This is what I was looking for!

The only thing I cannot figure out is how to use my input signal lenght, accumulated in a timer, to compare it with the system time. I miss a small piece of knowledge block I guess...

Any tip please?

------------------------------------------------------------------------------------------
Split from Why this doesn't work? S7-1200.

Hello,

you can also do the follow

• rising edge of the signal - read the system time and save them at a DB
• falling edge of the signal - read the system time and save them
• subtract both DTL and you have as result the signal length

The accuracy/precision is depending of your system and your senor etc.
If you need a very exactly time then is this solution not good enough, then you should use interrupts

regards
CaOd

Thanks for interest.

I think the method with rising/falling edge may no be the best for my application.

I have to measure a flow with pulses. The flow meter gives 6.33 pulses/s at demanded flow which, using a cyclic interupt and moving the data to a calc function, every 1s doesn't give me any accuracy, and I need at least 1s refresh rate for controlling my pump.

I thought to maybe count the lenght of the pulses by reading the time difference between the puls and system/local time and use some calculation function to average it over 1 sec period. 

Could it work? I spent the whole day yesterday trying to made it but fail miserably... 

Hi,

I would suggest using the TIME_TCK function to get a millisec timestamp at every rising edge.

Delta T = current TimeStamp - prev TimeStamp

The time function will overflow after +/- 40 days so ignore the Delta if it's negative.

It's probably a good idea to take an average of the last few seconds before using it to control a pump.

good luck,

Remark

Thank you both for the answers! I had a break and I'm continuing again.

I figured out that by using the hardware input, I actually need to enable and configure it in device config... I didn't know that. Now it seems more logical.

The problem I now have is that my T_DIFF block if always inactive. It worked once till I added a new CALC block in the OB1. Now, even after changing everything, emptying the OB1, re-dooing hardware config, it won't work...

I'm probably missing a small detail.... any tip?

Also, Remark, how can I find the function you're describing? 

This code you posted is in Cyclic OB or interrupt?

I'm not sure I',m following the question as it seem to have changed from measuring the time length of a pulse to measuring time between pulses. If is the latter you may consider HSC as in this example.

The code is in interupt OB40.

I want to measure the time between the pulses. 

Thanks for the suggestion, I'll check it later.

That probably is the reason it's not showing up correctly in monitor because you're trying to capture a millisecond  or so of execution.

Just make sure the result of the calculation is correct. 

Kal.A

That probably is the reason it's showing up correctly in monitor because you're trying to capture a millisecond  or so of execution.

Just make sure the result of the calculation is correct. 

 Changing it to seconds or minutes doesn't solve the problem... it stays always inactive so my calc block doesn't get any input and no calculation is possible... 

I just don't see what's wrong with it.

I'm trying the suggested method now.