hysteresis of 74HC04

It is most definitely NOT hysteresis. It’s not a Schmitt-trigger input. All you can infer is that above 3.15V is a HIGH and below 1.35V is a LOW. Anything in between is indeterminate.

azadfalah said:

Is the difference or gap between these two voltages specified in the image below , the same as hysteresis?

Click to expand...

No.
That difference is simply the manufacturing tolerance between different batches of ICs.

azadfalah said:

if not, then what is the hysteresis voltage of 74HC04 ?

Click to expand...

It is likely very low (millivolts) and not specified.

If you want deliberate hysteresis then you need to use a Schmitt-trigger type gate, such as the 74HC14.

crutschow said:

If you want deliberate hysteresis then you need to use a Schmitt-trigger type gate, such as the 74HC14.

Click to expand...

Thank you Crutschow,

according to 74HC14 datasheet at vcc 5V we have ~ 1V hysteresis, for bigger hysteresis I should use comparator? Is there any logic gate with bigger hysteresis voltage?

1V is a lot of hysteresis. If you need more, then you’ll probably need to use a comparator. What’s your application where you need such large hysteresis?

Hi,

I agree, if you ned more than 1V hysteresis .. it's most probably no logic signal (input).
If you use a non logic (level) signal, you should tell us. (Because it's a special usecase)
HC04 is a logic IC, designed for logic level signals.

Klaus

barry said:

1V is a lot of hysteresis. If you need more, then you’ll probably need to use a comparator. What’s your application where you need such large hysteresis?

Click to expand...

KlausST said:

Hi,

I agree, if you ned more than 1V hysteresis .. it's most probably no logic signal (input).
If you use a non logic (level) signal, you should tell us. (Because it's a special usecase)
HC04 is a logic IC, designed for logic level signals.

Click to expand...

I am trying to connect a sinusoidal voltage with amplitude of 0.7 - 5V to the microcontroller interrupt
It seems that due to the hysteresis problem, interruptions are happening several times. I am trying to solve this problem

In fact, the high and low of a 50hz sinusoidal signal should be detected

How much latency can you live with in the measurement of high and low ? Can you
live with averaging a number of samples ?

In-so-far as the hysteresis problem in a varying amplitude environment, sounds like you
need a lot of gain, minimal hysteresis (if actual noise is small), centered on zero crossing,
eg. ac coupled.......a comparator.


Regards, Dana.

Multiple crossings of 50 Hz detector are typically caused by switching interferences, hysteresis doesn't necesserally help against it. A low pass filter in combination with hysteresis might help, but it also impairs precise phase detection. Another possible solution is a time filter, block timer interrupt for a certain time period after first occurrence.

Interesting thread : https://forum.arduino.cc/t/measuring-ac-line-frequency/187212/5


Regards, Dana.

Hi,

I don't understand,
Hysteresis usually does not cause a problem (multiple interrupts) instead it avoids exactly this.

Show scope pictures. Input and output.
Are you using HC04 or HC14?

Klaus

danadakk said:

How much latency can you live with in the measurement of high and low ? Can you
live with averaging a number of samples ?

In-so-far as the hysteresis problem in a varying amplitude environment, sounds like you
need a lot of gain, minimal hysteresis (if actual noise is small), centered on zero crossing,
eg. ac coupled.......a comparator.


Regards, Dana.

Click to expand...

azadfalah said:

I am trying to connect a sinusoidal voltage with amplitude of 0.7 - 5V to the microcontroller interrupt
It seems that due to the hysteresis problem, interruptions are happening several times. I am trying to solve this problem

In fact, the high and low of a 50hz sinusoidal signal should be detected

Click to expand...

Trying to ‘detect the high and low‘ of a sine wave are quite a different thing than what you’ve initially presented.

Are you trying to detect the peaks, or the zero crossings? If it’s just zero-crossing then you absolutely need hysteresis.

FvM said:

Multiple crossings of 50 Hz detector are typically caused by switching interferences, hysteresis doesn't necesserally help against it. A low pass filter in combination with hysteresis might help, but it also impairs precise phase detection. Another possible solution is a time filter, block timer interrupt for a certain time period after first occurrence.

Click to expand...

barry said:

Trying to ‘detect the high and low‘ of a sine wave are quite a different thing than what you’ve initially presented.

Are you trying to detect the peaks, or the zero crossings? If it’s just zero-crossing then you absolutely need hysteresis.

Click to expand...

I want to detect the drop of sinusoidal voltage in the first cycle which was not present and I want to use interrupt
Here, the 74HC04 IC is used. According to the information that friends gave me in the first posts, the possibility of several interruptions is normal according to the schematic below.




It seems that by replacing the IC 74HC04 with 74HC14, the problem will be solved to a large extent

Hi,

Mind that standard signal flow in a schematic is: left to right. GND and negative supply voltages should point down, positive supply voltage should point up. Thus your schematic is unnecessarily hard to read

***
You talk about "sine" waveform, but the input to the HC04 is far from being sinusoidal.
The output of the optocoupler is not sinusoidal, the input of the optocoupler is not sinusoidal.
If there is a asinusoidal signal, then most probably at a circuit that you don´t show at all.
--> you need to tell us the whole story. Otherwise we will have to guess. With a good chance that our guessing is wrong.

The shown circuit looks like a part of a ZCD circuit.
If so, we need to see the optocoupler input to somehow be able to predict the optocoupler output.

In either case: the circuit is designed poorly:
* C73 should act as a filter, but the value is much too small to work as expected
* for C73 to work as a filter, it needs a series resistor. In your case the rise time of the signal (HC04 input) is dominated by the R255, but the falltime is dominated by the optocoupler and it´s drive signal.
* U17 needs proper power supply bypass. Not only a fast capacitor, but also a puk capacitor, since the HC04 drives relatively high LED current.
* U17 the whole circuit needs a proper GND plane.

All in all there is a lot of information missing to give good advice. False interrupt trigger may come from
* poor PCB layout
* poor GND
* poor VCC (and decoupling)
* poor optocoupler input circuitry
* poor filter design
* maybe more...

***

maybe more:
If it´s a ZCD circuit, then most probably the application drives an SCR or triac, and in the end maybe high load current .. which may cause a lot of noise in the ZCD ciruit. In this case even the wiring of the triac/SCR is important for proper circuit function.

Klaus

KlausST said:

Hi,

Mind that standard signal flow in a schematic is: left to right. GND and negative supply voltages should point down, positive supply voltage should point up. Thus your schematic is unnecessarily hard to read

Click to expand...

Mr. Klaus, the circuit is in laboratory mode and without any noise
The input of 74HC04 has a low slew rate and is definitely not a standard logic, that's why I said it is sinusoidal because it is very similar to the input signal. The opto input is connected to an AC voltage of 110 volts with a 30k resistor and a bypass diode.

This is a normal frequncy standard board, also the bulk capacitor is located near the 74HC04

Please focus on my question, according to my research, the problem is that the hysteresis is small
Don't you think that considering the low slew rate and lack of sufficient hysteresis range is the first big problem in this design? if I increase the value of the input capacitor, Slewrate will decrease and more interruption will be created

KlausST, with respect, but most of the time you don't pay attention to the post question or to the answers of other friends

Thank you

Hi,

A paid atttention to your input: So lets see what happened.
* in post#1 you asked about the HC04 and hysteresis. All questions are answered properly in post#2..and post#3. Nothing more to say.
* In post#4 you switched to HC14 and asked about hysteresis and comparator. Barry answered correctly in post#5

****
Now to my post#6:
I tried to help by solving the problem. And here I mean the root problem. The problem is not hysteresis. The hysteresis is to cure the symptom. And hysteresis higher than 1V is unlikely to cure the symptom reliably.
Thus I tried to find out the root cause of the problem. (you know: better remove the root cause than to cure the symptom). The root cause obviously is unusual high noise in the input signal. It´s the almost the only way you can get multiple trigger output from a schmitt trigger circuit.
I asked to show scope pictures - you did not cooperate in this regard.

****
In post#7 you talked about sinusoidal.
I still can´t imagine how the signal should be sinusoidal at all. You failed to prove this.

****
In post#11 I focussed on you saying "hysteresis is the problem". I tried to tell you that hysteresis is not causing the problem but reducing the problem of multiple triggers. I failed in this. Sorry.

****
Posts#8..10 and post#12 gave similar informations. They asked for really important informations. You failed to cooperate on this.

****
Now in post#13:
You provided an non standard, incomplete and poorly designed circuit/schematic.
****

In post#14 I focussed on your given schematic. I told you what it does, and how/why it failes to operate correctly.
But the root cause - the source of noise - and how it enters the trigger circuit is still missing.
So I tried to use my experience (I´ve done 2kV RMS, 3000A RMS SCR switching uints with a precise, noise free trigger down to 20us pp for a silicon production plant) to show you where to look and how to solve your problem.

So from my experience in troubleshooting high power electronic circuits (from 1970 .. now) I´ve learned it´s good to take a step back to get the (whole) better view of the total system. Or just change the point of view to get a better view of the real problem. No need for you to do it the same way - as said, it´s just my experience.

****
In post#15 you complained about my posts. Sorry that I failed to guess what answer you expect.
(And please mind, that I don´t neccessarily give the same answer that others already have given. Also mind, that I simply don´t have the knowledge to answer all questions. Sometimes I still try to give side informations.)
I failed to tell you that 1V is a huge hysteresis (in my eyes 100uV is a small hysteresis). And in my ZCD circuits I never used more than 50mV of hysteresis.
I failed to tell you that the opposite of "Slewrate will decrease and more interruption will be created" is true.
--> As FvM pointed out: proper filtering is the way to go (removing the root cause = noise) , then you may even reduce hysteresis.

So I´ve honestly tried my best to really solve your problem. Tried to give you side informations, tried teach you that hysteresis is not causing multiple triggers. I failed in this.

To avoid to annoy you with more posts I better leave the thread.

Good luck.

Klaus

KlausST said:

Hi,

A paid atttention to your input: So lets see what happened.
* in post#1 you asked about the HC04 and hysteresis. All questions are answered properly in post#2..and post#3. Nothing more to say.
* In post#4 you switched to HC14 and asked about hysteresis and comparator. Barry answered correctly in post#5

****
Now to my post#6:
I tried to help by solving the problem. And here I mean the root problem. The problem is not hysteresis. The hysteresis is to cure the symptom. And hysteresis higher than 1V is unlikely to cure the symptom reliably.
Thus I tried to find out the root cause of the problem. (you know: better remove the root cause than to cure the symptom). The root cause obviously is unusual high noise in the input signal. It´s the almost the only way you can get multiple trigger output from a schmitt trigger circuit.
I asked to show scope pictures - you did not cooperate in this regard.

****
In post#7 you talked about sinusoidal.
I still can´t imagine how the signal should be sinusoidal at all. You failed to prove this.

****
In post#11 I focussed on you saying "hysteresis is the problem". I tried to tell you that hysteresis is not causing the problem but reducing the problem of multiple triggers. I failed in this. Sorry.

****
Posts#8..10 and post#12 gave similar informations. They asked for really important informations. You failed to cooperate on this.

****
Now in post#13:
You provided an non standard, incomplete and poorly designed circuit/schematic.
****

In post#14 I focussed on your given schematic. I told you what it does, and how/why it failes to operate correctly.
But the root cause - the source of noise - and how it enters the trigger circuit is still missing.
So I tried to use my experience (I´ve done 2kV RMS, 3000A RMS SCR switching uints with a precise, noise free trigger down to 20us pp for a silicon production plant) to show you where to look and how to solve your problem.

So from my experience in troubleshooting high power electronic circuits (from 1970 .. now) I´ve learned it´s good to take a step back to get the (whole) better view of the total system. Or just change the point of view to get a better view of the real problem. No need for you to do it the same way - as said, it´s just my experience.

****
In post#15 you complained about my posts. Sorry that I failed to guess what answer you expect.
(And please mind, that I don´t neccessarily give the same answer that others already have given. Also mind, that I simply don´t have the knowledge to answer all questions. Sometimes I still try to give side informations.)
I failed to tell you that 1V is a huge hysteresis (in my eyes 100uV is a small hysteresis). And in my ZCD circuits I never used more than 50mV of hysteresis.
I failed to tell you that the opposite of "Slewrate will decrease and more interruption will be created" is true.
--> As FvM pointed out: proper filtering is the way to go (removing the root cause = noise) , then you may even reduce hysteresis.

So I´ve honestly tried my best to really solve your problem. Tried to give you side informations, tried teach you that hysteresis is not causing multiple triggers. I failed in this.

To avoid to annoy you with more posts I better leave the thread.

Good luck.

Klaus

Click to expand...

Hi,

Mr. KlausST
Thank you for your kindness and attention, and I'm sorry, I didn't mean to ignore your kindness and help, I really appreciate your honest guidance.

The problem started from where I thought that the difference between the two voltages Vih and Vil in the 74HC04 IC is the hysteresis, but this voltage is not hysteresis and the hysteresis of this IC is about a few millivolts.

I haven't tested this circuit with 74HC14 yet, my question was if there is a logic IC for this if I am looking for more hysteresis, I didn't mean that I tested the circuit with 74HC14 IC and the problem remains.

As soon as I get the 74HC14, I will test the circuit and report the results here

Thanks

Wow, yet another jerk.

“Most of the time you don’t pay attention.” How blatantly rude from someone who apparently doesn’t have a clue of what they’re doing, needs a LOT of guidance in performing a simple task, and doesn’t even know what he doesn’t know.

And here’s another bit of information for you, which you would know if you had exerted the minimum effort and actually read the data sheet: maximum allowable input rise time is 500 ns. A 50 Hz sine wave will exceed that.

barry said:

Wow, yet another jerk.

“Most of the time you don’t pay attention.” How blatantly rude from someone who apparently doesn’t have a clue of what they’re doing, needs a LOT of guidance in performing a simple task, and doesn’t even know what he doesn’t know.

And here’s another bit of information for you, which you would know if you had exerted the minimum effort and actually read the data sheet: maximum allowable input rise time is 500 ns. A 50 Hz sine wave will exceed that.

Click to expand...

In my previous threads, despite the schematics and complete information, they still did not focus on my question and questioned the main question. You can see it in the activity section of my profile.

Please keep the Subject, I will post the test result here and if the problem is not solved, we will deal with other issues one by one