Hi Speed Counting

controlled

Lifetime Supporting Member
C
Join Date
Nov 2005
Location
bowmanville
Posts
419
Hello

I am having a problem on a machine trying to count gaps in sheeted boxboard. I am using a slc 5/04, with a banner reflective photoeye detecting the gaps. The machine counts perfectly until we run the machine upto and above 80% running speed. At these speed the counter starts to miss some of the gaps. I think the problem has to do with scan time. Does anyone have any ideas how to easily correct this. I am thinking that maybe a hi speed counter card would work. Any suggestions would be appreciated.

Derek
 
Derek,

If the only parameter that is being changed is the Speed of the machine, then the most likely causes are either the scan time, or the frequency limitation of the photoeye itself.

A high-speed counter module would be a logical next step.
 
controlled said:
Hello

I am having a problem on a machine trying to count gaps in sheeted boxboard. I am using a slc 5/04, with a banner reflective photoeye detecting the gaps. The machine counts perfectly until we run the machine upto and above 80% running speed. At these speed the counter starts to miss some of the gaps. I think the problem has to do with scan time. Does anyone have any ideas how to easily correct this. I am thinking that maybe a hi speed counter card would work. Any suggestions would be appreciated.

Derek
Click to expand...

It probably is scan and speed related i.e. the turn-on time of the PE does not last long enough for the plc to capture it. Look at the datasheet for the PE to see what its effective turn on time will be, just in case you are not familar with what turn-on time may be it is a time period the PE must be blocked (or unblocked) to effectively transmit a signal.
 
I agree that it is more likely related to the switch, not the scan time of the PLC.

Just a shot in the dark... You're not using an AC photoeye, I hope. Those things are much less responsive than the DC models. People still specifiy them since "all the other I/O is 120VAC".

As suggested above, check out the data sheets of the photoeye for specifications.
 
The datasheet for the sensor that I am using states that has a 4ms off/on response time. The scantime in my program is 20ms.
 
1. What is the width of the gap?
2. What is the linear speed of the material at 80%?
3. This gives a theoretical ON time of the PE.
4. Subtract the response time.
5. This yields your probable ON time to the PLC.

This all assumes that the gap is the shorter time. Apply the reasoning above to the shorter of the two (ON or OFF)

To count a scan must take place when the input is off and another when the input is on. With a 20 ms can the signal must be on for at least 20 ms and off for at least 20 ms to ensure counting. (This is assuming that the signal is going to a standard input.)

If you find that either the on or off time is getting lower than the scan time then you need to slow the line, increase the gap or distance between them, or use a faster acting input, like a high-speed counter card. With the scan time of 20ms the 4ms response time of the eye is of secondary importance.
 
In order to reliably count, you must sample the input at a frequency at least twice the frequency of the input signal. Twice is adequate for an input with 50% duty cycle. If the duty cycle is anything else, you need to sample at twice the inverse of the minimum pulse width.

Note that the pulse width seen by the PLC is made narrower by the response time of your photo eye and the filter on the PLC input card.

If the I/O is not the limiting factor, then perhaps putting the counter in a STI routine with an immediate input will solve your problem.
 
Thanks everyone

The lineal speed of the line at 80% is 900 ft/min, or 15 ft/sec.
The gap size is about 4", and the sheets are a minimum of 30".
My input is on when the eye is blocked and off when the eye is clear. The sensor switches from on to off probably 4 or 5 times per second.
 
I think (I will be corrected soon if i'm wrong), your 4" opening would be true for .72ms, so that would be close to what Gerry has stated and would satisfy what Bernie said....again I think.


But I would think that your Photoeye would be the problem

What I did in the past...I don't know if it would work in this app. was install a timer to hold the input signal for x amount of time, it was a mechanical limit switch on a cam lobe...the drive was not seeing it
 
One thing that I haven't seen mentioned that will affect the response time is not only do you have the switching time of the photoeye, but you also have the switching time of the input module. For a 1746-IB16 this would be 8ms on and 8ms off. This would mean somewhere around 24ms on the switching in the worst case. This is above the 20ms scan time
 
Looked up a 1746-IA16 just in case it is an AC input and the switching times look very close to Genius's numbers

On 35ms, Off 45ms.

Worst case would be 88ms. Are you using an AC prox as Jimmy Ohio believed?
 
Last edited:
geniusintraining said:
I think (I will be corrected soon if i'm wrong), your 4" opening would be true for .72ms, so that would be close to what Gerry has stated and would satisfy what Bernie said....again I think.


But I would think that your Photoeye would be the problem

What I did in the past...I don't know if it would work in this app. was install a timer to hold the input signal for x amount of time, it was a mechanical limit switch on a cam lobe...the drive was not seeing it
Click to expand...


hmmm, how do you know a 4 inch opening is good for 72ms? 15 feet a second is 900 inches a second, that means a 4 second gap is....looks to me as awfully close to one inch a millisecond, y'all do the math, not my forte.

Alas I may be wrong but looks to me like the PE can not turn-on to fulfill the needs of the PLC.

dangit, done it again. ooops u said .72ms, definitely not enough time.
 
Last edited:
I get that the beginning of a piece will come every 188 ms and the gap will be open for 25ms at 80%

But I went to work at 5 this morning and just got home so what do I know
 

Similar Topics

B
hgh speed counting
I am converting a machine from a micrologix 1100 to athe compactlogix 5069-L306ER controller. I have small disk about 0.930" and need to count at...
Replies
14
Views
3,704
Geospark
Geospark
J
High Speed Counting with a PE
How does everyone count items at a higher speed using a PE? I traditionally just use a sequence to count such as PE gets blocked for a set period...
Replies
8
Views
2,088
JosephMont
J
J
Help with SLC high speed counting (IIM and STI)
Hello, I have a machine that is counting plates and stacking them. I've found that when the gap is too small, the PLC gets the wrong count. Each...
Replies
2
Views
1,851
jake789
J
TimWilborne
Twido maximum counting speed via ladder logic
I'm working with a TWDLCAA24DRF that is having trouble missing pulses off of a flow meter. I'm not real familiar with the Twido PLCs but it...
Replies
6
Views
3,587
Lare
L
J
ML 1100 High Speed Counter not counting in high speed
Hello everyone, I am using a ML1100 and a 2048 pulse incremental two channel encoder. I have set up the high speed counter and the ML1100...
Replies
3
Views
3,196
Mickey
Mickey
Back
Top Bottom