Rule of Thumb question

mrtweaver

Member
Join Date
May 2007
Location
Bloomsburg
Posts
329
I was wondering, what would be considered the best rule of thumb when it comes to PLC scan time versus input speed time?

An example would be measurement wheel, the measurement wheel provides 12 pulses per revolution, this translates to 1 pulse = 1 inch, the machine runs 243 feet per min, when all the math is dont it equates to about 1 pulse per 20mS. So if the PLC scan time is 20mS that would be bad because it would most likely miss pulses. So what would be the optimum speed at which the PLC should scan in order to keep up with the input?

Thanks and have a great day.
 
There is no rule of thumb for scan time, as far as I know. In your example, use a counter input input so that you won't missany pulse at whatever scan time PLC takes. Actually your program should not have too much dependency on PLC scan time
 
Your program must sense both the On and the Off-periods of the pulses.
And also consider that the on and off periods may not be of equal length.
So the PLC scantime must be approx maximum 90% of either the on period or the off period, whichever is the shortest. And this is not the mean value, it is the max value. Some PLCs reports the mean value when you check the scantime online.
If possible, catch the scan time by the program and store the max value in a register which you can check at any time.

You also have to check the inputs propagation delay in the same way as the PLC scan time.
 
Btw. you also have to consider the propagation delay of the output from the sensor that generate the impulses. It may well have different turn-on and turn-off times. If the sensor detects on a symmetrical tab on a shaft for example, but has 3 ms turn-on time, and 7 ms turn-off time, the 4 ms difference will mean that the PLC input will have 4 ms shorter time to sense the off-part than the on-part of the period.
 
My rule of thumb is to have the scan time at least 10 times faster than the fastest input. So for 20 ms pulses I would try to get 2 ms scan time. If it is really tight, I may go with 5 times rule (4 ms scan time in this case) but no more than that.

If such scan times are not achievable you may consider a faster processor, a high-speen counter input module or some kind of interrupt-driven processing.
 
My rule of thumb is to have the scan time at least 10 times faster than the fastest input.
That is a little bit of overkill but at least you are safe. I know our FPGA and most encoder chips use a frequency of at least 3 times the highest frequency. The idea is that the input goes through 3 D flip flops and all three D flip flop outputs must be high or must be low to change state. This is why the sample rate must be 3 times the input frequency. In a PLC this would require keeping track of the last three encoder inputs. 5 rungs should do the trick with 3 times over sampling. Three rungs keep track of the last 3 states. One rung sets the output if all are high and another run resets the output if all are low.
 

Similar Topics

Hi there Actually dont know whether shielded cable being used should be grounded on both sides or only at one side. Is rule same for both power...
Replies
14
Views
7,764
I do not use ferrules on my panel terminals. I use screw terminal. (Old school guy) For those of you that use ferrules on your stranded wire...
Replies
9
Views
2,544
I am getting off into using ferrules on some of our PLC IO that has spring and not screw connection terminals. Now a choice in crimping tools... I...
Replies
27
Views
10,322
I want to start using Ferrules in my panels but the thing is the DIN standard has a certain color for a certain size which means you may have a...
Replies
6
Views
2,015
Sorry for being off topic, but I'm looking for either a wire crimper that comes with a calibration certificate, or somewhere to send our crimper...
Replies
4
Views
1,798
Back
Top Bottom