Pulse width measurement with Arduino

[medelec35]

Noticed delay is off for some frequencies.
If change delay to 2 seconds will fix that.

[ChrisT66]

Hello Martin, incredibly it works fantastic!
And with so little code

Thank you very much! My (probably first) problem is solved

Now I have to monitor the pulse change, I don't care about exact ms, more about a fast or slow change of the high level and whether positive or negative to the origin.

Surely again something very simple...

Greetings from Bavaria Chris

[ChrisT66]

Big praise to the forum here, you both have always helped me quickly, I hope I can return the favor sometime

[mnf]

Glad to help..

Let us know how your project goes.

Martin

[ChrisT66]

Here comes the next question:

After the display output, I added another decision (see example) to output two PWM signals.

If the value is smaller, one LED lights up as it should be 100% at the PWM outputs. If the value is larger, one LED lights up 100% and the other one flashes slightly, why?

[ChrisT66]

since i changed to timer2 the hardware is also behaving strange. the read in values are starting to change themselves

[medelec35]

Glad I have helped also.
Looks like there is a bug with Timer2.
The frequency is double to what it should be.
The measure pulse correctly, you will need to have prescaler on 1:256

Bit confused with your pwm settings.

Since period overflow is 255, that means if duty is 255 the output will be near enough high all the time.
If duty is 0 the output will be low all the time.
So PWM is acting like a digital output.
If you want duty of 50% then set duty to 128:
Duty = PeriodRegister/(100/RequiredDuty)
Duty = 255/(100/50) ~128

[ChrisT66]

i thought i have to use timer2 because pwm also uses timer0? can i leave everything on timer0?

The PWM outputs were set to 255/0 to switch the lamps on and off for test purposes only. I noticed that one LED always flashes although it should be off.

[medelec35]

i thought i have to use timer2 because pwm also uses timer0?

What you did was correct.
As PWM only uses timer 0, you correctly changed timer to timer2

noticed that one LED always flashes although it should be off.

You are also correct here as well.
The issue you have is because the milliseconds resets to 0 for a new test.
Just place the decision branch containing PWM control, just above the 2 sec delay.
I.e withing the Pause branch that displays values on the display.

[ChrisT66]

the short blinking of the pwm2 led remains, although it should be off

[ChrisT66]

I'll have to add a delay to block out the short milliseconds reset...

[medelec35]

I have just added the 2 LEDs.
Working OK for me.
Freq:11.00Hz
LCD Display showing: 45ms
LED lit: CH2 (D5)

Freq:70.00Hz
LCD Display showing: 7ms
LED lit: CH1 (D6)

For me either one or other LED is solid.
As the tests could be different, giving unpredictable results
Can you let me know what frequencies are causing the issues, along with what the display shows.

The next state would be setting a max expected pulse length, so display could just show 0 for example.

[ChrisT66]

but not so easy. i think the whole period of time is affected until the value exceeds my threshold value again. is there any way to stabilise the millisecond value in order to hide this counting up phase?

[ChrisT66]

really no led blinking on you? weird.

I'm running the PWM at 6KHz, maybe that's the difference. I want to dim the LED using the PWM signal...

[ChrisT66]

now without a code change suddenly also the second one starts blinking when the other one is on

Maybe a poltergeist?

[medelec35]

If you post Flowchart, I will run it on my hardware.
Let me know what the frequencies going in D8 are?

[medelec35]

now without a code change suddenly also the second one starts blinking when the other one is on

Input floating?
Do you have a weak pull-up or pull-down on D8?

[ChrisT66]

here's the latest code. i'm using a servo tester as a signal source. the signal looks good on the oscilloscope

[medelec35]

You have not made the suggested modifications of

Just place the decision branch containing PWM control, just above the 2 sec delay.
I.e withing the Pause branch that displays values on the display.

:

Suggested Modifications..png (98.87 KiB) Viewed 7435 times

[medelec35]

I'm running the PWM at 6KHz, maybe that's the difference. I want to dim the LED using the PWM signal...

On flowchart you posted, PWM is set to 62.5kHz.

[ChrisT66]

but i already had, but it didn't work either. when i go to 2s it runs quite stable. but much too slow for my application. at one second there are mismeasurements (sometimes it counts down to 1ms instead of the smallest value of the servotester of 7ms in my display.
if you go from the smallest value (7ms) to the largest (17ms) the display remains at 2-3ms. only if you slowly adjust up and down the 17ms come back as maximum

[medelec35]

OK,

when i go to 2s it runs quite stable. but much too slow for my application.

OK, made changes so updates instantly.

[ChrisT66]

v3 does not work with my hardware


now the spook is gone. saved milliseconds after counting into the variable result

[medelec35]

v3 does not work with my hardware

Not sure why does not work for you but works well for me?
Anyway so long as your flowchart works for you, that the main thing.

[ChrisT66]

I gave up understanding these things a long time ago...

it still works, i'm just a bit optimistic, but only with your help, thanks!