Looking for two-input comparator with hysteresis

[Pooh Bear]

I'm concerned about connecting the lower power rail back to the
transformer winding - surely they won't be at the same voltage because
of the rectifier voltage drop?

You could go half wave and connect the CT to ground. That'll need a CT 24V sec
though.

In this case just take the ac signal to the schmitt from either end of the TX.
It'll aproximate a square wave at line frequency now.

Graham

 

[Pooh Bear]

Thanks Spehro. I think I see what you're saying, though I still need to
have a think to convince myself as to why a resistor is needed at both
ends of the transformer winding.

Only one resistor in series ( as part of the PD ) is needed between the schmitt
input and the transformer connection. Since the voltage now aonly goes about 1V
below ground and with the potential divider helping you also won't need a series
diode.

Graham

 

[Pooh Bear]

I've thought about this a bit more. If you connect the "top" end of the
transformer winding to the Schmitt trigger input through a 100k ohm
resistor, then connect the Schmitt trigger input to logic ground through
a 22 nF capacitor, you form a potential divider which also acts as a low
pass filter. Presumably you are suggesting connecting the logic ground
to the "bottom" end of the transformer winding through a 10k ohm
resistor

No. No additional connection is required. The power supply provides the return
part of the siganl path.

Graham

 

[Pooh Bear]

Thanks for the very useful advice. I'm looking for something simple
which just works reliably, and this seems to fit the bill. If I've done
my calculations right a potential divider consisting of a 100k ohm
resistor and 22 nF capacitor

An R and a cap are a filter not a PD.

gives a maximum voltage of around 10 V
between the Schmitt trigger input and logic ground (for a 50 Hz signal).

And the schmitt is running of how many volts ? 5 ? The job of the PD is to reduce
that 10V to a voltage within the supply rails of the schmitt.

Actually, yout 12V ac winding will produce about 17V pk-pk so a PD of just
greater than 3:1 should be used. Make that 100k into 330k and add a 100k from
schmitt input to ground in parallel with the 22n ( this won't alter the RC time
constant very much ).

Are the 100 k ohm and 22 nF standard values,
Yes.

or did you and John
calculate them specially?
No.

Should I connect another resistor between
logic ground and the "bottom" node of the bridge rectifier?

No. Not required.

Graham

 

[Pooh Bear]

The 60 Hz line is sort of the definition of clock time.

50 Hz where the OP lives actually. ;-)

Graham

 

[Pooh Bear]

The clock hands move once per minute. I would consider a 5 minute error
in 6 months to be acceptable accuracy, and from what I know the UK power
grid should be capable of providing this.

It's actually way more accurate than that. Remember that mechanically driven
clocks with ac synchronous motors have to keep time too.

Graham

 

[Pooh Bear]

And the problem with a CMOS schmidt trigger is...?

Now only the other week I asked if it's Schmitt or Shmitt and now you say
Schmidt ! Come on which one is it now ?

Graham

 

[Pooh Bear]

Not much, really, except for crappy predictability.

And the UK tendency to prefer to overdesign too.

Graham

 

[Pooh Bear]

OK, why not? I's been a couple of decades since people figured out
CMOS latchup and took suitable precautions. Modern HC parts are spec'd
to handle numbers like 50 mA without problems.

I still don't feel quite right doing it though. And for just *one resistor*
extra.

Graham

 

[Jim Thompson]

And the UK tendency to prefer to overdesign too.

Graham

Yep. Why use a 4x4 post when you have a Greek column handy ;-)

...Jim Thompson

 

[Pooh Bear]

I NEVER rely on ESD diodes for clamping... VERY BAD engineering
practice.

...Jim Thompson

I'm pleased to see I'm in rarefied company ! It's just so *cheap* doing that
too. ;-)

Graham

 

[Pooh Bear]

Hard to predict... but it'll be brand dependent. ESD structures are a
black art and are different for every foundry.

Exactly so. Unless a component has manufacturer specced data I'd avoid using it.

Why don't you try it and report back (and specify brand of 'HC14)?

The cost of which will be what ? vs 0.1c for a resistor in qty adding $1 total
BOM cost to a run of 1k pcs.

Graham

 

[Pooh Bear]

They divide potential, don't they?

On a frequency dependent basis ! It's called a filter !

I advise you not to get into the habit of using incorrect terms. It could prove
expensive in the long run and simply daft in the meantime.

The way they divide potential changes
with frequency, making them a filter, but I would argue that they're a
potential divider too.

No. A PD is quite specifically something else.

Graham

 

[Pooh Bear]

According to Wikipedia the inventor's name was Otto H. Schmitt.

That accords with the last info posted on the subject here too.

Ta, Graham

 

[Christopher Tidy]

Christopher Tidy wrote:




An R and a cap are a filter not a PD.

They divide potential, don't they? The way they divide potential changes
with frequency, making them a filter, but I would argue that they're a
potential divider too.

Chris

 

[Christopher Tidy]

Don Lancaster wrote:




Now only the other week I asked if it's Schmitt or Shmitt and now you say
Schmidt ! Come on which one is it now ?

According to Wikipedia the inventor's name was Otto H. Schmitt.

Chris

 

[Christopher Tidy]

Christopher Tidy wrote:




On a frequency dependent basis ! It's called a filter !

I advise you not to get into the habit of using incorrect terms. It could prove
expensive in the long run and simply daft in the meantime.

That's interesting. So two capacitors in series would be a potential
divider, because the division of potential isn't frequency dependent?

Chris

 

[John Larkin]

50 Hz where the OP lives actually. ;-)

How many seconds do they have in a minute?

John

 

[John Larkin]

Exactly so. Unless a component has manufacturer specced data I'd avoid using it.

Well, they usually do. TI specs +-20 mA for its HC14, with no
footnotes about duration.

John

 

[Pooh Bear]

That's interesting. So two capacitors in series would be a potential
divider, because the division of potential isn't frequency dependent?

Yes, that can indeed be done. It has another useful property. Capacitors don't
dissipate energy so it's effectively lossless unlike a resistive divider.

Graham