First move on X is approx 1/16th of an inch off

[ravenhearth]

My first test cut was cutting a series of holes on the x axis, one inch apart. Very simple gcode, Show here:

Code: Select all

g20
g1 z-.1 f1.5
g0 z0
g0 x1
g1 z-.1 f1.5
g0 z0
g0 x2
g1 z-.1 f1.5
g0 z0
g0 x3
g1 z-.1 f1.5
g0 z0
g0 x0

The result is a set of 4 holes spaces one inch apart between centers, however the space between hole 1 and hole 2 is approx 1/16th of an inch short, and the spacing between the three other holes is accurate.

I am measuring this by taking my calipers and measuring from the inside edge of one hole, to the inside edge of the next. The dia. of my end mill is 1/8th of an inch, so i should expect to see .875 between edges i believe, however the spacing between the first and second hole is .810 - .820 while the other gaps are what i expect and are .875 +- a few thousandths. This problem occurs on every test run i've tried.

If I alter my program slightly and add .0625 to only the first g0 on x, to 1.0625, then the spacing between edges is now properly .875 (+- a few thousandths)

these images help show the problem
http://i.imgur.com/yqaFy2l.jpg - this is the first set of holes, note my calipers appear to be angled up very slightly but no where near enough to account for the .0625 thousands
http://i.imgur.com/W1ZIsDH.jpg - this is the second
http://i.imgur.com/qsc3N2b.jpg - third

these images are some test programs i ran which appear to be out of scale and may be related to this x move issue
http://i.imgur.com/f7YnQyx.jpg - a gear shape i attempted to cut which is noticeably out of scale
http://i.imgur.com/LdBsthq.jpg - hexagon shape i cut, notice the corner at the top seems to be messed up

I am also experiencing the disconnect problem (shown here: http://carbide3d.com/blog/2016/disconnect-problems/) and i'm not sure if maybe this is related to my problem.

What should I do?

[WillAdams]

Possibly backlash — check / adjust belt tension? Try having the machine move, return to zero, then resume?

For disconnect, if the suggestions at http://www.shapeoko.com/wiki/index.php/ ... mendations don’t help, contact support@carbide3d.com

[ravenhearth]

I tightened everything and reset my grbl settings to their defaults and my strange 1/16th seems to have gone away. How tight is tight enough, though? I ran another program and it appears I now need to do some step adjustment, so I've ordered a test indicator off amazon, a mount for it, and some 1-2-3 blocks that should let me measure pretty well. but how do I calculate the appropriate settings from that?

Also, I am now trying to run my router on what should be another circuit(to be honest, I'm not sure if it is or not.) so that should do the trick.

[WillAdams]

Please see:

- http://docs.carbide3d.com/article/47-measuring-runout
- http://docs.carbide3d.com/article/42-ho ... lt-stretch

as well as http://www.shapeoko.com/wiki/index.php/ ... ting_Steps and the balance of that page.

[ravenhearth]

I seem to have been mistaken. My mysterious 1/16th of an inch is still there.

I ran three different tests, the right triangle formation as described in the belt stretch calibration article, the original set of holes, and a second set of holes on the x axis where i attempted to move 1 inch on x+, then back to x 0 before cutting my series holes, and again in all cases, the distance between edges is off approx 1/16th of an inch between only hole one and hole two, as described in my original post.

The y axis does not appear to have this problem.

I also received a disconnect once during these tests despite using a different outlet, I'll need to try something else, there.

[WillAdams]

Did you file a flat on the X-axis motor shaft? Perhaps the pulley set screw is loose / stuck high from that flat?

Could the X-axis motor be shifting in its mount?

Do slow side-to-side motions and watch each part of the machine until you identify which part is not quite moving properly?

If none of that helps, contact support@carbide3d.com ?

[cvoinescu]

Or even keep the machine still and push the carriage back and forth in the X direction, by hand, to see which parts move with it and which sit tight. If the play is 1/16", you should be able to see the movement pretty easily.

[ravenhearth]

There is no play, all parts are secure and move appropriately, and the pulley does not appear to be slipping. but it would be difficult to see the motor shaft not turning, and honestly i can't tell but there doesn't feel like a slip. I have contacted support@carbide3d.com in the meantime, but received a response telling me that the team is currently at maker fair, so no luck there for the time being.

I didn't file a flat on the motor shaft, as i recall it came with one. But the more I think about it, the more likely it seems that the set screw might be loose and shifting and binding side to side. It would also explain why i can still return to X0, because it picks up that play again when it reverses directions. I'll try tightening it when i get the chance.

[ravenhearth]

Welp, that fixed it. Thank you very much for the help.

It looked like everything was tight, and I did run it back and forth, but it was subtle enough that i couldn't see the motor shaft slipping because of how smooth/shiny it was under my lighting conditions but resetting the pulley worked. Fantastic, thank you.