Dynomotion

Group: DynoMotion Message: 12511 From: embraced338 Date: 11/12/2015
Subject: Axis losing steps

Hi All,


I've got a machine that appears to be losing steps after multiple moves.


It appears to only affect the X axis, but I have not thoroughly investigated this happening on other axes.


Transmission is a 25mm lead ballscrew powered by Yaskawa sigma 7 motors, through Yaskawa SGD7S-5R5A00A002 drivers in open loop step/direction mode.


This has been checked by running a test program that jogs the X axis back and forth (attached) and checking the motor rotor angle before and after the test. After the test, the motor has accumulated about 10 degrees of error (error is always positive), which in my system is equal to 555 steps of that axis (20,000 steps/rotation,). This causes the axis to be way out of position by the end of the test, and is far too much.


Attached is my initialisation file and trajectory planner settings, along with the test program used.


Also worth mentioning is that when the program is running, the commanded feedrate is 2000mm/minute. At the beginning of the test, the velocity maxes at about 1980mm/minute (this is what the actual feedrate in kmotionCNC says), but by the end of the test ends up at around 1500mm/minute.



Any assistance in solving this would be appreciated.



Thanks,


Lindsay



   @@attachment@@
Group: DynoMotion Message: 12512 From: jojodillinger Date: 11/12/2015
Subject: Re: Axis losing steps
You probably tried this, but do you get the same step loss for a lower feed rate of say 500mm/min?
Group: DynoMotion Message: 12513 From: TK Date: 11/13/2015
Subject: Re: Axis losing steps [3 Attachments]
Hi Lindsay,

One possibility is that the Yaskawa Drive isn't receiving all the Step Pulses.  How are your Step/Dir signals wired to the Yaskawa Drives?  Single ended?  How long?  Shielded?  How is the Shield Connected?  Can you reference a page in the Yakawa manual for the interface specs?


With a feedrate of 2000mm/min the Step Pulse Rate would be:
2000mm/min / 60sec/min / 25.4mm/inch x 20320 steps/inch = 26.666KHz

That shouldn't be a problem.

Some of your other Axes have too high of velocity settings.  ch2->Vel is 10MHz.  KFLOP's maximum rate is 2.5MHz.  Also with a 2us Pulse setting the pulses would all run together at 500KHz.  But with such a low acceleration setting of 30000 it would take 10000000/30000 = 333seconds to accelerate which is why you probably don't see a problem.

Regards
TK


On 11/12/2015 9:09 PM, embraced338@... [DynoMotion] wrote:
 

Hi All,


I've got a machine that appears to be losing steps after multiple moves.


It appears to only affect the X axis, but I have not thoroughly investigated this happening on other axes.


Transmission is a 25mm lead ballscrew powered by Yaskawa sigma 7 motors, through Yaskawa SGD7S-5R5A00A002 drivers in open loop step/direction mode.


This has been checked by running a test program that jogs the X axis back and forth (attached) and checking the motor rotor angle before and after the test. After the test, the motor has accumulated about 10 degrees of error (error is always positive), which in my system is equal to 555 steps of that axis (20,000 steps/rotation,). This causes the axis to be way out of position by the end of the test, and is far too much.


Attached is my initialisation file and trajectory planner settings, along with the test program used.


Also worth mentioning is that when the program is running, the commanded feedrate is 2000mm/minute. At the beginning of the test, the velocity maxes at about 1980mm/minute (this is what the actual feedrate in kmotionCNC says), but by the end of the test ends up at around 1500mm/minute.



Any assistance in solving this would be appreciated.



Thanks,


Lindsay
Group: DynoMotion Message: 12514 From: embraced338 Date: 11/13/2015
Subject: Re: Axis losing steps [3 Attachments]
I'll review the shielding and run the same test on some of my other machines that have identical drives.

Do you have any ideas about the velocity decreasing through the length of the program?


Thanks,


Lindsay
Group: DynoMotion Message: 12515 From: TK Date: 11/13/2015
Subject: Re: Axis losing steps
Hi Lindsay,

Changing the velocity and acceleration values to reasonable values and should eliminate the issue.

Its hard to explain but basically what is happening is that you are telling KFLOP that it may take up to 333 seconds to stop.  So when KFLOP sees only ~3 seconds of buffering available (maximum TP has been allowed) it considers this to be a risky situation and starts reducing feedrate.  Feedrate is reduced gradually over 333 seconds as any faster has the risk of stopping too fast. 

Regards
TK

On 11/13/2015 3:52 PM, embraced338@... [DynoMotion] wrote:
 

I'll review the shielding and run the same test on some of my other machines that have identical drives.


Do you have any ideas about the velocity decreasing through the length of the program?


Thanks,


Lindsay
Group: DynoMotion Message: 12526 From: embraced338 Date: 11/16/2015
Subject: Re: Axis losing steps
It seems the problem exists within the servo driver.

As this is the case, I think I'll connect the encoder output back into the KFLOP to close the loop, hopefully to correct the position.

Is this feasible?


Thanks,


Lindsay