BP Mill Quill Lock and Z inhibit

More
18 Jun 2017 22:17 #94683 by OT-CNC
The physical mechanicals of uncoupling are done. I don't need help with that.
The motors are SEM servo motors that came with the original anilam control. The control has been replaced with a pc running linuxcnc and mesa hardware. There is also no separate DRO. I'm reading position off the A10 linear encoders on the x,y and rotary encoder from us digital on the Z screw. No additional linear encoder on the quill.

Please Log in or Create an account to join the conversation.

More
19 Jun 2017 02:13 #94688 by Todd Zuercher
If it is a servo, you can use the pin "pid.N.enable" to disable the pid loop for one axis. That would make the analog output to that drive zero. With a velocity mode servo drive that should cause the motor to "stop" and be "locked" up with the drive still enabled. But this could easily result in drift or a low speed run away. If linuxcnc is disabled before you return the normal control of the Z pid enable. It should not cause a jump in motion. You may need to do something with following error settings to prevent alarms.

Please Log in or Create an account to join the conversation.

More
19 Jun 2017 13:57 #94698 by OT-CNC
Thanks Todd. If I use pin "pid.N.enable", will linuxcnc limit any motion to that pin if I inadvertently hit the jog keys or rotate the mpg for that axis or have a g code ready to move that axis? What about turning off the input side of that axis like the jog and mpg so not to trigger a joint following error? Does that make sense what I'm asking?

Please Log in or Create an account to join the conversation.

More
19 Jun 2017 14:21 #94699 by Todd Zuercher
It will not send ANY commands to the analog out pin. This is why it is little dangerous. If the drive starts to drift or runaway, Linuxcnc will not be able to send any corrective commands.

Please Log in or Create an account to join the conversation.

More
20 Jun 2017 17:07 #94734 by andypugh
I think the answer here might be to latch the pid.2.command input, That should leave the motor locked but active.

You can do this with a mux2. You connect one input to the position-cmd of the axis, and the other is looped back to the input. Then the select pin switches between the current command, or the latched value. Then send that output pin also to the PID Input.

However, you will get an f-error if the command has changed during the time the command input was latched.

Please Log in or Create an account to join the conversation.

More
18 Mar 2018 19:36 #107507 by OT-CNC
I'm coming down the home stretch on this retrofit...sorry to keep pestering you with a billion questions. One moment I feel like I'm understanding something the next I'm back to square one.

I'm trying to implement latching the pid.z.command input that was suggested above. I need to configure this correctly as I can't afford a runaway on the z. I don't think I'll be quick enough to hit estop on such short travels. I'll park it in the middle before trying the code and switch. I do have limits installed but hardly any over travel.

Anyways, I started with this which I know is not correct and need some guidance in completing it:

# --- manual Z-QUILL-LOCK ---
#net z-quill-lock <= hm2_5i25.0.7i77.0.0.input-14

loadrt mux2 count=1
addf mux2.0 servo-thread

net z-quill-lock mux2.0.in0 <= axis.z.joint-pos-cmd
net z-quill-lock mux2.0.in1 <= mux2.0.out
net z-quill-lock mux2.0.sel <= hm2_5i25.0.7i77.0.0.input-14
net z-quill-lock mux2.0.out => pid.z.command.input

##

Please Log in or Create an account to join the conversation.

More
19 Mar 2018 14:03 #107544 by andypugh
Actually, that looks right to me. Does it work?

(You could set the PID output limits very low, if you want to make sure that things go wrong slowly)

Please Log in or Create an account to join the conversation.

More
19 Mar 2018 18:36 #107557 by OT-CNC
No I got an error. I'll have to re-load to see the specifics. I don't remember what it was.

Please Log in or Create an account to join the conversation.

Time to create page: 0.411 seconds
Powered by Kunena Forum