I have 3/8" anchors and a piece of 3/8" threaded rod going down to the anchors. When not using my M4, the anchors are pretty unobtrusive or I can quickly unscrew the threaded rod and pop in a 3/8" hex bolt instead to keep the hole clean and then the only thing protruding up is the head of the hex bolt.
If needed later on, I can always get a bit of quick mix cement and fill the holes back in (drop in anchors and all) and it’ll be like nothing was ever there.
I wouldn’t worry about that offset for now. The difference the offset makes is likely pretty minimal and likely only an issue when making cuts closer to an anchor. Getting a frame set up and making cuts is much more important. You can always go back, make changes, and fine tune things later if you’d like.
I built a small version of this (each side is 2.5 ft, giving about a 12" square work area, it would be a couple inches bigger but the sled hits the corner braces) and got some pictures over the christmas break. I got a fitness of 0.7
currently two pictures and two videos of it calibrating.
I need to pull it out of the back seat of my car and try again with 0.88 firmware. I wasn’t especially happy with it, I think I need to reduce the calibration tension as it was causing the router to tip at times (which is why I tried recording video of the calibration)
Can this idea be extended to get rid of the belt length knowledge requirement?
Theoretically you can use 2d triangulation if the Z offset isn’t involved.
that doesn’t help, you still need to know the belt lengths to do the 2d
triangulation.
Would never then need to retract\extend the belts and things could use
relative movement…
relative to what? when you don’t know the belt lengths, you have to do something
to establish a starting point. That’s what the retraction is for.
Though I like the idea of the Z offset, it seems like it would pinch the sled to the workpiece good.
it also causes the sled to want to tip more at the edges, and the higher the z
offset, the more error you introduce if your z offset doesn’t include your
workpiece thickness, etc/
I deleted the post because it was complicated and didn’t want anyone to waste time…you must have mod rights and saw it. I might write some test code later, I believe there is a way to know 4 points in space if you know relative movement along each point. So the device would tighten all 4 motors somewhat, raise current to one motor at a time while measuring 1 dimensional movement on all magnetic encoders. By knowing how far each line moves when you have a known line moving a certain distance - you can model that after several movements to locate each anchor point in space. I think all information can be unknown except relative movement towards each anchor point during a calibration process.
I am not that great at math so it would take me a while. It might not be triangulation. It might be least squares or something then proofing with pythag. I feel somewhat confident it would work if the Z is on the same plane, it becomes more complicated when Z is offset by a known or unknown amount.
There’s also two other pieces of known information - current driving each motor, you guys seem to have something there in the hardware - and an intended frame design. In a perfect world, if you wind tr, br, and unwind br, bl, under tension, the device will move in a straight horizontal plane. In a non square non perfect frame - some motors may see more current draw as the belt tightens due to off plane movement. That’s also a piece of information one could use to see if anchors are offset from perfect config.
It’s not a complaint, more of a thought experiment. Once I am using my 4.0 I am likely going to order a 4.1 upgrade kit, and I am going to mess with porting everything to grblhal which looks a bit easier to modify. Might try it months down the line.
I don’t think anything needs to change right now, what is being done is pretty accurate and people are making great cuts.
Okay, so I copied this one and tried to use Onshape to modify it, but I need to take some lessons, I guess!
I was trying to turn this into a attachment point for the ends of the 2x4s for the Portable Horizontal frame.
I see this sliding over the end of the 2x4 flat, about 2 inches, with the attachment tower, starting with the same displacement at the bottom and ending at the farthest point.
I was considering removing the extruded tube segment in the center and seeing if a friend would consider making a test fit out of welded aluminum or steel.
I would add two mounting options: a hole in the center to align with an insert nut and a bolt, or four corner holes for the use of would screws for a more permenant installation.
that is what the current calibration routine is trying to do. Please keep
working on your idea, it would be great if you came up with a simpler process.
Okay, so I copied this one and tried to use Onshape to modify it, but I need
to take some lessons, I guess!
I was trying to turn this into a attachment point for the ends of the 2x4s for
the Portable Horizontal frame.
I see this sliding over the end of the 2x4 flat, about 2 inches, with the
attachment tower, starting with the same displacement at the bottom and ending
at the farthest point.
so instead of the anchors being within the height of the 2x4 they are above it?
I’m not sure I would trust the plastic to be rigid enough to do that.
I was considering removing the extruded tube segment in the center and seeing
if a friend would consider making a test fit out of welded aluminum or steel.
I would add two mounting options: a hole in the center to align with an insert
nut and a bolt, or four corner holes for the use of would screws for a more
permenant installation.
my first version had such mounting holes, but when I assembled it I found that
just pulling everything tight and then screwing down the angled corner piece
made it very solid, so when I did this new version, I didn’t bother.
