EMC2 Fallowing errors, Supermax/Anilam retrofit

Printable View

10-08-2009, 04:51 AM
jvanatta

EMC2 Fallowing errors, Supermax/Anilam retrofit

I am having trouble with "fallowing errors" on my Z axis and I am running out of hair to pull out.

I have retrofitted an old Anilam Crusader M to EMC2 using the fallowing hardware configuration.
5I23 i/o card
7I33T servo interface.
Z axis
Encoder is Renco 662237-72 and is mounted on a .200" per/rev ball screw
Drive is Westamp 30060-75 velocity control with resolver feedback
Motor is SEM MT304-59
Motor sheave is 10 tooth
Ball screw sheave is 18 tooth

These are some pics of my machine.

http://i619.photobucket.com/albums/t...r/P1010688.jpg

http://i619.photobucket.com/albums/t...r/P1010687.jpg

http://i619.photobucket.com/albums/t...r/P1010686.jpg

This is my Z axis parameters file

Quote:

#Z AXIS
[AXIS_2]

TYPE = LINEAR
HOME = 0.0
MAX_VELOCITY = 2.8
MAX_ACCELERATION = 28.0
BACKLASH = 0.000
INPUT_SCALE = -2540
OUTPUT_SCALE = 1
OUTPUT_OFFSET = 0.0
MIN_LIMIT = -6.0
MAX_LIMIT = +6.0
FERROR = 50.0
MIN_FERROR = 0.01
HOME_OFFSET = 0.10
HOME_SEARCH_VEL = 0.10
HOME_LATCH_VEL = -0.01
HOME_USE_INDEX = YES
HOME_IGNORE_LIMITS = YES
MAX_OUTPUT = 0.0
# PID tuning params
DEADBAND = 0.000015
P = 10
I = 0
D = 0
FF0 = 0
FF1 = 0
FF2 = 0
BIAS = 0

The idea of this project is to use the Mesa boards with the existing linear encoders and servo motor/drives. I have discovered that to use linear encoders I will have to set up a dual feedback system,Not interested in doing that right now. So I am trying to set up my Z axis with one encoder on the the ball screw. If this works I will install encoders on the x and y axis ball screws. I prefer not to mount the encoders on the motors due to space constraints.

Please advise before I go bald!

John
10-08-2009, 04:00 PM
PCW_MESA

Quote:

Originally Posted by jvanatta

I am having trouble with "fallowing errors" on my Z axis and I am running out of hair to pull out.

I have retrofitted an old Anilam Crusader M to EMC2 using the fallowing hardware configuration.
5I23 i/o card
7I33T servo interface.
Z axis
Encoder is Renco 662237-72 and is mounted on a .200" per/rev ball screw
Drive is Westamp 30060-75 velocity control with resolver feedback
Motor is SEM MT304-59
Motor sheave is 10 tooth
Ball screw sheave is 18 tooth

These are some pics of my machine.

http://i619.photobucket.com/albums/t...r/P1010688.jpg

http://i619.photobucket.com/albums/t...r/P1010687.jpg

http://i619.photobucket.com/albums/t...r/P1010686.jpg

This is my Z axis parameters file

The idea of this project is to use the Mesa boards with the existing linear encoders and servo motor/drives. I have discovered that to use linear encoders I will have to set up a dual feedback system,Not interested in doing that right now. So I am trying to set up my Z axis with one encoder on the the ball screw. If this works I will install encoders on the x and y axis ball screws. I prefer not to mount the encoders on the motors due to space constraints.

Please advise before I go bald!

John

A following error is a very general indication that the servo system is not working or not working well (in other words the actual motor shaft position is not following the commanded position within the ferror limits. To get specific help you will have to post more detailed information about the current Z axis behavior. A verbal description would help:

Does the servo hold position (resist applied force) and move when commanded?

Do the FERRORs happen at low or high speed moves?

Does the servo system oscillate when at rest?

The best information to diagnose the problem would be a HALScope plot of commanded position and feedback position from the encoder when doing a quick "step" , that is a short move at maximum accleration.
10-08-2009, 07:07 PM
ramac

im afraid i can't help you with your problem as i am just starting to look at a mesa/emc conversion. I am hoping maybe you can help me get started though as it looks to me you have done exactly what im hoping to do, convert crusader m to emc. I see how you mounted ti 7i33, but what is the board mounted above it and where is the 5i20?
10-08-2009, 08:31 PM
jvanatta

Quote:

A following error is a very general indication that the servo system is not working or not working well (in other words the actual motor shaft position is not following the commanded position within the ferror limits. To get specific help you will have to post more detailed information about the current Z axis behavior. A verbal description would help:

Does the servo hold position (resist applied force) and move when commanded?

Do the FERRORs happen at low or high speed moves?

Does the servo system oscillate when at rest?

The best information to diagnose the problem would be a HALScope plot of commanded position and feedback position from the encoder when doing a quick "step" , that is a short move at maximum accleration.

This is the scope of the z axis moving both directions with JOG.

http://i619.photobucket.com/albums/t.../supermaxz.jpg

The motor does resist movement
Ferrors happen during commanded moves and jogs both directions.
Servo system does not oscillate at rest.
10-08-2009, 10:00 PM
jvanatta

Quote:

Originally Posted by ramac

im afraid i can't help you with your problem as i am just starting to look at a mesa/emc conversion. I am hoping maybe you can help me get started though as it looks to me you have done exactly what im hoping to do, convert crusader m to emc. I see how you mounted ti 7i33, but what is the board mounted above it and where is the 5i20?

The board above the 7I33 is a micro ATX computer.
The 5I23 is mounted in the PCI slot on the motherboard. I built a bracket to support the card. I mounted the computer and the 7I33 with stand offs from Allied electronics directly to the enclosure back board. The computer and I/O cards are powered from a basic computer power supply. Between the mother board and power supply is the MESA CFADPT10 SATA compact flash adapter. This holds my hard drive which is a 4g CF card. So far this setup is stable.

http://i619.photobucket.com/albums/t...r/P1010690.jpg
10-09-2009, 12:08 AM
ramac

I really appreciate your help getting me started. Im a machinist by trade and any electronics skills have been self taught as needed in order to to keep making parts! I am anxious to get started on this project though, it should be a great learning experience. guess Ill just order the mesa boards and download EMC and see what I can do with it. Your problem with linear encoder feedback will also be an issue for me since I have the same setup. So ill be watching for your posts. Thanks again
10-09-2009, 01:19 AM
123CNC

troubleshooting

First, I would assume the Westamp drives and motors are all original and as OEM installed, and therefore well tuned at the drive.

If this is the case, I would temporarily increase my following error limits, FERROR = 500.0
MIN_FERROR = 01.00
and hand drive the z to mid travel. Then jog or MDI the z down 0.5 (in the minus (-) direction) and see which way it actually moves. If it goes up (positive) instead, problem solved. Either rewire your encoder A/B channels, flip the analog signal to the amp, change the sign or your axis scale (-2540 to +2540), or change the direction in your hal pin connections.

If the drives and or motors are all newly added, and direction is correct, you may need to tune your amps/drives as well.

PS I like your PC flashcard setup.
10-09-2009, 04:21 AM
jvanatta

Quote:

Originally Posted by 123CNC

First, I would assume the Westamp drives and motors are all original and as OEM installed, and therefore well tuned at the drive.

If this is the case, I would temporarily increase my following error limits, FERROR = 500.0
MIN_FERROR = 01.00
and hand drive the z to mid travel. Then jog or MDI the z down 0.5 (in the minus (-) direction) and see which way it actually moves. If it goes up (positive) instead, problem solved. Either rewire your encoder A/B channels, flip the analog signal to the amp, change the sign or your axis scale (-2540 to +2540), or change the direction in your hal pin connections.

If the drives and or motors are all newly added, and direction is correct, you may need to tune your amps/drives as well.

PS I like your PC flashcard setup.

Hello 123CNC

Awesome! It works well! Thank you!

Is there a specific or general calculation for FERROR?

I set the FERROR and MIN_ERROR just like you said and it first ran away but I changed the sign on the axis scale and the z axis works like it should except it shows the positive direction as down. Any ideas an how to correct it? Change the hal pin?

I think I will need to re tune the amps and drives at some point because the motor is experiencing some instability in one direction but I am happy with the results so far. Not bad for a 20yr old drive!
I moved the z axis .100" in opposite directions and ruffly measured a backlash of .0085. I set the comp and now it's within .001". I love this stuff!

Tomorrow I order two more encoders for the x and y ball screws. Is there any reason I should not stay with 127 cont/rev encoders?

Again, thank you for the information!

John
Owasso OK
10-09-2009, 10:07 AM
123CNC

E-stop

I know I thought about it when previously posting, but failed to mention be ready with the E-stop when disabling some of the automated stop features such as following error. Its also nice to check your other safeties such as soft limits and hard limits prior to such 'seeing if she is in/out of control.'

Regarding setting FERROR and min-ferror, the EMC2 default values are generally a good starting point, and then it is basically trial and error. In Axis IIRC you can play around with this values directly (meaning it will adjust your ini file) through the axis calibration menu. Otherwise, you may want to just try some no-load (no spindle) travelling and find the threshold where your machine trips over its own limitations, manually adjusting your ini settings. Adjust slightly back from the unstable/stable value in your ini and run with it. You may need to readjust some if under load you experience to much nuisance interruption or not enough stopping before its too late. Try to error on the side of caution.

Regarding encoder counts, I'll try not to offend the disciples of 'as much as you can / more is better / or Supersize its.' A general rule of thumb in calibration is to use a reference standard at least 4 times greater than the desired accuracy. With a 127 quad encoder, and if 0.001" is your desired end point your at about 2.5 times greater. Not bad, but 200 would get you closer to 4, 250 closer to 5. You may want to look at some adjustable resolution encoders, I think Digi-key carries some. You may want to search the zone for specific mfg/model.

Last, regarding positive down. It sounds like you now have the drive and encoder in agreement with one another, but they disagree with your third party - Axis. You could soft change it with your hal pin assignments, or hard change it by now reversing both your analog drive signal AND encoder feedback.
10-13-2009, 06:18 PM
jvanatta

123CNC

Quote:

Last, regarding positive down. It sounds like you now have the drive and encoder in agreement with one another, but they disagree with your third party - Axis. You could soft change it with your hal pin assignments, or hard change it by now reversing both your analog drive signal AND encoder feedback.

any idea which pin I should change. Not sure which file,.(hal, .var, or .ini?) to find the pins. I would like to change the pins to keep my wiring the same throughout my build.

another question. I am wiring a home switch on this axis but I don't have a clue were to start in the software. I have a 7I37-ta I/O block to use but I am not sure how to set this up. I am curious if there is a place people have posted their .ini files with examples of a similar setups?