy mailing list
> >      > Submit email to: Synth-diy at synth-diy.org
> >     <mailto:Synth-diy at synth-diy.org>
> >      > View archive at: https://synth-diy.org/pipermail/synth-diy/
> >     <https://synth-diy.org/pipermail/synth-diy/>
> >      > Check your settings at:
> >     https://synth-diy.org/mailman/listinfo/synth-diy
> >     <https://synth-diy.org/mailman/listinfo/synth-diy>
> >      > Selling or trading? Use marketplace at synth-diy.org
> >     <mailto:marketplace at synth-diy.org>
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20241208/1323dbbe/attachment.htm>


More information about the Synth-diy mailing list