> Doesn't "schmidt trigger" imply some positive feedback to "snap" the
> output/internal gates to a digital state especially quickly, in addition
to
> the hysteresis?  (rather than leaving transistors/etc in a possibly linear
> region of operation, prone to oscillations and/or excessive power
> consumption.)  In fact, I thought the hysteresis was a byproduct and/or
> method used to achieve this capability, and the "important" part of a
> "schmidt trigger" was its ability to correctly operate with slowly
changing
> (ie analog) input levels...

No. It's all the same thing. "Schmitt trigger" (not Schmidt fwiw) means "has
hysteresis" {{as the circuit originally popularised by M. Schmitt had}}. No
other magic employed. The positive feedback is what causes the hysteresis.
The ability to operate with slowly changing levels occurs because the
input's trigger level rapidly moves away from the slowly changing input as
soon as the trigger level is crossed.

The input never gets to an intermediate state because there isn't one. When
you get past the trip point by even the minutest amount the trip point then
recedes rapidly and you are effectively instantaneously in the other input
state. The internal logic sees swift rail to rail transitions and has
minimal opportunity to draw large currents through simultaneously-on high
and low drivers.

Other schemes may be used to deal with slowly changing inputs but if they
don't involve hysteresis they are not Schmitt triggers. (Arguably positive
feedback is necessary to implement hysteresis - even if the decision making
is done digitally).



        Russell McMahon

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.