Recording a Servo Axis Move Profile

[robertmee]

I have to rollout a common CLX PLC program to a dozen or more installations, but each installation uses a different HMI front end. Some WW, some Ignition, Some FTSE and Some FTME. So what I'm needing to do I would normally do in a historian within a SCADA, but don't have that option here.

I have a rapid move of a rotary axis, open to 50 degrees, close to 0 degrees. The rise and fall of the move is limited to 40ms. The sedentary duration is probably a few seconds open, and 10 seconds closed. So a very rapid trapezoidal move.

I'm using adaptive tuning on 5700s, and the tuning is rock solid. However, just as with the HMI's this code has to easily adapt to 6500s as well, which will require manual tunes.

Part of the install of this Axis is a mechanical pully system. Historically, the maintenance techs have a hard time setting the tension. I've been asked to display a baseline of what a proper tensioned axis looks like, and then to run a test move to show what a newly tensioned axis looks like. Overlay the two to show any gross discrepancies.

Since no historian is going to record 40ms rise times, and even the CLX will struggle to poll samples within that rise and fall time, I'm first wondering if even possible. The idea would be to record position, torque, velocity and position error and generate my own trend with an array of samples. Conceivably a poor mechanical setup would show itself within one of these parameters.

Setting up a fast periodic routine and sampling the data into an array is the easy part. Displaying it depending on HMI might be a challenge....especially a FTME platform. But before I worry about the HMI side, is this a waste of time or is there a better means to alert to a potential mechanical setup issue? Anyone done anything similar to this?

 

[JTCat]

Are you able to program in a maintenance / calibration routine so that the rotary axis rotates slower for this purpose only? Then record this slower speed and chart it over time? I understand that the tension may not share a perfect linear relationship to speed, but some relationship may be able to be established that suits the objective.

 

[JTCat]

also, is it possible to use a strain gauge sensor in this application for setting tension?

 

[robertmee]

Are you able to program in a maintenance / calibration routine so that the rotary axis rotates slower for this purpose only? Then record this slower speed and chart it over time? I understand that the tension may not share a perfect linear relationship to speed, but some relationship may be able to be established that suits the objective.

Interesting....Outside the box thinking. I like it. It would certainly give me more resolution to note any anomalies.

As for the strain gauge, I'm not sure how that could be incorporated. The drive system is a pulley to pulley with a very short poly chain belt. The load pulley is stationary on the device, and the servo motor/pulley sits directly below on a cradle with jack screws for adjusting the tension. The tension is between the two pulleys and in the belt....tighten the screws which lowers the motor, which increases the tension of the belt. The problem is they tighten it too much and put too much strain on the MPL shaft bearing and prematurely causes failure. This motor cradle stays in place and they roll out many variations of the load side where no two are alike. So small variances in height creates the need for adjusting the motor cradle up and down.

Where could a strain gauge be incorporated that measures the tension in the belt between the pulleys? It's the deflection in the motor shaft that becomes the problem.

 

[robertmee]

Maybe something like this:
https://www.youtube.com/watch?v=m2inwbm9Q6E&t=186s

 

[JTCat]

I've installed strain gauges on a laminator before with the pulley belting on the feed rollers.

I would install a strongback to anchor to one end of the gauge and attach the other end to the adjustment mechanism, such as the motor mounting bracket if the belt is tensioned by movement of the motor mounting plate. I used very high tensile strength poly cord to anchor to the motor mounting plate for attaching to the strain gauge.

I would properly tension a belt then take the reading from the strain gauge sensor. Id adjust a high end and a low end and that small range was considered the acceptable range of tension. This allowed us to properly tension the belt even with the manufacturing variability between the belts.

 

[robertmee]

I've installed strain gauges on a laminator before with the pulley belting on the feed rollers.

I would install a strongback to anchor to one end of the gauge and attach the other end to the adjustment mechanism, such as the motor mounting bracket if the belt is tensioned by movement of the motor mounting plate. I used very high tensile strength poly cord to anchor to the motor mounting plate for attaching to the strain gauge.

I would properly tension a belt then take the reading from the strain gauge sensor. Id adjust a high end and a low end and that small range was considered the acceptable range of tension. This allowed us to properly tension the belt even with the manufacturing variability between the belts.

Yeah, that's because you have a slotted motor base and are using gravity to your advantage. The motor's not going to move in the horizontal direction and the earth is supporting the vertical load. In my case, rotate your picture 90 degrees. The motor base is sitting on vertical jack screws supporting it. Crank the screws down, moves the motor base down in the -Y direction and pulls the belt tight. A strain gauge won't work on the opposite side. Gravity, the weight of the motor, etc. are in the same axis as the strain gauge.

I did think about putting a strain gauge like your setup on the belt side with a turnbuckle. Install the motor and adjust the base until the belt is just tight, then zero the strain gauge with the turn buckle. Now lower the motor base using the strain gauge as the measurement. That would somewhat work but it's still subjective as to the original tightness of the belt and where the gauge is zeroed.

 

[JTCat]

Right. Your application is different, but this is just something I have done in the past that has worked well. In that configuration the strain gauge would be mounted above the motor and tension is examined when it is moved in the downward direction to tension the belt. It may not work if there isn't an unobstructed path above the motor.

If you are using a ball screw or jack screws to tension from a vertical standpoint you may be limited to using a deflection gauge or find out what the newton / meter (inch/lb, ft lb, kg m, etc) torque is on the jack screws for the correct tension an see if you can arrive at the correct tension by torque value on the jack screws.

That's just off the top of my head. There may yet be an easier way.

 

[robertmee]

Right. Your application is different, but this is just something I have done in the past that has worked well. In that configuration the strain gauge would be mounted above the motor and tension is examined when it is moved in the downward direction to tension the belt. It may not work if there isn't an unobstructed path above the motor.

If you are using a ball screw or jack screws to tension from a vertical standpoint you may be limited to using a deflection gauge or find out what the newton / meter (inch/lb, ft lb, kg m, etc) torque is on the jack screws for the correct tension an see if you can arrive at the correct tension by torque value on the jack screws.

That's just off the top of my head. There may yet be an easier way.

I'll be in one of the plants tomorrow. Going to take a closer look at the mech arrangement, now that I have an idea of the strain gauge. Appreciate the direction.