Thanks a lot

Do I understand this in the right manner ? You want to count impulses in a time of 9 seconds and three times with different counters. These counter chains should then be offset by three seconds each ?

For this I would use a second clocked counter that resets itself at the ninth second. A drop-out delay of 9 seconds is then triggered via the three analog differential threshold value switches for the values "0", "3" and "6". The fall delays then activate the individual counters.

The whole thing is quite time-critical and requires a delay of one cycle between the counting periods. However, this means a certain inaccuracy when counting.

greeting Ella_68

(Translated with Google)

ixo65

I don't understand the point of your reasoning.

Hi. Exemple with this topic : Flowrate with pulses Logo 8

In my case, i have a water flowmeter with puls output, variable between 0,33 and 5,55 pulses per second and I want the momentary flow every 10 secondes. The problem is a 10 second sample is not enough for accuracy. a 30 second sample is better.

Is it possible to make 3 samples delayed by 10 seconds between them.

At the end of each sample, so every 10 seconds, calculate the flow rate with the result of each of these samples

a.piquerel35

Ella_68

Do I understand this in the right manner ? You want to count impulses in a time of 9 seconds and three times with different counters. These counter chains should then be offset by three seconds each ?

 

Ella_68

For this I would use a second clocked counter that resets itself at the ninth second. A drop-out delay of 9 seconds is then triggered via the three analog differential threshold value switches for the values "0", "3" and "6". The fall delays then activate the individual counters.

Sorry I do not understand

A program often speaks more than 1000 words.

Here's an example.But I worked with the reset pulses for the individual counters, which leads to the same counting on the first run. This would then possibly have to be adjusted according to your own ideas.

Kind regards Ella_68

(translated with Google)

Thanks !

How did you do the test? On a logo or in the LSC simulator mode? The LSC not only runs slower than on a Logo, but is also heavily affected by the fluctuations caused by the program processing of the operating system. Normally this is not noticeable, but with time sensitive programs this can be a big problem. You also have a "dead" area where you don't count. E.g. the 0.01s for the clock generator or the delay that some blocks need for internal processing. This comes into play with high numbers of pulses per time unit.

But why are you stressing yourself ? Just get a threshold switch for the fast inputs with a capture for 3 seconds and then extrapolate to 9s .

Ella_68

(translated with Google)

 Yes I also just noticed that the 0.01s added a dead time. I replaced the asynchronous pulse generator with a timer on delay. The result is slightly better but yes I'm in simulation because I don't have the material with me. Thank you for this information, it reassures me!

Ella_68

But why are you stressing yourself ? Just get a threshold switch for the fast inputs with a capture for 3 seconds and then extrapolate to 9s .

 Look at the attchment of my first post of this topic :

If it's possible, it could be very easy.

 Hello a.piquerel35,

a.piquerel35

In reality, the period will be 30sec, maybe even 1 minute.

But if it's 30 seconds, then the user will have to wait 30 seconds to know the bitrate, which is too long. The objective is to see the effect of an adjustment on the flow.

By creating 3 periods of 30 seconds which are staggered by 10 seconds each, then the user will only have to wait 10 seconds.

I don't quite understand the meaning, or I'm making a mistake :

If the user only sees the first 10 seconds of a measurement time of 30 seconds because the value is then replaced by another measurement, the remaining 20 seconds are of no interest to the user.

It would be easier if the impulses were always recorded for 10 seconds and then extrapolated to the minute (impulses multiplied by 6), because the user is only interested in the instantaneous value.

This is also much easier and more stable to program and you can even shorten the sampling time to one second (impulses multiplied by 60). At this sampling rate, provided there are enough pulses, the user always knows the current count per second.

Or what is the deeper meaning of implementing it with three offset measuring sections, two thirds of which are not used ?

Ella_68

(translated with Google)

The 30 seconds are necessary because the pulse frequency is quite low. The longer the duration, the more accurate the result. but the refreshment will be long. By making 3 in my example it is not for 10 seconds that I count the pulses but for 30 seconds. By doing it staggered, I therefore have 3 periods and their results every 10 seconds.

I don't know if it's clearer