Help with Logix 5000 motion camming / jerkiness

phuz

Member
Join Date
Jun 2008
Location
Mohnton, PA
Posts
1,043
Setup: Kinetix 5700 servo w/ MPM motor, CompactLogix L33ERM
Was Yaskawa now AB.

I'm working with a mill that has a longitudinal axis with an MPM motor. This axis tracks an incoming material via a 2500ppr encoder and mills slots into the material as it traverses through the mill. The machine that feeds the incoming material is belt-driven and pneumatically squeezed so that it can't slip. The material is also clamped in the mill during the "milling" part of the sequence. After the milling, a simple MAM instruction moves the axis back to waiting position for the next trigger.

The problem: I am getting jerkiness during testing without even having the piece clamped in the mill. I am simply letting it pass through while rolling over the encoder. I am also wondering how I can activate a camming profile successfully while the piece is essentially clamped in both machines. I feel like there will be a lot of stress from the feeder machine trying to push the material while it is clamped in the mill.
Current camming profile attached, too.

Thanks in advance.

Capture.JPG
 
The accel & decel are smooth, but the linear phase is jerky. I can only assume this has to do with the encoder positioning. Any way to help alleviate this without swapping encoders? I will try to get a trend. I just got to the site to do some more testing.
 
For testing you can create an imaginary axis that runs at steady speed and try camming to it instead of the encoder. If it is still jerky, then your servo is the issue - not the camming to the encoder. If it is smooth, then the encoder is suspect.
 
I am doing a somewhat similar thing with a very similar cam to what you show.

I am using an "AB 842HR sine cosine serial encoder - size 25 High Resolution Auxiliary Servo Feedback". It feeds into a Kinetix 7000 auxiliary port.

Mine seems to be steady, but my process is probably more forgiving than yours, so I don't know that I would notice a little jerkiness.

How are you getting the encoder - via aux port or from an ethernet encoder?
 
This is from a wired encoder to the Aux port of the Servo drive. When I trend position vs. time, it "appears" to be linear, but I have only trended it in FTView, which has a limit of 50ms polling. I have recommended to the customer to invest in a high-resolution ethernet encoder. We'll see.
 
I forget why right now, but I think you will struggle to cam a servo to an ethernet encoder. I know I tried and couldn't do it but I don't remember the details. I think the RSL software just wouldn't let me cam to it??? (sucks to get old and forget stuff). Aux port is the way to go. Not sure if the sine/cosine encoder would help (suspect it might).
 
So far running the cam with a master axis of a virtual position is performing very smooth. Doesn't look good for the encoder. So how did it perform better when the Yaskawa system was in place? Was it less picky about noise?
 
Here's a Logix 5000 sample of position vs. time.
It looks really smooth, despite the actualvelocity bouncing around as much as it does.

Capture.JPG
 
Here's a capture of the Encoder.ActualVelocity
Look at the way the velocity changes in discrete levels.
It look like the encoder is changing by 2 counts most of the time with sometimes changing by 1 count and other times changing by 3 times. This is not enough resolution to gear to.

Also, the cam profile issue seems to be ignored. The cam profile shown looks like the acceleration rate is constant during the ramp. It is impossible to change the acceleration instantly as show in the cam profile.

Ditto the comments about gearing to an Ethernet encoder.
 
I have had similar plots with an AB servo and ethernet encoder. After a few days I finally figured out that the feed forwards in the axis setup were at 100%. I didn't put them there, it happens when tuning the loop. Once I lowered them down, it settled down. A word of caution, the axis tuning section has a slider for adjusting the loop bandwidth. When you move that slider, it puts the feed forwards back to 100%.

Also, I know in the setup of the ethernet encoder, you can filter out some noise which cleans up the master signal, but if you apply too much the resolution is not real good.
 
Last edited:

Similar Topics

Hi all, Has anyone found a quick way to patch in the functionality of the Windows help files in Windows 10 so it can be utilized in Logix 5000?
Replies
0
Views
594
Hello Everyone, I have tried to access the help function from my RSLogix 5000 but everytime I select a function block on the PLC rung and press...
Replies
3
Views
1,222
I’m working with indirect addressing and trying to make an HMI where each Rectangle Object has a tag of “A[0-127]” with a correlating Push Button...
Replies
15
Views
4,142
I have been tasked to upgrade some SquareD PLCs at our plant. I have done several of these upgrades without any issues. However, I have now met my...
Replies
4
Views
2,407
Hi There, When i did FRD instruction in SLC 500 its working, when i try to simulate in RSLogix 5000 same way as i did in SLC 500 which is not...
Replies
14
Views
2,651
Back
Top Bottom