simple oneshot(2)

didn't look at the pic, but first thought, "What happens when the unexpected happens?"

Like power up, up then off then up again, or very slow up. those kinds of things.

Oh just an RC high pass and diode to ground to "eat up" the reverse voltage spike.

Hmm that shouldn't be a problem. It only "sees" internal sharp pulses. I was more worried about too much current in the diode, or if the -0.3 volt from the diode clamp would be an issue. It seems like this must be (or must have been*) either a standard trick.. or something that is frowned upon.

George H.

Naw, today the 'component' is an iPod, or such. people are writing apps as the product and think in terms of a super complex system as the component. Whole new ballgame.

=_NextPart_000_002D_01CF59B2.47F14A40 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable

"George Herold" schreef in bericht = news: snipped-for-privacy@googlegroups.com...=20

Hi all, OK a silly basic question. I'm doing some digital stuff. =20 I wanted a one shot to reset some things, but didn't want to put in an = entire chip (74HC123.. or whatever the number is.) So I just used a cap, resistor, diode and inverter.

Thanks, George H. (This didn't appear the first time I posted it... trying again.)

If you need a clean reset puls like a 74HC121 or similar circuit will = provide, this is *not* OK. You will need some more components like in = the schematic below.

|| D2 |\

If the inverter weren't CMOS, and you can tolerate multiple edges near the end of the delay, yes.

For CMOS, slow-slewing inputs are highly discouraged (can cause high power consumption) unless the gate is one that has hysteresis (i.e. a Schmitt inverter).

At/near the logic threshold, a slow ramp voltage might result in power supply noise getting amplified (and that means lots of hash on the output), so hysteresis is your friend. Two thresholds, so you're never near both of 'em.

Either a '555 or a SN74LVC1G123 (both in SO-8 nowadays) would work. In CMOS, a slow-slewing input that approaches threshold will result in lots of hash (because power supply noise gets amplified), and possibly in high power consumption, UNLESS you have a Schmitt gate. The hysteresis of a Schmitt trigger solves both problems. Switching to TTL solves the transient power consumption issue.

Either a '555 or a SN74LVC1G123 (both in SO-8 nowadays) would work. In CMOS, a slow-slewing input that approaches threshold will result in lots of hash (because power supply noise gets amplified), and possibly in high power consumption, UNLESS you have a Schmitt gate. The hysteresis of a Schmitt trigger solves both problems. Switching to TTL solves the transient power consumption issue.

(having posting problems, hope this isn't a repeat send!)

This is a prop-delay one-shot:

That one makes a pulse that is one gate delay wide, but you can add any delay element in series with the lower inverter: coax, RC, more gates, or just a cap from the inverter output to ground, if you're feeling tacky.

It's all "logic legal" electrically.

Use an XOR gate, and it fires on both input edges.

Found of any old text book.

...Jim Thompson

The book that you need is "Killfiles for Dummies."

A week or so ago, JT claimed that he installed a script to killfile me and all followups. Of course, he's claimed that many times before, and lied every time. He claims he kills me, and then lurks. I bet he peeks in his neighbors windows, too... that's his personality.

Metastability? What can go metastable?

As far as width goes, it's similar to George's RC glitcher. Better, actually, in not having long time constants that will introduce pattern dependencies.

Hi guys, Well a follow-up on the oneshot. So this is with a 74HC14. (Schmitt trigger inverter.) Here's a link to a dropbox file. (I'm really loving dropbox.) (It's turning into a notebook supplement.)

George H.

OK that seems to work too. you might want to add some series resistance so as to not load the input as much when D1 conducts.

Hmm..If you take out C1 and D1 then it like a pulse stretcher. (I've done that sometime in the dim past.) I guess that's also part of mickey mouse logic.

George H.

OK that'll be a fairly short pulse.

Once I've put in enough other stuff I might as well just stick in a "real" one-shot. I had a few extra inverters, so the above was what I thought of.

Grin.. Ya mean my logic aint legal? The digital police will be after me :^)

OK I'll try and remember that trick.

George H.

Hmm google groups? It's a bit weird lately, my original post appeared (to me) several hours after I posted in.. and long after Phil responded.

George h.

The RC glitcher can have problems if C is small, and undershoots zero volts. Misdemeanors, probably not a felony.

I suspect you may be over-thinking this. ;)

Cheers

Phil Hobbs

AKA "Creeping Elegance" ?>:-} ...Jim Thompson