EEVblog #941 - The Amazing Schmitt Trigger

[EEVblog]

Fundamentals Friday
What is a Schmitt trigger and how does it work?
What is hysteresis?
And how do they fix two common problems in electronics, namely slow slew rate signals on CMOS digital chip inputs casuing metastability, and noise on comparator inputs.
The issues are demonstrated first on the breadboard, then the whiteboard explanation, and then the hysteresis fix is added and demonstrated.
And how are Schmitt Triggers physically implemented in CMOS logic?

[tggzzz]

Low slew rate and metastability are completely unrelated issues. Schmitt triggers cannot reduce metastability.

Low slew rate causes outputs to "oscillate" up and down whenever the input is near the threshold voltage, because the input is an invalid input voltage.

Metastability occurs when a flip-flop's input transitions at exactly the wrong time (in the metastability window), and the flip-flop "can't decide which way to go". Metastability can happen when both the clock and input signal are correct well behaved logic signals, but there is no defined phase relationship between them. In a properly designed system that should only occur in synchronisers.

Metastability is rare and difficult to demonstrate, so much so that it took the best part of a decade for it to be accepted. The reason for the rarity is that the metastability window is very very small, much much less than the maximum clock frequency.

OTOH, Low slew rate problems will probably occur on every transition.

[EEVblog]

Metastability occurs when a flip-flop's input transitions at exactly the wrong time (in the metastability window), and the flip-flop "can't decide which way to go". Metastability can happen when both the clock and input signal are correct well behaved logic signals, but there is no defined phase relationship between them. In a properly designed system that should only occur in synchronisers.

Yes, and I've explained that in another video somewhere I'm sure. Obviously it's not metastability in terms of setup and hold times for clocked system, that's a different thing entirely. I should have made that clear.
Actually, I think I might edit the video again to make this clear.

[Tom45]

What is that riding on the top of the square wave display at about the 5 minute mark? Appears on the high and low flat of the square wave, but only after the trigger.

[vitormhenrique]

video is not working here... marked as private video...
 

[lazarusr]

video is not working here... marked as private video...
 

Ditto.

[Andy Watson]

video is not working here... marked as private video...
 

Ditto.

It was working, but now it isn't. I suspect Dave has taken it down to make some edits.

[EEVblog]

A separate video on metastability  maybe?

Maybe.
I could have sworn I've talked about it somewhere...

[EEVblog]

It was working, but now it isn't. I suspect Dave has taken it down to make some edits.

Yep, some edits to tidy it up.

[nctnico]

It was working, but now it isn't. I suspect Dave has taken it down to make some edits.

Yep, some edits to tidy it up.

Well then I'll watch it after taking a nap first.

[ez24]

Blocked for me too

[EEVblog]

It's back up.

[Jaunedeau]

Haha, this night was the first time I used a schmitt trigger (to clean the output from a phototransistor).

It did not work well, I still hade some 1-10 wuick state switch before the signal stabilise (74hc14 from TI).

I then made a very simple circuit with a potentiometer, the schmitt trigger, and a led, set the input to 1.95V and... tada, the led is blinking !!

Lesson learnt : dont trust circuits you bought on aliexpress

[hedley]

Dave - Thanks  again . From your early days fundamentals Friday has been the highlight of your channel for me . In South Africa there are +- 52 Fridays a year so please keep them coming . Worked examples with circuit , datasheet and scope measurements are unique to your channel and really add insight to viewers.I even forced my spousal unit to watch the 34063 episode I think #110 which got me hooked.

[EEVblog]

I even forced my spousal unit to watch the 34063 episode

Don't do that! 

[bktemp]

Nice cheating on the practical demonstration 
Does it make any difference if you add capacitors across Vcc&Vss?
You can see a small glitch in the middle of the ramp, this is probably the point where the gates are switching. As you probably know, CMOS gates can draw a significant current (compared to the idle current) while switching. Whitout any capacitors on the supply rail, the voltage will drop and this will shift the input threshold voltage causing the input gate to toggle while the input voltage is near the threshold voltage. Adding a capacitor may reduce help with a slow clock (I'm saying it the correct way to do it, always use a fast clock egde, but even then you need a good decoupling of the supply voltage).

[Jaunedeau]

No one makes a cent by selling fake 7414.

No, but by selling parts that has been wrongly marked because or a mistake, yes.

They are already down to cent range even form TI, let along from many cheap but legit Chinese companies.

Btw, if I can send you a picture of the IC, with a badly printed TI logo, you'd know there is a broblem

Before questioning parts, first question your decoupling. If your naked eyes can see it blinks, it is definitely not "IC speed", therefore think more on external circuit.

I did, and when my circuit came down a 3.7V lithium battery, one 74hc14, one resistor, one LED, and one potentiometer, when my (admittedly cheap) multimeter showed 0V on AC voltage on the input, 1.9V DC voltage on input, and the led was randomly blinking, my conclusion is : this part HAS a problem. They were sold by 10 pieces and none of them have any hysteresis.

[EEVblog]

Nice cheating on the practical demonstration 
Does it make any difference if you add capacitors across Vcc&Vss?

A little, but still does it.

[jonovid]

like the video  but you did not show 555 timer as schmitt trigger or transistor logic trigger circuit only FETs

[sibeen]

Nice video, but, are you sure that the Schmitt trigger symbol comes about as you stated.

I was taught, many moons ago, that it comes from the hysterisis loop (B-H curve) exhibited by magnetic material.

[rch]

Nice video, but, are you sure that the Schmitt trigger symbol comes about as you stated.

I was taught, many moons ago, that it comes from the hysterisis loop (B-H curve) exhibited by magnetic material.

I would think your teacher was mistaken. Magnetisation hysteresis may well have been historically one of the first examples plotted like that, but it was a long time ago.  It may also be one of the first examples taught in school, but the hysteresis of heater thermostats was also a popular example.  (I expect it is used to show the advantage of  PID now, with much more maths!)

By the time a symbol for Schmitt triggers was needed it was a pretty commonplace engineering idea, so I am pretty sure the logo was based on the hysteresis loop of Schmitt trigger!

[Brumby]

Nice video, but, are you sure that the Schmitt trigger symbol comes about as you stated.

I was taught, many moons ago, that it comes from the hysterisis loop (B-H curve) exhibited by magnetic material.

Hysteresis is a descriptive term - the graphical representation of which on an X-Y graph gives the shape we all know.
 * You can find it with magnetic materials - and this is the most commonly known example.
 * Plotting the function of the Scmitt trigger also gives it.

... and there could be numerous other examples.


Hysteresis is not defined by the B-H curve.

[bills]

Thanks Dave even a dumb a-- like me got it, I have been trying to understand this for some time now.
bill