Z-speed switch

I'm having trouble getting a simple zero speed setup to work. when the machine is running my prox is sending pulses at 90hz and the plc only scans at 50hz. Is this a problem.

It shouldn't be a problem if all you are trying to do is detect if the shaft is turning or stopped.

If your not using a hi speed input, your machine may not be able to see the pulses, aka the signals alway on. If the target shaft stops on the sensor, as far as the program's concerned there's no difference. I suspect that your program sees the shaft as always running.

What is the duration of the target (in deg of rotation)?. is there more than one target? What is the shaft speed, and target length etc?

How are you reading this in your program?

The bottom line is that you need to be able to detect the pulses. When the pulses stop, either on or off, the shaft has stopped. In order to do this you need a high speed counter, or a pulse train frequency thats lower than your scan time.

Mike.

I just finished a couple of projects where I used a Red Lion IFMA pulse to frequency module to convert the prox pulses to an analog signal, in my case 4-20 mA. If you already have anaolg I/O this may be the way to go.

I dont know the size of the target in degrees. the prox is seeing three 3/4" bolt heads equally spaced about 4 inches apart. I estimated that the prox would be operating at 90 hz and the prox says it can operate at 300 hz. the program is set up to reset a timer every time it see's a transition on the prox. the plc acts like its not seeing any thing happen at all. i'll probably just get a red lion unit. but i would like to know how you set up a z speed switch with just a prox and a plc. thanx for all your input.
geoff.

Questions, questions, questions...

Are you using a High Speed counter?
By 90hz I assume you mean that it pulses every 11ms or so?
OR is it pulsing 3 times @ 90hz which would be 3.7ms per pulse?
Can your plc scan time handle this?

Simplest solution to me is use just one bolt to sense. Using your hz scenario that should place it at 30 hz or 33ms per pulse which should be easy for most plcs. Use just one to test it.

If you know that it should sense a pulse say every 100ms (34 ms in this case) or less than you can set an alarm to detect missed pulses...x number of missed pulses triggers the alarm OR know it has stopped. ALSO if it stops ON a bolt then use time to determine it is stopped...this could also be used for an alarm if suppose to be running.

You know me, I get confused so have to ask alot of questions.

Ron, keep in mind what Mike stated about the duration of the target. The prox has to be ON for some minimum duration in order for the PLC to 'see' it as ON. Same holds true for OFF. A simple bolt head may have too short of an ON time for the PLC to read it, especially when you get up in the RPMs.

I'd try for a 50% duty cycle on the switch. Therefore, with one target, one 180° flag. Two targets, two 90° flags, 180° apart. Ideally, the more flags the better, up to the frequency limits of the prox or PLC input (at your max. RPM). This will allow the PLC to reliably see both ON and OFF states of the switch. Then it's just a matter of using timers to monitor the ON and OFF duration of the switch. If either timer expires, the shaft is at (or damned close to, depending on timer presets) zero speed. More targets = better 'guarantee' of TRUE zero speed... :nodi:

You could get by with one timer if you use the transitions to reset the timer. The 50% duty cycle becomes more important with this method though, as you only have ONE timer preset!...

beerchug

-Eric

Instead of a high rate input from a proxy, if you are just looking for zero-speed state as compared to being in motion, use an AirPax zero-speed switch. There are also other brands.

This looks like a 12mm proxy, but is a magnetic switch. It requires a 5VDC input and if at least 5 teeth per second pass by, the output goes on & stays on as long as there is movement - no pulsing.

I'm sure there are other ratings for these, but the ones I have come across are all on Minster stamping presses and they use them to make sure the gearshaft is turning when the clutch is engaged.

Also, Banner makes a Zero-Speed detector that monitors regeneration of electric motors to see if they are still moving. They commonly call them Anti-Plugging Devices. (No reverse until the motor stops going forward - or "plugging" the moving motor) http://www.bannerengineering.com/

Greg

grobertson

Based on your earlier statements, 3 targets, and frequency of 90 hz, your shaft speed is approx 1350 rpm. This would require the scan time of a late model Cray III Super PLC, or a counter module. However one target, with a 180 deg duration should provide a low enoough freq, and wide enough pulse width for your program to pick up, (Depending on the scan time). Max program scan time for your shaft speed, and 180 deg pulse would be about 40 ms.

Greg, Your taking all the fun out of it...It's called reinventing the wheel, or programming around missapplication of hardware.

Seriously though, they do make devices for this purpose. check Gregs link.

Mike

One big 180 degree target isn't going to give him much resolution when it comes to detecting zero speed. I believe he is going to have to buy a speed sensor like Tom or Greg used.

I tested one from a company called electro. It had a trim pot on the back to adjust your frequency threshold. The output was a single on/off. It was approx US$150.

Eric I was thinking of that, I have done things with bolts/prox that worked fine if you allowed OFF time, the ON time could be fast/short. The bolt thing works ok with 1750 rpm shafts. Has for me that is. If you need to SEE the prox for a longer period, just weld a piece of key stock across the bolt head.

Except fot the RPM part I am close to Mike in my calculations. I expect he uses the math more so being close is ok for me.

Sometimes you buy components to do something, sometimes you experiment with what you have to see what works.

What PLC?

You might be able to use an input interrupt for change of state. In the interrupt routine, set a flag every change of state to reset a timer in the main program. Or take a time-stamp value

Rick Densing said:

One big 180 degree target isn't going to give him much resolution when it comes to detecting zero speed.

Click to expand...

Is absolutely correct, and that's why...

Me said:

Ideally, the more flags the better, up to the frequency limits of the prox or PLC input (at your max. RPM).

Click to expand...

But then again, you can always crank up the timer's preset to wait as long as needed...

beerchug

-Eric

grobertson has a couple of problems. The first is frequency - without a high speed counter input his program won't pick up the pulses, and a standard input will appear to be always on or always off.

The second problem is pulse duration and prox response time. If the bolt is moving too fast, the prox won't pick up the bolt head as it goes by. I recently had this problem on a project, where the bolts were on an 18" diameter bolt circle at 3,600 rpm. I ended up replacing the magnetic prox with a photo switch and putting reflective tape on 1/2 the shaft circumference.

Even with a high speed counter grobertson will have to make sure the prox response is fast enough to pick up the bolt head.

Tom,

Thanks for the equations,