Jerky Shopbot from Aspire Resulting in Unwanted Sanding and Potential Machine Damage

[JeffPantheon]

Hello all,

I have been struggling with using projected vectors onto 3D surfaces to cut profile paths for going on a month now. Asking for help from the Universe (of fellow Shopbot users).

For visualization purposes, we are cutting skateboards out of pressed panels on vacuum fixtures. The important pieces of this puzzle are that we are projecting vectors onto said surface to cut the boards. I am getting jerky motion in the files. Depending on our VR settings, they will either be violent jerks (when "3D ramp threshold" is set closer to 1000 and "Minimum Distance to Check" is set closer to .01) or smooth ramped starts and stops (when those same settings are set to defaulted 100 and .100, respectively). Bottom line is we are getting lines in the side of our workpiece aka skateboards which is taking way too much time on the drum sander and actually changing important aspects of the shape at times (points can no longer be points, etc).

So this is what I've tried:
-Obviously tried adjusting VR settings. The settings which seems to make the biggest difference are those listed above, but neither change the final product
-I have tried running the vector without projecting to the surface, so no Z change. This results in a smooth motion with the machine. No problems whatsoever. The Z changes are most definitely related to the jerkiness, though it does APPEAR at least that the jerky motions are happening on the X and possibly the Y axis. Touching either motor, it is unclear whether the motors are jerking or not, but they do not feel like they are. The entire table jerks, though.
-I have run the file without the workpiece. The jerks remain. So this is not vibration happening on the workpiece or a lack of vacuum. The jerk motions are in the exact same spot every time, and seem to be most closely related to changes in Z height.
-I have changed my Z feed rates, thinking that the Z may be limiting the XY. This is not the case.
-I have changed the stepover on the bit info. I don't know why I thought this might make a difference, but it didn't.
-I have changed the cut type from Outiside/Right to "On" and received no change.
-I have tried 3 different surface types from Rhino, importing native 3DM, STL up to .001 resolution, and the highest resolution of OBJ. All three file types still result in jerky motions and appear to be in the exact same spots.
-I have tried reducing the cutting speed. Doesn't not change. The jerks almost appear to be written into the code.
-I have tried multiple surfaces and multiple vectors. Happens to every file.
-I have tried "Fit Curves to Selected Vectors" and reduced each vector down to the absolute minimum amount of points possible, thinking that there might be micro step-overs happening from Rhino to AI or DWG to Aspire. No change.
-As a weird side note, our machine has two Y-Cars and 2 Spindles (so I have to be careful not to bump them up against each other). I have moved my primary Profile Spindle out of the way and brought over my detail Spindle (which I use with a big 2" ball for detail cuts like wheel wells) and ran the file with that spindle/y-car. Jerks in the same places. As stated, appears to be in the code.

I'm running out of ideas. Well, yea I'm totally out of ideas. I believe the table to be perfectly square. It's far more square than the last Shopbot I was running, actually. And if it weren't, you'd think adding in the Z change vs running flat would make no difference, so I don't think that's the problem either. Any ideas at all are welcome. I'm told by Tom at Shopbot that their Shopbot Buddy is jerking in the file too. Please help!

Thanks

Jeff from Pantheon Longboards

[JeffPantheon]

Attached a sample file.

Also, we have tried several different posts. Shopbot (Inch) and ShopbotTC (Inch) are the ones we fall back on for their simplicity but nothing else we have tried has worked either.

[Gary Campbell]

Jeff...
I am pretty sure I have seen this before. Can you email me the 1st 50-100 lines of the file that produces the error?

GCnC411@gmail.com

Got it!

[Gary Campbell]

Jeff...
Do an aircut test with the ramps back at the defaults, and your speeds set at 2ips (XY) and 2ips(z) and repost.

[JeffPantheon]

Just tried it. Still getting jerks in the file. They appear to be in the exact same places, but the jerks themselves seem like they're causing the XY motors to ramp a little differently. Doubt it would affect the work piece based on what we've done so far.

[Gary Campbell]

Have you reset the ramps to the defaults?

[JeffPantheon]

Yes. At the time when I bumped the ramps all wonky, I was actually getting better results out of one particular file. Thought I had solved the problem, and then I went to another file and it the jerks are just super harsh with those values. So I was advised to take all the ramps back down to defauls, and I did. We still get jerky motion with the cuts, but it's not violent. It's just lots of starting and stopping, ramping up and down in speeds in places where the motion ought to all be smooth (no corners, just smooth curves).

Here's a kinda weird side note: I'm also getting these weird waves on the corners when I cut rectangles. Specifically, I'm cutting a male rectangle on all my fixtures to act as a sort of lock and key to my table, which is outfitted with female rectangles. At the corner of each rectangle, the machine can't seem to keep a straight line, it's making a really small S curve instead of holding a straight line and then making the turn at the corner. Corner itself is straight, but every single rectangle I'm making is doing this. I have no idea if these are interrelated, but if they are and you've heard of this issue, maybe it kills bird/s with stone. At any rate, truly appreciate your efforts! Thanks!

[srwtlc]

Jeff,

I'm thinking that the cause of the problem is the result of projecting the vectors onto the 3D surface. You could be picking up pixels either around the perimeter of the edge of the model or just due to low model resolution. If you were to make sure that your job dimensions are close to the size of the part and jack up the modelling resolution, it may make it a bit smoother. If the profile is the edge of the model (drops off to nothing beyond that point), you could be picking up stray pixels along the way. Possibly over sizing the model to be outside the projected profile will help. It's the change in Z height from pixel to pixel that's causing the XY to slow down for the Z and making the jerky motion. Higher resolution, smoother model to follow.

That aside, if you adjust the ramp values for 3D ramp threshold and slow corner speed way up, it might make it acceptable for you. Also, match the XY and Z feedrate, making it 3,3. By doing that and setting the 3D ramp threshold to 500 (an outrageous amount) and the slow corner speed to the max of 99, I was able to just about dial it out.

[Gary Campbell]

Not by a machine, but if you look at the code, there are some longer than expected segments

[srwtlc]

Here's a zoomed in shot of the toolpath from the -X end looking towards +X. You can see a lot of Z axis action as it travels right to left and then as it goes away from you in the upper left part is a compressed view of all the Z moves that are abruptly up and down along the long side (shown in the side view toolpath shot). This is the 'projected on model' profile toolpath following the pixels in the model. A higher res model might make it better.