RPM gone wild

I am using a ML1400 to count the RPM of several items. They all use the same math to determine the RPM after 10 counts of the counter. The one that has begun giving fits should never be over 35 rpm but in the last month has begun given results up to 32767, 2 times, but only at start up. It may be several days between unusual events. Once past a 10 second period, count is where it should be. I only graph the RPM every 5 seconds, but the calculations are done in less than a second.

I use a prox sensor to count the targets on a rotating wheel. The wheel has two targets. I use this count to warn or stop the process if it is to low. Chance of actual over speed is basically none and it causes no grief other than it is there. The other two items never show over speed.

What would you use as a formula to determine the rpm within 10 counts, or 5 revolutions?

I would say add in a routine that clears the registers and timers associated with the RPM calculations at start-up, that should fix the erroneous numbers at start up

Is there possibility for dividing by zero or math overrun error?

Do you zero your rpm counter when axle is stopped. I mean is there possibility that your counter has something like 7-9 counts allready when axle start to spin, so you would get false rpm reading because math is executed earlier than should? (from 1 count instead of 10)

I discovered that my logic was not exactly what it should have been for the 35 rpm, while putting in a CLR command for the counter as suggested. The result being the counter was holding an ACC value at start up

No div by zero, but the math was done before it should have been.

Thanks

Sounds like the counter is counting whilst the shaft is stopped. Probably something going on with the input circuit.

I use a prox sensor to count the targets on a rotating wheel. The wheel has two targets.

Click to expand...

Probably something going on with the input circuit.

Click to expand...

I agree. The wheel could stop with the prox on one of the 2 targets, then rock back and forth a few times. I would not enable the prox and counter until the wheel has been started up.

My mistake was having the counter reset at 12. I changed it to 11. Occasionally rotation would stop at count of 10 or eleven so at startup next day, for one shot, the count was already high with a very short time to reach it. The other items I monitor have at least 3 times the RPM and 12 works for those.

No counting or timing is done until I have an input from the motor starter. No motion within 1 second Stops the motor. Equipment is such that no rocking of the large targets can occur.

Got my data this AM, still not perfect, but at least no math errors, and the max count is below 250, but it takes 10-15 seconds for it to settle down.
Target to sensor distance varying maybe. Hmmm.

I can always make it not accept anything over 35 but that won't answer why.

You could try this method:

http://www.plctalk.net/qanda/showthread.php?t=76007&highlight=stroke+counter&page=2