Page 1 of 2

Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 2:48 pm
by Camel
I see other people are having similar issues, but I didn't want to hijack another thread.

I'm having issues with my Z axis where the mill either is told to plunge too deep or in other cases too shallow. I've running the stable GRBL v8 and the gcode was generated with makercam with an inkscape svg. I've tried many many different designs and they all seem to do similar things.

Interestingly, I've tried Universal Gcode Sender, Gcode Sender, and GRBL Controller and they all seem to interpret the same gcode's Z commands differently -either going too deep or stuck air cutting.

What's the problem? Thanks
aam tag 2.nc
Gcode
(29.5 KiB) Downloaded 72 times

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 3:28 pm
by RobCee
From reading the GCode and the description, I would suggest that your Z movements are too fast for the machine. You are plunging down in the Z direction at 3ipm or 5ipm and moving up in Z at 30ipm (the F commands in the code).
I would try limiting your Z moves to 2ipm in both directions and see what happens. Once you have that working repeatably you can try increasing it again.

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 4:29 pm
by Camel
Thanks.

I've tried limiting the Z speeds using GRBL controller and also with a recam in Makercam. Same problem. Are you saying the issue is a mechanical limitation on the Z axis, or a software issue?

I've also noticed another, possibly related, bug with the Z axis. After I hit the cancel button in UGS or with GRBL Controller, moving the z axis with the software is very glitched. Asking the machine to move Z up 1 step moves Z up 10 steps, or more. Odd.

Luckily I fitted an E stop...

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 4:36 pm
by RobCee
Could it be an inches/mm problem when coming out of your code?

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 4:58 pm
by Camel
Maybe, how would I check that? Or avoid that?

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 5:28 pm
by WillAdams
Manually issue a G20 or G21 to change to the unit system which you wish to use for jogging the machine.

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 6:44 pm
by Camel
Would a metric/imperial issue make these issues though?

I just tried a simple circle pocket cut made in Makercam. The machine cut the first .05" pass, retracted Z, moved to the center of the circle and plunged about half way back and started the 2nd pocket pass, cutting in the air.

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 7:18 pm
by WillAdams
My suggestion was meant to address the case of having sent a .nc file in Imperial measure to the machine, but wanting to jog in metric.

The plunge rate in the G-Code seems fine (3in ~=75mm), so my suggestion would be to check the current and if there isn't any active cooling, add a fan.

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 8:43 pm
by Camel
Thanks for the help Will.

So you think the Steppers are missing steps? Always possible I guess. I have the drivers cooled with a fan and have some experience with Pololus from back in my 3d printer days.

The thing is the Z height is screwy in repeatable ways. In UGS it does one thing at the same point and with GRBL controller it does another at the same time.

Can someone point me to some simple Gcode that is known to work for you? Something with more Z movement than the Hello World shapeoko drawing.

God this is frustrating... I've spent two weekends trying to get this thing to work.

Re: Z axis issues - either too deep or too shallow

Posted: Sat Mar 29, 2014 9:56 pm
by WillAdams
There's a hand-coded G-code file which I put on the wiki:

http://www.shapeoko.com/wiki/index.php/ ... quare_Test

The actual file is here: http://www.shapeoko.com/wiki/images/6/6 ... e-50mm.txt

Very conservative feed and advance values --- takes ~40 minutes or so to cut.

I think I might have a simpler version which cuts a bit faster somewhere....