Following Error after Converting Config to Metric
- mjohnston
- Offline
- New Member
-
Less
More
- Posts: 14
- Thank you received: 2
12 Sep 2017 22:58 #98903
by mjohnston
Following Error after Converting Config to Metric was created by mjohnston
Hi all.
I've run into a strange issue, and I'm wondering if anyone is able to shed some light on it.
I've got a working machine configuration using inch linear units, and I wanted to make a metric version - same machine, just open a different config and see all the units in metric. (I know there's a menu option to do this in Axis, but I'm using my own GUI and I haven't programmed a fancy feature like that ... yet.
) I made a copy of the entire configuration, then edited the .ini file, multiplying all occurrences of linear units by 25.4, except the step scalings, which I divided by 25.4.
When running the machine with the metric configuration, I now get steadily increasing following error unless I'm going at a very slow speed. The inch configuration gives no significant following error regardless of commanded speed (I've set my max velocity within the motor's capabilities, of course.) The real-world speeds are the exact same, only the units have changed. In the inch configuration, the following error remains essentially constant. Since I'm running open loop "steppers" (actually servos that take step & direction signals and give no position feedback to LinuxCNC), having any significant following error basically means the trajectory planner (or stepgen?) can't keep up to the commanded velocity - is that correct?
Anyone have any idea why I would get following error in metric but not in inch?
I've run into a strange issue, and I'm wondering if anyone is able to shed some light on it.
I've got a working machine configuration using inch linear units, and I wanted to make a metric version - same machine, just open a different config and see all the units in metric. (I know there's a menu option to do this in Axis, but I'm using my own GUI and I haven't programmed a fancy feature like that ... yet.

When running the machine with the metric configuration, I now get steadily increasing following error unless I'm going at a very slow speed. The inch configuration gives no significant following error regardless of commanded speed (I've set my max velocity within the motor's capabilities, of course.) The real-world speeds are the exact same, only the units have changed. In the inch configuration, the following error remains essentially constant. Since I'm running open loop "steppers" (actually servos that take step & direction signals and give no position feedback to LinuxCNC), having any significant following error basically means the trajectory planner (or stepgen?) can't keep up to the commanded velocity - is that correct?
Anyone have any idea why I would get following error in metric but not in inch?
Please Log in or Create an account to join the conversation.
- PCW
-
- Away
- Moderator
-
Less
More
- Posts: 18592
- Thank you received: 5113
13 Sep 2017 01:26 #98908
by PCW
Replied by PCW on topic Following Error after Converting Config to Metric
If you are using Mesa hardware and used pncconf to create the hal file,
you need to change the PID max_error settings to 0 ( or say .0125 ) for al axis
Pncconfs default of 0.0005 is fine for inches but 0.0005 mm
doesn't allow enough velocity correction headroom for the PID
to fix timebase differences between the hardware stepgen and LinuxCNC
you need to change the PID max_error settings to 0 ( or say .0125 ) for al axis
Pncconfs default of 0.0005 is fine for inches but 0.0005 mm
doesn't allow enough velocity correction headroom for the PID
to fix timebase differences between the hardware stepgen and LinuxCNC
The following user(s) said Thank You: mjohnston
Please Log in or Create an account to join the conversation.
- mjohnston
- Offline
- New Member
-
Less
More
- Posts: 14
- Thank you received: 2
13 Sep 2017 14:52 #98923
by mjohnston
Replied by mjohnston on topic Following Error after Converting Config to Metric
That did the trick. Thanks!
Please Log in or Create an account to join the conversation.
Time to create page: 0.125 seconds