Another encoder problem..

[scouringpad]

Hi - I'm new to this site, but have looked over past posts and don't see a definitive answer for my current problem. I'm using HSC1 in a Micrologix 1500 and an Allen-Bradley 600 PPR 2 channel encoder to measure rubber matting on a shear line. The encoder has a 1 foot wheel on it, and rides on the matting. About 33 feet of matting come by the encoder every 7 minutes or so, and are cut into rough lengths entered by an operator into the panelview hooked to the PLC. The problem is in the 6 minutes or so that the matting is essentially sitting (it creeps slowly backward as it shrinks while cooling). The encoder counts fine forward or backward while things are moving along, but while sitting still, or almost still, I'm getting erratic false counts. Sometimes none, sometimes a few inches forward or backward. There is slight vibration from an air table under the matting, but not much, and the rubber will normally shrink back about 3 inches between pulls out of the press. Encoder is connected with Belden shielded cable about 15 feet long.

As sometimes the reading is just what it should be, I suspected some noise on the line, but am not sure what else I might look at. Filters, scan times? Another encoder of the same type yields the same results. Though I did make the program that runs this shear line, I am fairly new to this, and have little experience with HSCs. The operators not knowing whether the first cut of the cycle will be long or short makes them generally cut on the long side for the whole pull, and this generates a lot of wasted trim. If anyone has an idea that would steer me in the right direction, I'd sure appreciate it!

 

[bkottaras]

If the encoder works fine while things are moving and only has issues when the material is sitting...
IF it only counts backwards (down) while the material is shrinking that is the nature of your process.
Counting forwards (up) doesn't sound good at all.
Sometimes counts up, down or nothing definetely leads me to believe that it is the vibration that goes through the mounting hardware into the encoder.
How is the encoder mounted and how it detects material rolling through?
Is it possible that the encoder wheel loses contact with the material while stopped and is freewheeing?
If it does, the vibration will be your issue and produce false counts in whichever direction.

 

[scouringpad]

It is on a hinged arm about three feet long, and gravity is what keeps it in contact with the rubber, so it's not losing contact. Would a small amout of vibration cause erronious counts? The air tables the rubber is conveyed on have numerous blowers connected to them, but I don't see how I could isolate from that..

 

[bkottaras]

That's all I can think of for now.
If it works well while moving then it is either the process that changes (but cannot count both ways regardless) and affects the count or the vibration induced through the mounting hardware.
What if you run, stop the process, lift and lock the wheel and observe?
If the count doesn't change while the wheel is locked then I would bet that the vibration does the trick while the material/encoder is idle!

 

[OkiePC]

Are ths signals single ended or differential?

What is the voltage?

 

[bernie_carlton]

Is the 2 channel encoder a quadrature type or a count up / count down type?

See if you can mark the wheel when it will be stopped for a while to see if there is any real movement.

Check any linkages for slippage.

 

[scouringpad]

That's all I can think of for now.
If it works well while moving then it is either the process that changes (but cannot count both ways regardless) and affects the count or the vibration induced through the mounting hardware.
What if you run, stop the process, lift and lock the wheel and observe?
If the count doesn't change while the wheel is locked then I would bet that the vibration does the trick while the material/encoder is idle!

Yes, it does appear the vibration is doing it. If I went with a coarser resolution encoder, I would expect the problem to remain, maybe on a lesser scale.. Maybe weighting the material down with a heavy wheel next to the encoder would damp the vibration enough..?

 

[scouringpad]

Are ths signals single ended or differential?

What is the voltage?

It's 11-24vdc in, 11-24vdc dld out.

 

[scouringpad]

Is the 2 channel encoder a quadrature type or a count up / count down type?

See if you can mark the wheel when it will be stopped for a while to see if there is any real movement.

Check any linkages for slippage.

Differential line driver. I have marked the wheel, and it moves backward as the matting shrinks. the shrinking mat has to pull two sets of pinch rolls back toward the press, so it is a somewhat jerky affair as it builds tension to move the rolls. Most of the 6 minutes it appears static. There is no linkage - the wheel is mounted on the encoder shaft, and is secure.

 

[Steve Bailey]

Picture a rubber band one foot long, fixed at one end, the other end free to move. If you stretch the free end by one inch, a point halfway between the fixed end and the free end will only move half an inch. I think that may be what you're seeing.

 

[scouringpad]

Picture a rubber band one foot long, fixed at one end, the other end free to move. If you stretch the free end by one inch, a point halfway between the fixed end and the free end will only move half an inch. I think that may be what you're seeing.

That would be true, but in this case the encoder is very near the free end - just a few feet from the shear. There is approximately 50 feet of rubber back to the fixed end (press). The movement under the encoder is about three inches backward per cycle. The problem is that sometimes it counts back 10 inches, or counts up two or three inches instead of the three less I should be seeing.

 

[OkiePC]

Well, I am not sure what the root cause is in your case.

Usually, higher voltages are less succeptible to noise (I have switched from 5 to 12vdc on some VFDs to help with noise).

A differential line driver quadrature encoder should be pretty immune too.

If the backward travel is jerky, could some slippage or backlash be occuring then? Is there anything you can do mechanically to smooth out that jerk?

Could there be some error in the logic that does the calculations? Can you post the code?

If no problem is found there, perhaps you can add a registration sensor and input interrupt to make a correction to the length value at some point during each feed cycle. For example, a sensor mounted 30" from the shear connected to an input with the interrupt programmed to poke in that value to the accumulator when it's triggered.

 

[bernie_carlton]

Ensure again that the signaling type from the encoder is quadrature and that it is set up that way in the 1500. Secondly review the filter settings for the inputs. It is possible that the sudden movement is generating pulses faster than the filter setting will allow. You should be able to get correct readings even with very high pulse rates.

 

[scouringpad]

Well, I am not sure what the root cause is in your case.

Usually, higher voltages are less succeptible to noise (I have switched from 5 to 12vdc on some VFDs to help with noise).

A differential line driver quadrature encoder should be pretty immune too.

If the backward travel is jerky, could some slippage or backlash be occuring then? Is there anything you can do mechanically to smooth out that jerk?

Could there be some error in the logic that does the calculations? Can you post the code?

If no problem is found there, perhaps you can add a registration sensor and input interrupt to make a correction to the length value at some point during each feed cycle. For example, a sensor mounted 30" from the shear connected to an input with the interrupt programmed to poke in that value to the accumulator when it's triggered.

The backward travel is not really too jerky as it only amounts to 1 to 3 inches in 6 minutes, and during the pull the encoder sees 33 feet in 40 seconds. I'm quite sure it is not slipping.

I'm not sure how to post the code other than as an .RSS file. Is there a simple converter to .pdf or some other readily readable file available?

I thought about a registration sensor, but the matting could be anywhere from 3 inches in front of the shear (if last cut was made at end of pull), to 10 feet past it (depending on what length they are cutting, and where the counter is at when the pull ends). The process involves making a continuous element of rubber matting that is then cut to size. If they are cutting 75 inches of product and the pull ends with 74 in the counter, The cut will be made right after pull start up.

Thanks for the help!

 

[OkiePC]

I'm not sure how to post the code other than as an .RSS file. Is there a simple converter to .pdf or some other readily readable file available?

Zip it first (right click, send to compressed (zipped) folder).

Then you should be able to attach it to a post.

Check out what Bernie has to say about quadrature type and setting up the 1500. I think he has quite a bit of experience with encoders and the Micrologix.