Servo communtation problem?

[rainman]

Hi all,

I have a geared servo motor (incremental encoder + halls), a servo amplifier and a controller card.

The servo amplifier parameters had been configured manually to cater for the servo motor.

From the controller I am able to read the encoder counts and also the transitions of halls (ABZ) when turn the shaft.

When the amplifier is enabled, the motor is able to energized but when gives it a slight turn on the shaft deliberately, the shaft jerked to a certain angle. This happened almost everytime.

Most of the times the jog command will fail to move the motors, but at times, I'm able to jog the motor for around +-0.4 deg and it will stopped and the following error will starts to build up and the motor halts.

I'm just asking in general, what could have caused the "motor enabled->slight push->jerk" problem?

- There is no load on motor.
- PID (Only P is set to some values which gives sufficient stiffness to the motor, which I dun think its PID issue, of course I had tried various PIDs also)

- Had swopped UVW motor phase and "played" with various communtation offset angle.

- Motor poles should be correct based on motor specifications, which is 10 poles.

Thanks.

[Al_The_Man]

You may have the encoder phasing wrong, I assume it is a differential encoder, if so swap A & /A if single ended, swap A & B .
Al.

[mike_Kilroy]

Originally Posted by rainman .... I have a geared servo motor (incremental encoder + halls), a servo amplifier and a controller card. ....... I'm able to jog the motor for around +-0.4 deg and it will stopped and the following error will starts to build up and the motor halts. I'm just asking in general, what could have caused..... - Had swopped UVW motor phase and "played" with various communtation offset angle. - Motor poles should be correct based on motor specifications, which is 10 poles.

You have 2 sets of phasing to get right: 1) encoder direction as al said, & 2) halls aligned to motor phases.

if it runs and runs away each time to full speed, it can be enc phase and fixed as al says.

if it moves + or -.4 degrees and locks up then it is motor/hall phasing issue.

assuming motor poles of 10 is right, then u need to phase it. there are 6 (SIX) motor connection choices: only 1 will make the motor run right. set ur phase offset back to 0 and try all 6 one at a time to find the one that works. remember, after u get halls phased right, THEN u can phase encoder: ie., if u find a set of motor connections that makes motor run away fast instead of go .4 degrees and get stuck, it may be phased right but encoder backwards; any motor lead combo that runs away needs encoder direction reversed to see if that also is needed....

[rainman]

Originally Posted by mike_Kilroy You have 2 sets of phasing to get right: 1) encoder direction as al said, & 2) halls aligned to motor phases. if it runs and runs away each time to full speed, it can be enc phase and fixed as al says. if it moves + or -.4 degrees and locks up then it is motor/hall phasing issue. assuming motor poles of 10 is right, then u need to phase it. there are 6 (SIX) motor connection choices: only 1 will make the motor run right. set ur phase offset back to 0 and try all 6 one at a time to find the one that works. remember, after u get halls phased right, THEN u can phase encoder: ie., if u find a set of motor connections that makes motor run away fast instead of go .4 degrees and get stuck, it may be phased right but encoder backwards; any motor lead combo that runs away needs encoder direction reversed to see if that also is needed....

Thank you very much for the kind reply.

Unfortunately I tried all the 6 ways (UVW,UWV,VUW,VWU,WUV,WVU) and all give me the same result. It always enabled motor, give it a slight turn on shaft and the motor "jerk" to a certain angle and remained energized. I would have to say 99% of the time I can't even jog the motor.

The halls looks OK and encoder seemed to be reading in the correct value/direction. I have checked the wiring and they seemed alright too.

I have no chance to see the motor run away, as the fault is always current overshoot.

[Al_The_Man]

Originally Posted by rainman The halls looks OK and encoder seemed to be reading in the correct value/direction. I have checked the wiring and they seemed alright too. .

Have you tried the encoder reversal?
You mention you have a controller card, is there a loop also closed back to the controller?
What controller card is it?
If you also have a loop closed back to the controller also via the encoder, this is where the conflict maybe (double PID loop).
What is the nature of the control signal? step/dir or analogue?
Al.

[mike_Kilroy]

so u tried all 6 and all go 1/2 degree and stuck...... u shud be doing a few basic things; turn ur peak current limit down to 1 or 2 amps so u stop faulting. be from Missouri: prove u have 10 pole motor...(how u say? short 2 motor leads and rotate shaft - count the humps; 10 pole = 10 humps.

enc right one? halls on it (comm channels) or in motor?

turn off pos loop and vel loop; make drive just current loop to simplify it until u get control.

can u set drive to just use halls for commutation only and ignore the enc until u get control? if so, do.

u can see enc count; does it count correct no for 1 motor rev (not gearhead output just motor? u gotta have right enc resolution set in drive or phase goes off quickly. does drive ask for enc counts per pole or per motor rev? 10x difference..... do u need to put in x4 on the resolution?

halls. since u can scope them, u get 10 transitions per motor mech rev?

[rainman]

Originally Posted by Al_The_Man Have you tried the encoder reversal? You mention you have a controller card, is there a loop also closed back to the controller? What controller card is it? If you also have a loop closed back to the controller also via the encoder, this is where the conflict maybe (double PID loop). What is the nature of the control signal? step/dir or analogue? Al.

I only did "reversed" the encoder by changing over the controller side by means of software parameter, unfortunately changing of A,A/,B,B/ are impossible with the right pin extracting tool.

I believed the loop is closed in controller side, there isn't an option for me to run position loop over at the controller side (this amplifier is only meant to behave like a dumb drive)

The controller is a synqnet card, not running the usual analog or step/dir signals.

Thanks.

[rainman]

Originally Posted by mike_Kilroy so u tried all 6 and all go 1/2 degree and stuck...... u shud be doing a few basic things; turn ur peak current limit down to 1 or 2 amps so u stop faulting. be from Missouri: prove u have 10 pole motor...(how u say? short 2 motor leads and rotate shaft - count the humps; 10 pole = 10 humps. enc right one? halls on it (comm channels) or in motor? turn off pos loop and vel loop; make drive just current loop to simplify it until u get control. can u set drive to just use halls for commutation only and ignore the enc until u get control? if so, do. u can see enc count; does it count correct no for 1 motor rev (not gearhead output just motor? u gotta have right enc resolution set in drive or phase goes off quickly. does drive ask for enc counts per pole or per motor rev? 10x difference..... do u need to put in x4 on the resolution? halls. since u can scope them, u get 10 transitions per motor mech rev?

Hi Mike,

Thanks. There's a very good way to confirm on the motor poles but unfortunately this is a harmonics motor with built in gear, I doubt I can feel the poles.

I had confirmed the encoder counts to be correct, as per the specification. I can read 400000 counts/rev from my controller. which is 2000ppr from the motor, gear ratio of 1:50, and quadrapture factor of 4. 2000*4*50=400000.

I think you have raised a very good point in establish the correct current loop first etc.

I'm hoping to seek opinions on what could be the possible cause of the following scenario.

1) Motor enabled, with holding torque, no fault
2) Give it a slight turn on the shaft, motor "run-away" to a small degree (seemed like skip a pole), still with holding torque no fault.
The strange thing is if I turn it CW, it will run away CW, turn CCW, it will run away CCW.

3) Try jog motor, motor either don't run at all and encounter following error, controller fault. OR motor run to a small degree and start building following error then controller fault. Amplifier also fault, over current.

Thanks.

[mike_Kilroy]

Originally Posted by rainman I'm hoping to seek opinions on what could be the possible cause of the following scenario.

100% guaranteed hall/motor phase wrong (or mismatch of halls & motor pole count - same difference)

why not tell us mfgr & part nos?

prob MEI ZMP controller, kollmorgen S200 drive, anyones guess on motor....

[Al_The_Man]

Originally Posted by rainman I believed the loop is closed in controller side, there isn't an option for me to run position loop over at the controller side (this amplifier is only meant to behave like a dumb drive) .

That appears to read to be a contradiction?

The way I solve a problem such as this is to remove the controller from the equation and apply a variable command signal manually to the drive input, this then will eliminate or confirm the drive and feedback, commutation etc from being the problem.
Al.

[mike_Kilroy]

Originally Posted by Al_The_Man That appears to read to be a contradiction? The way I solve a problem such as this is to remove the controller from the equation and apply a variable command signal manually to the drive input, this then will eliminate or confirm the drive and feedback, commutation etc from being the problem. Al.

Al, I wonder if it isn't a typo rather than contradiction?

Unfortunately, with at least our Kollmorgen Synqnet drives, w/o a synqnet connection, the drive won.t even enable so simplifying the loops may be hard for him.... synqnet typically uses a current loop drive and closes vel&pos in the controller....

AL, where IS SG? (his location)?

[Al_The_Man]

Originally Posted by mike_Kilroy AL, where IS SG? (his location)?

The ISO code for SG is Singapore?
Al.