StartAble1-13379.c

Hi David,

It looks like your P I D parameters are all zero

Regards

TK

 

Hi Tom,

I'm just finishing a new 2 axis machine build which uses 3 phase brushless servos and KFlop/SnapAmp.

For some reason I am not getting any motion when trying to move either axis and the Analog screen in KMotion shows no activity for the PWMs.

The motors turn during the initialization sequence (using the Write3PH () command), but not when trying to jog, move or tune. Autophasefindfast.c seems okay too.

Init file is attached. Any suggestions?

Thank you,
David.

 

Hi David,

It looks like your P I D parameters are all zero

Regards

TK

On 6/23/2016 11:30 AM, david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

I'm just finishing a new 2 axis machine build which uses 3 phase brushless servos and KFlop/SnapAmp.

For some reason I am not getting any motion when trying to move either axis and the Analog screen in KMotion shows no activity for the PWMs.

The motors turn during the initialization sequence (using the Write3PH () command), but not when trying to jog, move or tune. Autophasefindfast.c seems okay too.

Init file is attached. Any suggestions?

Thank you,
David.

Virus-free. www.avast.com

Hi David,

That shouldn't be a problem.  The Autophase.s should determine

 

Hi Tom,
Thank you.

Something else I just realized is the servo motors are 3 phase brushless, but they have 10 poles as opposed to the usual ones I use with 4 poles. Can the Snapamp manage these okay?

Regards,
David.

On 6/23/2016 5:32 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

It looks like your P I D parameters are all zero

Regards

TK

On 6/23/2016 11:30 AM, david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

I'm just finishing a new 2 axis machine build which uses 3 phase brushless servos and KFlop/SnapAmp.

For some reason I am not getting any motion when trying to move either axis and the Analog screen in KMotion shows no activity for the PWMs.

The motors turn during the initialization sequence (using the Write3PH () command), but not when trying to jog, move or tune. Autophasefindfast.c seems okay too.

Init file is attached. Any suggestions?

Thank you,
David.

Virus-free. www.avast.com

 

Hi David,

That shouldn't be a problem.  The Autophase.s should determine

the correct number of counts per cycle regardless of the number of cycles (poles/2) per rev.

Regards
TK

On 6/23/2016 5:17 PM, David Stevenson david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,
Thank you.

Something else I just realized is the servo motors are 3 phase brushless, but they have 10 poles as opposed to the usual ones I use with 4 poles. Can the Snapamp manage these okay?

Regards,
David.

On 6/23/2016 5:32 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

It looks like your P I D parameters are all zero

Regards

TK

On 6/23/2016 11:30 AM, david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

I'm just finishing a new 2 axis machine build which uses 3 phase brushless servos and KFlop/SnapAmp.

For some reason I am not getting any motion when trying to move either axis and the Analog screen in KMotion shows no activity for the PWMs.

The motors turn during the initialization sequence (using the Write3PH () command), but not when trying to jog, move or tune. Autophasefindfast.c seems okay too.

Init file is attached. Any suggestions?

Thank you,
David.

Virus-free. www.avast.com

Virus-free. www.avast.com

> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

StartAble1-13388(1).c

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

StartAble1-13391(1).c
AutoPhaseFindFast-13391(1).c

 

Hi David,

That would indicate incorrect Commutation. 

Post your AutoPhaseFind.c program, the results, and your initialization program. 

Regards
TK

On Jun 24, 2016, at 4:58 PM, David Stevenson david.m.stevenson@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Virus-free. www.avast.com

AutoPhaseFindFastRev2-13392.c

Hi David,

The AutoPhaseFind should rotate until it finds the index twice then reverses until it finds the index twice.

If things are working such that the motor is rotating based on the phase angle, and properly detecting the Index pulse, then the reported positions where the Index pulse was found should be equally spaced and at  similar fractions of phase angle.  Yours are not. 

Do you know how many counts/rev your encoder should be?

Have you tested the encoder?  If you rotate one rev how many counts does the position change?  If you rotate back does the position return to the original value?

Most likely noise on the Index pulse, encoder not working, or index pulse being too short and being missed.

Attached is your AutoPhaseFind Program modified so it should rotate 4X slower.

HTH

Regards

TK

 

Hi Tom,

Here is the report for axis0:

REPORT
------
0 Position =  -1701 PhaseAngle = 0.583900
1 Position =  -3390 PhaseAngle = 1.097700
2 Position =  -1792 PhaseAngle = 0.581100
3 Position =  11310 PhaseAngle = -3.418900
Counts per rev =  -1689
Counts per cycle =  -3287
Counts per cycle (rounded)=  -3290
invDistPerCycle (rounded)= -0.000303951368
Commutation offset =  -1939
Input Gain Specified =  1.000

and axis1:

REPORT
------
0 Position =  -4998 PhaseAngle = 1.098700
1 Position = -21382 PhaseAngle = 6.098700
2 Position =  -4998 PhaseAngle = 1.071700
3 Position =  11386 PhaseAngle = -3.928300
Counts per rev = -16384
Counts per cycle =  -3277
Counts per cycle (rounded)=  -3280
invDistPerCycle (rounded)= -0.000304878049
Commutation offset =  -1099
Input Gain Specified =  1.000

The init file and autophasefind file are attached.

Thanks for your help,
David.

On 6/24/2016 9:19 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

That would indicate incorrect Commutation. 

Post your AutoPhaseFind.c program, the results, and your initialization program. 

Regards
TK

On Jun 24, 2016, at 4:58 PM, David Stevenson david.m.stevenson@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Virus-free. www.avast.com

 

Hi David,

The AutoPhaseFind should rotate until it finds the index twice then reverses until it finds the index twice.

If things are working such that the motor is rotating based on the phase angle, and properly detecting the Index pulse, then the reported positions where the Index pulse was found should be equally spaced and at  similar fractions of phase angle.  Yours are not. 

Do you know how many counts/rev your encoder should be?

Have you tested the encoder?  If you rotate one rev how many counts does the position change?  If you rotate back does the position return to the original value?

Most likely noise on the Index pulse, encoder not working, or index pulse being too short and being missed.

Attached is your AutoPhaseFind Program modified so it should rotate 4X slower.

HTH

Regards

TK

On 6/24/2016 7:34 PM, David Stevenson david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

Here is the report for axis0:

REPORT
------
0 Position =  -1701 PhaseAngle = 0.583900
1 Position =  -3390 PhaseAngle = 1.097700
2 Position =  -1792 PhaseAngle = 0.581100
3 Position =  11310 PhaseAngle = -3.418900
Counts per rev =  -1689
Counts per cycle =  -3287
Counts per cycle (rounded)=  -3290
invDistPerCycle (rounded)= -0.000303951368
Commutation offset =  -1939
Input Gain Specified =  1.000

and axis1:

REPORT
------
0 Position =  -4998 PhaseAngle = 1.098700
1 Position = -21382 PhaseAngle = 6.098700
2 Position =  -4998 PhaseAngle = 1.071700
3 Position =  11386 PhaseAngle = -3.928300
Counts per rev = -16384
Counts per cycle =  -3277
Counts per cycle (rounded)=  -3280
invDistPerCycle (rounded)= -0.000304878049
Commutation offset =  -1099
Input Gain Specified =  1.000

The init file and autophasefind file are attached.

Thanks for your help,
David.

On 6/24/2016 9:19 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

That would indicate incorrect Commutation. 

Post your AutoPhaseFind.c program, the results, and your initialization program. 

Regards
TK

On Jun 24, 2016, at 4:58 PM, David Stevenson david.m.stevenson@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Virus-free. www.avast.com

Virus-free. www.avast.com

Emerson Servo 089EDB400BACAA-13395.pdf

 

Hi David,

The AutoPhaseFind should rotate until it finds the index twice then reverses until it finds the index twice.

If things are working such that the motor is rotating based on the phase angle, and properly detecting the Index pulse, then the reported positions where the Index pulse was found should be equally spaced and at  similar fractions of phase angle.  Yours are not. 

Do you know how many counts/rev your encoder should be?

Have you tested the encoder?  If you rotate one rev how many counts does the position change?  If you rotate back does the position return to the original value?

Most likely noise on the Index pulse, encoder not working, or index pulse being too short and being missed.

Attached is your AutoPhaseFind Program modified so it should rotate 4X slower.

HTH

Regards

TK

On 6/24/2016 7:34 PM, David Stevenson david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

Here is the report for axis0:

REPORT
------
0 Position =  -1701 PhaseAngle = 0.583900
1 Position =  -3390 PhaseAngle = 1.097700
2 Position =  -1792 PhaseAngle = 0.581100
3 Position =  11310 PhaseAngle = -3.418900
Counts per rev =  -1689
Counts per cycle =  -3287
Counts per cycle (rounded)=  -3290
invDistPerCycle (rounded)= -0.000303951368
Commutation offset =  -1939
Input Gain Specified =  1.000

and axis1:

REPORT
------
0 Position =  -4998 PhaseAngle = 1.098700
1 Position = -21382 PhaseAngle = 6.098700
2 Position =  -4998 PhaseAngle = 1.071700
3 Position =  11386 PhaseAngle = -3.928300
Counts per rev = -16384
Counts per cycle =  -3277
Counts per cycle (rounded)=  -3280
invDistPerCycle (rounded)= -0.000304878049
Commutation offset =  -1099
Input Gain Specified =  1.000

The init file and autophasefind file are attached.

Thanks for your help,
David.

On 6/24/2016 9:19 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

That would indicate incorrect Commutation. 

Post your AutoPhaseFind.c program, the results, and your initialization program. 

Regards
TK

On Jun 24, 2016, at 4:58 PM, David Stevenson david.m.stevenson@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Virus-free. www.avast.com

Virus-free. www.avast.com

Hi David,

I don't see anything problematic.

Based on the back-emf spec of 42.75V/KRPM I would expect you to be able to get around 1500RPM with an 80V supply.

Regards

TK

 

Hi Tom,

The commutation seems to be working okay now, but I still can't move the motors normally.

Will you please have a look at the attached specs for the servos and let me know if anything obvious jumps out at you as to why they are not moving properly? I know the SnapAmp voltage is only 80, but my understanding is this will just limit the max speed of the motors.

Thank you,
David.

On 6/25/2016 1:03 AM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

The AutoPhaseFind should rotate until it finds the index twice then reverses until it finds the index twice.

If things are working such that the motor is rotating based on the phase angle, and properly detecting the Index pulse, then the reported positions where the Index pulse was found should be equally spaced and at  similar fractions of phase angle.  Yours are not. 

Do you know how many counts/rev your encoder should be?

Have you tested the encoder?  If you rotate one rev how many counts does the position change?  If you rotate back does the position return to the original value?

Most likely noise on the Index pulse, encoder not working, or index pulse being too short and being missed.

Attached is your AutoPhaseFind Program modified so it should rotate 4X slower.

HTH

Regards

TK

On 6/24/2016 7:34 PM, David Stevenson david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

Here is the report for axis0:

REPORT
------
0 Position =  -1701 PhaseAngle = 0.583900
1 Position =  -3390 PhaseAngle = 1.097700
2 Position =  -1792 PhaseAngle = 0.581100
3 Position =  11310 PhaseAngle = -3.418900
Counts per rev =  -1689
Counts per cycle =  -3287
Counts per cycle (rounded)=  -3290
invDistPerCycle (rounded)= -0.000303951368
Commutation offset =  -1939
Input Gain Specified =  1.000

and axis1:

REPORT
------
0 Position =  -4998 PhaseAngle = 1.098700
1 Position = -21382 PhaseAngle = 6.098700
2 Position =  -4998 PhaseAngle = 1.071700
3 Position =  11386 PhaseAngle = -3.928300
Counts per rev = -16384
Counts per cycle =  -3277
Counts per cycle (rounded)=  -3280
invDistPerCycle (rounded)= -0.000304878049
Commutation offset =  -1099
Input Gain Specified =  1.000

The init file and autophasefind file are attached.

Thanks for your help,
David.

On 6/24/2016 9:19 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

That would indicate incorrect Commutation. 

Post your AutoPhaseFind.c program, the results, and your initialization program. 

Regards
TK

On Jun 24, 2016, at 4:58 PM, David Stevenson david.m.stevenson@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Virus-free. www.avast.com

Virus-free. www.avast.com

 

Hi David,

I don't see anything problematic.

Based on the back-emf spec of 42.75V/KRPM I would expect you to be able to get around 1500RPM with an 80V supply.

Regards

TK

On 6/26/2016 8:45 AM, David Stevenson david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

The commutation seems to be working okay now, but I still can't move the motors normally.

Will you please have a look at the attached specs for the servos and let me know if anything obvious jumps out at you as to why they are not moving properly? I know the SnapAmp voltage is only 80, but my understanding is this will just limit the max speed of the motors.

Thank you,
David.

On 6/25/2016 1:03 AM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

The AutoPhaseFind should rotate until it finds the index twice then reverses until it finds the index twice.

If things are working such that the motor is rotating based on the phase angle, and properly detecting the Index pulse, then the reported positions where the Index pulse was found should be equally spaced and at  similar fractions of phase angle.  Yours are not. 

Do you know how many counts/rev your encoder should be?

Have you tested the encoder?  If you rotate one rev how many counts does the position change?  If you rotate back does the position return to the original value?

Most likely noise on the Index pulse, encoder not working, or index pulse being too short and being missed.

Attached is your AutoPhaseFind Program modified so it should rotate 4X slower.

HTH

Regards

TK

On 6/24/2016 7:34 PM, David Stevenson david.m.stevenson@... [DynoMotion] wrote:

 

Hi Tom,

Here is the report for axis0:

REPORT
------
0 Position =  -1701 PhaseAngle = 0.583900
1 Position =  -3390 PhaseAngle = 1.097700
2 Position =  -1792 PhaseAngle = 0.581100
3 Position =  11310 PhaseAngle = -3.418900
Counts per rev =  -1689
Counts per cycle =  -3287
Counts per cycle (rounded)=  -3290
invDistPerCycle (rounded)= -0.000303951368
Commutation offset =  -1939
Input Gain Specified =  1.000

and axis1:

REPORT
------
0 Position =  -4998 PhaseAngle = 1.098700
1 Position = -21382 PhaseAngle = 6.098700
2 Position =  -4998 PhaseAngle = 1.071700
3 Position =  11386 PhaseAngle = -3.928300
Counts per rev = -16384
Counts per cycle =  -3277
Counts per cycle (rounded)=  -3280
invDistPerCycle (rounded)= -0.000304878049
Commutation offset =  -1099
Input Gain Specified =  1.000

The init file and autophasefind file are attached.

Thanks for your help,
David.

On 6/24/2016 9:19 PM, Tom Kerekes tk@... [DynoMotion] wrote:

 

Hi David,

That would indicate incorrect Commutation. 

Post your AutoPhaseFind.c program, the results, and your initialization program. 

Regards
TK

On Jun 24, 2016, at 4:58 PM, David Stevenson david.m.stevenson@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:

 

Actually after bumping the P to 0.1 from 0.01 the motors are definitely servoing. Both axis though will allow a small (~20 degrees) rotation in one direction and then they lock up tight.

David

On 6/24/2016 7:31 PM, David Stevenson wrote:

Hi Tom,

After running the init file (attached) I don't think the commutation is working. There is slight resistance to turning, but not as much as I would expect. I re-ran the autophasefind.c to confirm the offsets, but that didn't make any noticable difference. The axis do show as enabled in the axis screen window.

Thank you,
David.

On 6/24/2016 5:53 PM, tk@... [DynoMotion] wrote:

 

Hi David,

There should have been no hardware changes from our end.

For a servo to work the commutation has to work correctly. Then the
Servo should hold position. Then you can begin tuning. Have you
verified any of the steps in order to determine what is going on.

For example you might enable the servo with:

#1 limited output to a small amount.
#2 infinite following error
#3 Only a small P Gain
#4 phase find and enable the servo

In this case the motor should act like a spring. The motor should hold
position and as it is forced away from target the torque should increase
up to the set limit. Then it should remain constant as the motor is
moved an additional few revolutions. If this works then most likely the
commutation is being performed correctly.

HTH
Regards

On 2016-06-24 13:43, David Stevenson david.m.stevenson@...
[DynoMotion] wrote:
> Thanks Tom.
>
> I switched out the servos to the old style with the 4 poles and there
> is no way I can tune this setup. Today is the 3rd straight day for
> this and I'm no closer now than I was 3 days ago. I've read at least
> 10 different how-to-set-your-PID articles but the process doesn't seem
> to work here.
>
> This is a brand new KFlop/SnapAmp and I have gone through the wiring
> at least 5 times. This is machine #10 with this design, but something
> is not right with this one. I even tried going back to an earlier
> software version.
>
> Have there been any fundamental changes to the hardware that might
> need to be considered? All the power supplies check out okay and I
> have replaced the ribbon cables just to be sure.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
> David.
>

Virus-free. www.avast.com

Virus-free. www.avast.com

Virus-free. www.avast.com

Virus-free. www.avast.com