define my stock and be sure to select center X,Y and Z
create my supports
set program zero to be the mid line of the model
generate toolpaths.
I have a Shapeoko 3 with the threaded table, so I drill two holes on a center line and bolt my stock to the table and only flip in the Y direction. I’ve emailed grz for support to see if they know of anything but thought I would reach out to the forums too. Anyone else run into this sort of problem?
Example generated gcode of a simple rectangle and picture attachedbox_bot.nc (25.7 KB) box_top.nc (25.6 KB)
It’s the carbide 3d threaded table, so it has holes drilled every 50 mm in a grid. I pick two holes in a straight line, drill in that line 200 mm apart and then bolt through the holes to the table. The holes are 6mm as are the fasteners, so it’s minimal clearance.
How are you centering your Machine between the two threaded holes?
How much clearance between the drilled holes and the fastener?
(Normally a number is the correct answer)
Measure the actual hole diameter: ______
Measure the fastener: ______
Difference: ________
EDIT:
Have you checked the squareness of your Z axis?
Have you checked the squareness of your spindle?
I do a LOT of double sided (and 4 sided) machining. All of these above items will contribute to the outcome of your part.
got ya. yeah I misunderstood your question. To get to the center of the holes I use a 1/4 inch bit and center it to the hole itself. I check to make sure that the machine has stayed registered after every op. The holes are M6 and the bit is 6.25mm and I end up with around .5mm of slop between the fastener and the hole, but I compensate for that by making sure the stock is pulled (or pushed depending on the op) tightly against the fastener.
Z is square, spindle is doing great. I do a fair bit of work on this machine but mostly tiling and 3d carving, this is the first time I’ve tried 2 sided with it and it really seems to only be related to the flip. What’s peculiar to me is that it’s always 2mm, no matter how large or complex the piece. I found the issue after trashing several 15 inch x 14 inch masks that I have been trying to cut that kept coming out with a 2mm drift. The rectangle is just a small test piece because I got tired of wasting so much wood
To elaborate on what @RichCournoyer said about the center setup - by leaving it in the middle, you don’t need to set zero again after flipping, so it will be in exactly the right place with respect to the first side.
No luck moving things to center 0. This is a picture of my basic setup. M6 bolts go through those .25 inch holes and fasten everything to the table along the model centerline, then I just flip front to back along the center line and carry on.
I thought maybe my flipping was not working, so I put a .5mm endmill on the center line, cut 12mm down, flipped the part, re centered on the cut line and went another 12mm down on the back side and got a perfect match on the .5mm hole. I have no idea what I’m doing wrong here, but it’s making my brain hurt I’m going to try Fusion360 tomorrow and see if I run into the same trouble to help narrow it down a bit more.
2 sided machining is definitely a thorn in my side when it comes to absolute accuracy. But I’ll throw in my 0.02 as I have been working on it and while I am not @RichCournoyer accurate I am within 0.05mm when I really really need to.
(1) As all said before I will say getting that 0 on the dead centre of the model in the flip axis (both sides) is pretty critical. Not sure how much the slop is playing here (sounds like you have it nailed anyways based on your 0.5mm endmill test) but I went to using 2 pins in the wasteboard along the x axis (carefully positioned so I can zero the y axis to the centre of one of the pins) and 2 holes in the stock and using that to rotate/locate.
(2) If you are using the MeshCam ‘2-sided’ workflow in all honesty I didn’t like it as it never seemed to work the way my brain said it should work. I ended up doing two seperate one sided jobs in MeshCam - one for each side. This required more thinking but it put me in the drivers seat and made my head hurt less. Once I learned and moved on to Fusion360 there was no going back.
hmm two one sided jobs is another interesting idea I’ll test that tomorrow as well as fusions cam . I mostly use meshcam for heightmaps and 3d carving anyway so a different tool isn’t out of the question if it solves the issue. Thanks for the advice!
got called into work and didn’t get to try much except for the 2 individual ops, which still got me with a 2mm shift. A bit beat today so I’ll try again tomorrow morning and experiment with fusions cam
Something certainly rotten here - two seperate one sided jobs should not have any dependency on each other and its hard to mess up the flip on a rectangle - it should be the equivalent of running the exact same gcode file twice - once on each side. How do the two nc files compare?
I think we’re on the same chase. I just went through the two files looking for every Y move and the coordinates are identical (for Y). I’m starting to think I have fundamentally missed something about how this should work at this point. I’m flipping on the middle line of the model, I can repeatedly flip my stock, home and jog back to X0Y0 and put a .5mm connecting hole through the stock with no problems, but this darn 2mm drift when I’m running a file continues to pop up. I’m using CNCJS, I wonder if it’s worth even trying a different sender tomorrow.
Today’s troubleshooting was…confusing. I 100% know it’s not the software . Everything seems to be tram and square but I went ahead and stripped the S3 down and tuned everything up. I still for the life of me can’t get that 2mm drift to go away so this is 100% user error. I just gotta figure out what I Don’t know I did figure out that using a tapered bit was not helping me at all (see that earlier statement about mostly doing 3d carving) and I switch to a 1/8 straight . thanks a ton to everyone that weighed in !
I think you mentioned that you pull the material towards you when you mount it to remove slop? Maybe I read that wrong? If so, when you flip, push away to remove slop. So pull on initial mounting, then push after the flip. I use 1/4” steel pins to center for a flip. I dril two holes using my SO3 in the waste board on job center, then use the same tool path on my stock to drill two more holes. Insert the two pins in wasteboard and stock, then clamp down. I use those same holes when I flip. Threaded holes and bolts will have some slop, the pins are a friction fit, very tight on both stock and wasteboard.
yup yup , same page. I pull on the top side and push on the back side. I’m going to 3d print some nylon cones this weekend to fit over the bolts and nest into the hole to try a self centering idea I’m kicking around.
Coolness, I just wanted to make sure you weren’t manually inducing the offset. If that doesn’t work try adding the steel dowel pin method and add the two drilled holes to your model centerline. Probably don’t wanna drill additional holes in your aluminum bed, but you could always bolt a piece of MDF to the top and use that. Jogging to the hole centers also isn’t exactly accurate if I’m understanding your process correctly, machine drilled holes should be near perfect assuming you machine doesn’t have loose pulley set screws or something else mechanically wrong.
I’m going to try Fusion360 tomorrow and see if I run into the same trouble to help narrow it down a bit more.
I did figure out that using a tapered bit was not helping me at all (see that earlier statement about mostly doing 3d carving) and I switch to a 1/8 straight . thanks a ton to everyone that weighed in !