CINCINATTI MILACRON ARROW 500 (BDS4 + VFS5) WITH MESA 7i77 + 6i25 WITH LINUXCNC

Hi All,

This is the file mesact generated. I change the "channel 4" to channel "1" manually, and it loads to linuxcnc. I'll be wiring the 7i77 to the servo drives to test if the servo drives will respond to the jog commands. Wish me luck.

Regards,
Kelvin

Make sure the drive enables work so a runaway is stopped
before damage occurs. Alternatively, decouple the linear axis
from the motors.

A blind setup has about a 50% chance of a runaway because
either the encoder direction or analog output polarity are reversed
resulting in positive feedback = runaway.

Hi all,

I tried one axis first to be safe.

If encoders are disabled in HAL file, I can move the y-axis freely. no shaking. everything is smooth.

But if encoders are enabled, I have varying results.

a lower scale (819.2) - joint 1 following error
a higher scale (4096) - no error but shaking occurs

May PID settings are
P - 50
I - 0
D - 0

Ballscrews are around 10mm

Regards,
Kibin

 

I would use the higher resolution, perhaps increase the following error limits
and proceed to tuning.

If you have oscillation, reduce your P term.

For velocity mode drives, most of the tuning is setting the correct FF1 term

If you set the analog scaling to machine units per second at 10V, FF1=1.00

You can change the PID values live with the "Calibrate"  window.

Hi PCW,

Noted on this. Ill first wire the enable of mesa 7i77 to the enable of the servo drive bds4 since I found out that there is a difference in bias/balance (adjustment for zero movement upon drive enablement) if I force enable (jumping the enable of the servo drive (Daredevil)) instead of the mesa board.

I'll first connect the ENAn of the 7i77 to the servo drives so that once the drives are enabled, there will be no movement.

After that, will follow your instructions and use 4096 as a scale starting point to fine tune one axis.

Thank you for your advise.

Regards,
Kelvin

Hi All,

Good day. Managed to wire the ENA to the servo drive enable.

My current problem is overshoot. It is overshooting. lets say I command it y5, it will go until y6 then go back to y5. Overshoot increases as feed speed increases. I estimate it to be 1mm per 100mmpm. so 100mmpm has an overshoot of 1mm, 300mmpm has an overshoot of 3mm.

PID settings are
P = 1.5 (If i add more, it shakes at higher speed)
I = 0
D = 0 (any increase and the system shakes)

Scale : 1225 (0.01 accuracy even with overshoots)

Hi All,

Good day. Managed to wire the ENA to the servo drive enable.

My current problem is overshoot. lets say I command it y5, it will go until y6 then go back to y5. Overshoot increases as feed speed increases. I estimate it to be 1mm per 100mmpm. so 100mmpm has an overshoot of 1mm, 300mmpm has an overshoot of 3mm.

PID settings are
P = 1.5 (If i add more, it shakes at higher speed)
I = 0
D = 0 (any increase and the system shakes)
FF0 = 0
FF1 = 1
FF2 = 0
Scale : 1225 (0.01 accuracy even with overshoots)

Regards,
Kibin

forum.linuxcnc.org/10-advanced-configura...to-example-mesa-7i77

Hi all,

After countless revisions. managed to make it work.

PID settings are
P = 1
I = 0
D = 0
FF0 = 0
FF1 = 0
FF2 = 0
Scale : 1225

It was staring at my face the whole time. On a whim i just decided to 0 everything, and it worked!

Regards,
Kibin

Hi, would you be willing to share your linuxcnc config files?
I have a Cincinatti Milacron Sabre 750 with similar drives but the heidenhain control and powered it up for the first time & the hard disk has just failed with no backups.
I believe replacement drives are NLA, so it might be as well just just bite the bullet and go straight to a retrofit with linuxcnc.

I have done a linuxcnc retrofit before and have spare mesa boards, so would be using a similar mesa hardware as yourself, and any configurations would potentially be very helpful.