z axis reversed

mariusl wrote: Andy,
Will it be the OUTPUT_SCALE setting under each individual AXIS section?

Is it stepper or servo?

Servo system Andy.

mariusl wrote: Servo system Andy.

Then you need to make the encoder scale negative. Or possibly swap the encoder A and B phases.
This might also need you to negate the PWM scaling factor too.

Andy,
I made the encoder scale negative and the output_scale negative and got a huge runaway on that axis. I did not swap the A and B signal yet.

So I will have to keep the ENCODER_SCALE as is and make the OUTPUT_SCALE negative and swap A and B signals on the hardware connection.

mariusl wrote: Andy,
I made the encoder scale negative and the output_scale negative and got a huge runaway on that axis. I did not swap the A and B signal yet.
So I will have to keep the ENCODER_SCALE as is and make the OUTPUT_SCALE negative and swap A and B signals on the hardware connection.

This is a situation where things tend to cancel out. For example if you swap A and B and then change the encoder scale you will end up exactly back where you were.
(except that the PID might get confused).

I suggest you need to undo one of the changes, not add more.

what final parameter does OUTPUT_SCALE actually change in your system?

Sounds like it did nothing (so you ended up with positive feedback=runaway)

PCW wrote: what final parameter does OUTPUT_SCALE actually change in your system?

Sounds like it did nothing (so you ended up with positive feedback=runaway)

That is correct, it changed nothing when I made it negative. It does make sense though to swap the feedback and make the scale negative. Then the axis motion should be reversed with proper feedback.

I will be back at the machine on Tuesday and then I will test what Andy suggested. I will give feedback.

just recently ran into a similar issue because I have a strange slant bed lathe with the spindle on the right from operator's point of view. Once I got my axes properly inverted so that away from the spindle was +Z and away from the spindle rotation axis was +X I ran into a new problem. To go left from my point of view I had to use the right arrow. To go Right the left, up the down.... you see my predicament.

Fortunately, I was able to dig through axis.py and flip these keys solely for the purpose of jogging the axes. You could even swap the axes if you like... say because your computer is at 90 degrees rotation relative to your router/mill/lathe. The relevant lines of code are here with my mod:


Hope this proves useful to someone
-Marc

PS: if anyone knows how to mirror the display I'm still trying to figure that out... see the display assumes that my spindle is on the left, but mine is on the right.

grossman83 wrote: PS: if anyone knows how to mirror the display I'm still trying to figure that out... see the display assumes that my spindle is on the left, but mine is on the right.

The display is in glcanon.py
github.com/LinuxCNC/linuxcnc/blob/master...274/glcanon.py#L1090

If I was in your position I would experiment with adding a
glRotatef(180, 0, 1, 0)
around that point.

I think that your lathe is basically conventional except rotated through 180 degrees on Y. glrotate is (angle, X-amount, Y-amount, Z-amount) In fact I think that those two lines there could be (90,1,0,-1) but with rotations order matters, to there might be a reason for the current two-line method.

glcanon is Python, so just changing the file and restarting LinuxCNC will work, no need to compile anything.

Thank you, thank you! This knowledge about Geometry= x-yz saved me a ton of time!!!