[87kevin]

Hi guys,

Just working on retrofitting an old mill with Mach3 and servos/drives from Machmotion. Got everything up and running great including x/y/x axis. My problem is with my A axis which is a 400w ac servo motor on a ATC carosel. It is a 20 tool changer, so it's pretty big. Carosel is gear driven by the motor at a reduction of 8-1.

Problem is that as soon as I try any movement on that axis from Mach3, no matter how big or small, slow or fast, it will move then start shaking(rotating) back and forth. I can see the numbers on the drive going up and down, while the M3 DRO stays still. It will occasionally settle itself down then stop, but usually it will escallate (rotate further and further) in each direction untill I get a motor overload error. Sometimes instead of going back and forth, it just runs away in one direction.

I've bench tested it, and it is fine. Problem occurs when mechanically joined to the load. Advice from machmotion was to change the frequency filter parameter on the drive. This didnt really work, I was at one pt able to stop the jumping, but then the jogging was very jumpy and not smooth.

There is hundreds of parameters on the servo drive to change but I'm new to this and have no clue where to start.

Any ideas/suggestions? Swapping drives/motors wont work since they are different sizes. These were all brand new never used.

Thanks!

-Kevin

[mike_Kilroy]

A axis motor prt no?
A axis drive prt no?

machmotion sells different brands and have sold more different in past too; no clue how to help w/o right data.

[87kevin]

400w TECO ac servo as found here:

TECO 400W AC Servo Drive and Motor

tsta-20c drive
TSC06401C motor

Thanks!

[mike_Kilroy]

Kevin,

It is my belief that if you call machmotion.com and speak to their tech folks (Andy, Dave, or Joe) they will tell you the TWO or THREE parameters that may need tweaking to solve your issue.

The process will be simple: you take ONE parameter at a time and make it bigger and smaller and watch results: if better, keep going in that direction until it gets crappy then back off till good again. Then do next parameter and next. when done you should have a decent system.

The parameters are called velocity loop proportional gain, velocity loop integral gain, and position loop proportional gain.

good luck

[87kevin]

OK Mike, thanks for your help. I will try that tonight.

I just heard back from Machmotion and they think the problem is too much inertia on the motor and they suggest a gearbox.

Anyone want to take a look at the setup and tell me if you think I need more gear reduction?

[mike_Kilroy]

sorry to hear that reply (:

If pix u posted is the motor with integral 8:1 gearhead into 3"dia pinion onto gear turning the huge tool changer, then they are ultimately probably theoretically correct that you should have more ratio to make it even better.

but that is not the right answer for you today. the right answer for you today is to adjust the proper velocity loop gain parameters to calm it down. let me say it another way: there is NO system you can build - even really really wrong - that you cannot set tuning parameters to operate with out your descibed oscillation. that said, the settings may be so low and spongy that it takes way too long to make the move and settle but it can be made to operate! so. back to gains. read your servo drive manual. find the velocity loop gain parameters. adjust prop and integral terms until your oscillation is gone. response may not be good enough but it wont shake rattle and roll.

rule of thumb: when you have huge inertias, the prop gain normally goes up (more, bigger) and the culprit is 99& of the time for shakes like you say the integral term is way way too big - make it smaller - if inertia reflected is way big the intergral term will be near zero but not zero. that will stop the shakes.

the proper way to test tuning and adjust it is WITHOUT THE MACH MOTION CONTROLLING IT. disconnect the analog input command from the machmotionn and use a small 1.5v battery instead. click it on analog input to move, off th stop. adjust the velocity gains in the drive to make smooth with it. THEN hook up mach motion..... if it oscillates again, THEN call macmotion and tell em it is smooth with battery box but shakes with their control - then they will tell u which paramete in their system to fix....

now to your challenge: i would do the motor inertia sizing for you if you post all the mechanical and performance data: mech size (dia and thickness) of the pinion, its mating gear, weight of all the stuff moving, a couple more overall pix to see the whole gidget, speed you want in rpm, any accel rate desires, is it vertical or horisontal, anything else that would let calculation of inertia, mass, etc.....

[87kevin]

Mike,

I've taken your advice and played with the settings a little. I've got it to behave normalish. It seems quite slow and sluggish and it still jerks in the opposite direction after making a move sometimes but for the most part seems stable now. What I found works was to crank up the speed loop integral time parameter. I also turned up the gain but at a point the motor started making terrible screeching sounds so I backed it off till just below that point.

There is no internal gearing on that motor, I was referring to the gearing between that pinion and the large gear. I'm going to see if it will hold up to tool changes and if not, MM sells a direct bolt on gearbox I can use. Another option would be to get a 14 tooth pinion but not sure that it would make enough of a difference.

I'd love it if you'd let me know how far off base I am with my hardware in this area. The gears mating surface is 3/4" and the pinion is about 1 1/2" in diameter. More importantly, it has 18 teeth while the big gear has 144 which is where I got the 8-1 figure. The carousel is horizontal, and no way of knowing what it weighs but it is heavy. It's 23" in diameter and appears to be made from 1/2" steel with 20 tools. (Empty for the testing so far).

Thanks in advance and for your help so far.

-Kevin

[mike_Kilroy]

OK, quick calcs in Motioneering (free at Home | Kollmorgen) shows your 11#-in motor is 27,295 times LOWER inertia than your load! We try to keep load inertia within 5-10 times max larger than motor.... this means it will move very spongy as you found and as I said it would. As I said and you found out, it does not mean you have to have an oscillating system, just supper spongy. If that spongyness is not acceptable to do the job required, you need to make the load inertia look a lot smaller to the motor.

If you want snappy response you need a motor mounted gearhead. Not sure what ratios mm sells; I know ours go from 2:1 to 1,000's:1.

For a 5sec move tool to adjacent tool, your motor only gets up to 7.2rpm! you can use a 100:1 planetary gearhead to get speed up to 720rpm and load inertia down by square factor or 10,000!

Re working the Motioneering sizing, adding the 100:1 gearhead gets load inertia to only 2.75 motor inertia. THIS will likely fix your problem.

email me if you want a cc of my motioneering file to load into your free copy to play additional what-if games or ideas on available gearheads that will just bolt right onto your motor.

[87kevin]

Yikes! Guess I'm farther away than I thought. BTW - 5sec between tools is way to slow. That could take over 1.5 min to get a tool from the other side of the carousel. I was thinking more like 1sec tool to tool. I think I have it running pretty stable but spongy at 12deg/sec right now.

Thanks for doing all that figuring out for me Mike. Greatly appreciated. Once I get the pneumatics on my machine going so I can try changing tools, we will find out if it will work as-is or not. I will let you know if it doesnt work so we can talk about gearboxes.

-Kevin

[mike_Kilroy]

Yikes! Guess I'm farther away than I thought. BTW - 5sec between tools is way to slow. That could take over 1.5 min to get a tool from the other side of the carousel. I was thinking more like 1sec tool to tool. I think I have it running pretty stable but spongy at 12deg/sec right now.

Thanks for doing all that figuring out for me Mike. Greatly appreciated. Once I get the pneumatics on my machine going so I can try changing tools, we will find out if it will work as-is or not. I will let you know if it doesnt work so we can talk about gearboxes.

-Kevin

my 5 sec was based on top motor speed of only 0.9rpm (your 12 deg/sec is 2rpm). and I just let it take 1/3 of that time to accel, 1/3 to decel, so all the way around would not have been 1.5 min more like 1 min at 1rpm.... your 2rpm means 30 sec around 1 rev. close enough for comparison; I just grabed a no out of the air. anyway, like we both are saying: try it and see if you can live with the spongy response. If so, your good to go, if not, then you need the gearbox to solve it. Good luck!