Need help - gcode not running tool path correctly

More
12 May 2020 20:22 #167545 by dokwine
I am running a CNC router using 2.7.15 on stretch. The machine has been in operation for about 4 years without issues (other than occasional user error :-/ ). I've done lots of projects, but this problem has me stumped.

I am executing a simple slotting routine (generated from Fusion 360) beginning with a ramp cut and then reversing to cut the full length slot. It simulates correctly in F360; it looks right on the screen in Axis; the gcode looks correct. When I run it however, the 1st commanded cut in -y direction (line N60 in attached file) stops short of y1.99 at y3.75, then reverses and cuts the full length slot. The slot is the correct depth causing me to question how y stopped short, but z didn't.

The machine then proceeds to do same thing from the other direction - the ramp cut starts in +y direction, stops short and reverses to the full length cut ending at y1.99.

I have run this same program in the x direction without trouble. same feed heights, same work setup; I 'm cutting a grid. There are no lost steps, errors, limits to motion, etc. The gcode and active codes all look fine to me and the same in relevant aspects as the correctly cut horizontal slots, but I don't consider myself expert. I've rebooted everything just to be sure - runs _exactly_ the same way every time :))

I'm really stumped, and could use some help. I'm sure this is some stupid user trick too - just can't figure out what.

I've attached a screen shot of axis showing my active codes and the spot where the gcode isn't running as expected (cursor). I've also attached the gcode FWIW.
Attachments:

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

More
12 May 2020 20:50 #167548 by tommylight
That is due to low acceleration of the machine and trajectory planer compensating for it.
A simple solution is to add
G64 P0.1
to the beginning of every gcode file, or set it in fusion post processor so it adds it automatically.
Where 0.1 is the distance it can veer of the path, in this case 0.1mm, so set that as you need it.

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

More
12 May 2020 20:50 #167549 by aleksamc
I suppose that problem could be in hardware (PC, servodrive, wires, etc), sources of problems could be a lot.

What machine do you have? What drivers, etc?
You give weak information.

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

More
12 May 2020 21:12 #167558 by dokwine
Thanks, I run G64 P0.001 (imperial here :( as a default, but I'll double check.

Also, the same tool path in the x direction runs fine, and just running the program without the machine enabled gives the same results. I'll double check, but I believe my accelerations are set the same for both x and y.

Thanks for looking
The following user(s) said Thank You: tommylight

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

More
12 May 2020 21:20 #167561 by tommylight
Keep in mind that when using "run from here" it needs that to be added as MDI, otherwise it reverts back to path blending.
There is a way of disabling path blending tha i do not know, and there is also the option of adding it to the ini file so the machine uses that always until told otherwise.
The following user(s) said Thank You: dokwine

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

More
12 May 2020 21:28 #167564 by dokwine
Just tying to keep my post short, in the belief I was missing something basic :)

This is a gantry CNC router. Parts (8020, linear rails, rack & pinion) are from CNC router parts. I modified with Technic step and direction servos (love these!), a Parker Z stage, an ATC ISO20 spindle and of course LCNC running on a Skylake I7 with Mesa 5i25/7i76. It is in excellent repair and has been/is a wonderfully stable and precise machine.

This little slotting program is probably one of simplest things I have done with the machine. The fact that LCNC/Axis shows the same behavior without the machine enabled; runs the tool path in exactly the same way on repeated runs, and the same tool path in the x rather than y direction runs correctly without drama leads me to believe the hardware is still ok.

Thanks for taking a look!

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

More
12 May 2020 21:44 #167567 by dokwine
Thank you so much! That was it!

I expected it had to be something with the active codes. I was burned by G64 years ago, and do use it as MDI generally setting each time I bring up the machine, BUT a) axis was showing G64 as an active code b) the same tool path ran in the other direction just moments earlier without issue and c) I thought that G64 only pertained to path blending on curves. I guess (now) that the 3/8" ramp down over 9 inches is still a curve, and it took the short cut.

Thanks for making me go back and look. I don't need to come to this forum often, but I always get solid help when I need it. Thanks again.

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

More
12 May 2020 22:34 #167571 by dokwine
As a quick addendum: turns out the slots cut in the x direction weren't perfect either. They were all short, but because the cuts were all started from the same end, I hadn't noticed. G64 P0.001 and everythings better.

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

Time to create page: 0.080 seconds
Powered by Kunena Forum