Sign in to follow this  
Followers 0
Rich M

MO2AE Gearing

34 posts in this topic

I need some HELP.. First I will explain what I'm trying to do. Hope nobody gets bored. This is for a planermill that trims the boards prior to sorting them. The boards get loaded into lugs via a lug loader upsteam. We will call this section of chain the Cut-N-Two Chains which has a powerflex 700 on it. The other motor is the Sorter Drive, another Powerflex 700. In normal operation these two drives travel at the same speed (1 to 1). They both have encoders on them. 0 - 4000 counts per revolution. The max speed they travel is 135 RPM. That's on the output of the gearbox which the lugged chains are attached. What we need to do is create a gap or empty lug, we have to slow down the Cut-N-Two chains and then catch back up creating a single empty lug. The reason for that is once the board gets handed off to the Sorter drive it goes through a trimsaw where we can cut the board in half. The empty lug will be behind this board and after we cut it we will raise up some liftskids and set the other half into the empty lug so we can sort it into bins. I have on my test bench two motors with encoders and two drives. I have them wired up to a MO2AE. I've got to the points were I can gear 1 to 1 and it does a good job. It's the slowing down to create that empty lug I'm having problems with. Any help would be appricated… Thanks everyone. Richard

Share this post


Link to post
Share on other sites
I presume not every board gets cut in half, but some action selects a board to be cut. You can use that action to trigger either: a change in the gear ratio in the MAG instruction. Note that the MAG instruction is transitional, which means it only operates on a false-to-true transition on its input. Therefore, if you use a variable for the gear ratio, you can alter the value while gearing is in progress and control when to use the new ratio by transitioning the rung with the MAG. You will probably want some clutch action.a reverse relative MAM on the slave axis while maintaining the 1:1 gearing. Note that all motion instructions are additive, so the sum of the reverse MAM and the MAG will be the command to the slave. This method offers separate accel & decel rates as well as profile and jerk settings for the speed change vs the single setting for the MAG clutch action. Also, at the end of the move, the slave is back to 1:1 with the master - no need to calculate / time when to change the ratio in the MAG back to 1:1. Edited by Gerry

Share this post


Link to post
Share on other sites
On further consideration, I think a better solution would be to superimpose a position cam on the gearing. This would automatically compensate for speed variations in the master during the adjustment and also make it easy to keep the chains in sync. Gerry
1 person likes this

Share this post


Link to post
Share on other sites
Definitely this is the way to go Step 1 is a MAG that locks the two chains together The MAPC that you call depends on how the product handles overlaps The easiest is a cam that moves the master axis 2 lugs while the slave moves 1 lug Harder is one that stops the slave movement waits then restarts at the right time - all depends on accel / decel rate that the slave axis can cope with Easiest Cam 0 0 Cubic 0.5 -0.083333333 Cubic 1 -0.5 Cubic 1.5 -0.916666667 Cubic 2 -1 Cubic Reason for the -1 is that you are slowing down slave which has a 1:1 gear

Share this post


Link to post
Share on other sites
The product can't handle overlaps. Needs to come back into position and on target...Tracking the master. I'm trying your suggestions but having some trouble. First the MAG instruction works when they are locked together. What happens is when we turn the vfd off and one coasts more than the other, when we start, they are not locked at the same encoder position. This can be a real problem becasue if the operator hits an estop, we take the command away from the drive. It's a safety issue. When we start back up, the slave needs to get back and lock up with the master. Any idea's to help with that portion? Thanks,

Share this post


Link to post
Share on other sites
How do you get the chains in sync now? You will need to establish a reference (home) position for each chain - preferably a prox to sense a lug. Actual home position is irrelevant - you just need to know where it is relative to some critical point (i.e. where you want to initiate the slow-down). On start-up, home each axis and move one as / if necessary to achieve sync. If you always maintain power to the MO2 and the encoders, the system will know the positions even when the drives are off (axis disabled) so on restart after a stoppage only a move of one axis should be required to re-establish sync. I would set the chains up as rotary axes with an unwind count equal to the lug spacing. BEWARE: if the unwind count is not an integer, you are headed for trouble.

Share this post


Link to post
Share on other sites
That profile actually does stop the slave. I understand the point of softening the speed change and controlling acceleration, but I suspect the powerflex drives wouldn't notice the jerk at each end of a simple 2-point cubic profile which would only slow the slave to minimum 25% of master speed.

Share this post


Link to post
Share on other sites
Two Point Cubic Instant change in demanded accel at profile start and stop - It all depends on the product and how stiff the system is and how close you want the system to follow - I am use to not having following errors when I start or finish a profile so I can have higher gains Agreed the Power flex probably will not see it the MO2AE position control will - Try trending the Position error bumps at start and stop of profile I did not look to see if it stops the axis. - Worse is if it tries to go backwards ok it just stops the axis for a moment Edited by michael G

Share this post


Link to post
Share on other sites
I put them in sync by looking at each encoder and determining how far off one is from the other. Very simple. What I need to do is say the master is stopped and it's position is 2000 (out of 4000) and the slave is 1000 (out of 4000) when I start I need the slave to catch up to the master is at (2000 and climbing) and vise versa. If the master is 2000 and the slave is 2500, then the slave will wait for the master to reach 2500 then follow. Thats what I really need some help on. I can't seem to get passed this point. Trust me, I'm trying....Hey, wanted to say thanks to all that has posted. It's nice to have someplace to be able to ask questions and get responses....I reallly appricate it guys. Richard

Share this post


Link to post
Share on other sites
Am I missing something here? The MO2AE takes position feedback from an incremental encoder, yet your description above implies that you have absolute position feedback. On initial power-up of the system, how do you know where the lugs are? On power-up the MO2AE will report both positions as zero, regardless of reality. You need to define actual positions before you can do anything useful - usually by a homing sequence. Once homed, the system will track position in real terms as long as power is on. Once you're dealing with real positions, to get the chains in sync simply command an appropriate move of one or the other before enabling gearing.

Share this post


Link to post
Share on other sites
Yes, since this isn't an actual servo system, I wouldn't expect the system to be very stiff at all. Although the profile only touches zero velocity, if there is following error would you anticipate a possibility of actual output going negative?

Share this post


Link to post
Share on other sites
We have absolute position feedback. If the system ever gets turned off, which it hardly ever does but if we loose power, I record the actual position it's at and then when the power comes back on I write the last known position. This isn't a problem and has worked great for some time. See my last post. That's what I really need help with. Getting the MO2AE to Sync the slave to the master after the operator hits an estop. Because when they hit an estop, we take the command signal away from the drives and they stop. Never do they stop at the same encoder count. Sometimes it's only off by 300 counts. When they start they system, the slave has to catch up or wait depending on if the slave is behind or ahead of the master. We never reverse the drive. If the slave is ahead of the master, the slave needs to pause until the mater catches up then follow. or vise versa. There isn't a problem with feedback ever. Very solid. I just need to know how to get the slave to catch up to the master if it's behind on startup....

Share this post


Link to post
Share on other sites
A number of methods which one you use depends on the results you want Simplest is advancing the behind (lagging) axis using MAM to meet the axis in front Then engaging MAG and starting the machine Harder is if the slave axis is in front and you must start the master axis and the slave axis must merge to the moving master Pointer is to use the MAPC waiting for the Master axis to move to the correct position Then have calculated a CAM that accelerates the slave to merge into the correct ratio

Share this post


Link to post
Share on other sites
Leaving aside the method of feedback.... If you know the distance by which the axes are out of sync, why can't you simply do as Michael and I have suggested and move one of the axes by the necessary amount before engaging gearing and starting the machine?

Share this post


Link to post
Share on other sites
That's what I did today and it works great. The slave moves right into position and then they both start. I do notice that they stop for a brief period and it's a littel jerky but I'm sure I can get that smothed out. When I do the MCCP then the MAPC to issue the slow down for the board where the master moves 2 revolutions and the slave move one., Could you go into a little detail about what I should start with. What I would like is a smoth curve not jerky..... Sorry guy's that I'm not all that up to speed on all this but I assure you, you guy's help more than you realize....I appricate it a lot. I'll send you guy's a video of what we are doing and how it looks when were done.

Share this post


Link to post
Share on other sites
I don't know if this helps, but here's a trick with the MAPC command that I have used. There are two parameters that you probably have set to "1": Master Scaling and Slave Scaling. If your camming motion is always the same, just at differents speeds/directions with relation to each other, then you can use the scaling instead of the MCCP command. What I have done is set up a basic zero acceleration cam: Easiest Cam 0 0 Cubic 0.25 0.083333333 Cubic 0.5 0.5 Cubic 0.75 0.916666667 Cubic 1 1 Cubic Then set the Master Scaling equal to the number of units of master distance that I want the cam to operate over. For example, if my units is inches, and I want the slave cam to operate over a distance of 12 inches on the master, set the Master Scaling to 12. Next, I calculate on the fly the distance that I want the slave to move during the cam, and use this as my Slave Scaling. This number can be negative, which results in a slow down or even a reverse move on the slave. So, if I want the slave to retard by six inches during the a master move of 12 inches, then I set my Slave Scaling to -6. It's just another approach, and one that I personally find easier than constantly recalculating the cam itself. It won't work for all situations, but might simplify things for you.

Share this post


Link to post
Share on other sites
To smooth things out, in the MAG and MAPC specify command position rather than actual position for the master reference. This will remove effects of quantization errors and mechanical noise from the encoders. Note that you should only need to execute the MCCP once. What profile are you using? The one Michael suggested in #4 is about as smooth as you will get. The 2-point one I suggested is smooth except for abrupt transitions at beginning and end. I'm assuming that you're using standard induction motors with the powerflex drives and I'm guessing that they would be unable to respond to the discontinuities in acceleration at each end of the 2-point profile. The 5-point profile eliminates the discontinuities but the magnitude of acceleration and the peak velocity are greater than those of the 2-point profile. I would appricate appreciate seeing the video.

Share this post


Link to post
Share on other sites
I used the 5 point profile. Works pretty good. Here is a video of it. What I ended up with is getting the two in sync with a MAM, then gearing them together 1- 1. Then when it's time for the slow down, I used the profile. I had to put the Exacution Mode to once and the Schedule to immidiate. I have to be able to move the poijnt that it enters at so I couldn't use pending. Everything is working really well. The one problem I have is when I need to do more than one lug in a row. If I put the Mode to continuous, it just stays in the cam. Sometimes we need to do more than one in a row. Sometimes it may be 20. So that's the only problem left right now. This looks like it is going to work. After I get it to do several in a row, I'll send a video of it actually working and you will be able to see it seperate the boards. Any Idea's?... Here is the video..... Master Slave Moves.wmv

Share this post


Link to post
Share on other sites
Two ways to do it 1) Simple way to get it to change board lengths Add a pending CAM while the add slowdown is occurring Pending CAM is 0,0 linear 20,0 linear for 20 lugs (Note relies on 1:1 MAG in progress) While the board length CAM is running redo the slowdown cam as a pending CAM Repeat 2) Another way is to perform the CAM you have but change the MAPC to use Master Start position (cannot remember the exact mode but it is in the same location as immediate) Edited by michael G

Share this post


Link to post
Share on other sites
I hope I don't sound like I'm beating a drum... If I understand you, this is an even better place to use the Slave Scaling parameter. Assuming you did not use my suggestion above, then you have a cam that is set up to retard one lug exactly. Say you need to retard 20 lugs, just change the Slave Scaling from 1 to 20.

Share this post


Link to post
Share on other sites
What he wants to do is effectively run the slave at half speed for some variable number of boards (= lugs on slave). Scaling won't achieve that.

Share this post


Link to post
Share on other sites
The specification expands.....any more items to come? Here's my suggestion: You'll need 2 profiles - the one you already have plus a "do nothing" profile - 2 points [0,0]; [1,0] You will need 3 MAPC instructions: This uses the "do nothing" profile and is enabled at start-up at the same time as gearing. Execution mode is continuous, schedule immediate or forward only to use a master lock position.This uses the retard profile and is triggered when one or more boards are to be retarded. Ex mode is continuous, schedule pending.This is the same as number 1 except the schedule is pending. Trigger this during the last slow-down.Ideally, you would just run the slave at half speed for the required time, but that raises the complication of synchronising. Using cams keeps everything in sync and produces an average speed equal to half that of the master, but the repeated cycling may look odd to an observer.

Share this post


Link to post
Share on other sites
+1 and use scaling on the scaling on the "do nothing" profile for the board length

Share this post


Link to post
Share on other sites
No offense, but this talk of scaling is confusing the issue. Board length is not a consideration except for making a decision to cut it in two - which requires the upstream (slave) chain to retard, creating a gap on the master. The boards do not travel length-wise, rather they lay across the chains, lugs pushing the edge.

Share this post


Link to post
Share on other sites
This works and I can do as few or as many as needed. The only problem is like you said, "the repeated cycling may look odd" which it does and we wouldn't want the chains speeding up and slowing down if we had more than one cut-n-two. Not changing the specs either, just need it to slow down and then stay that way until we need to track 1 - 1 with the master. ...I know you may be getting tired of all my problems but it's darn close. Any idea's. Thanks guy's

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0