Wednesday, July 07, 2010


selective laser sintering part 7: x-y axis

Hi folks,

Just a quick post -- I've been tinkering with the idea of making an entirely laser-cuttable x-y axis using nothing that isn't laser cut except bolts, nuts, and steppers (essentially things that wouldn't be immediately available at the simplest of hardware stores, plus the steppers).

I thought I'd post my progress so far, even though the latest iteration isn't going so well. Basically driving a long axis from one side introduces a shear force, so you typically would drive it from the center (like with a belt, or lead screw), drive it from both ends (either with two motors, or a long shaft that transfers the drive from both ends), or try and come up with some very precise linear rails and bearings that prevent all the shearing and torquing from happening. I chose to try that last bit, since all the other options involve extra parts (like shafts, belts, or screws), or extra motors. This isn't working out so well, but I still thought I'd post some pictures and my progress for informative purposes (sometimes design iterations that don't work out so well are just as informative as ones that do).

Happy to accept design thoughts from folks who have attempted this before?

thanks for reading!

[part 1] [part 2] [part 3] [part 4] [part 5] [part 6] [part 6 video] []

Labels: , ,

Hi Peter, glad to see your XY axis is coming along, even if it's not perfect yet!

Here are some ideas:

I cannot tell form the pics, but I am guessing you are using a rack and pinion system to drive the axes (just a pinion on the stepper motor, and a rack in the groove)

All your wheels/pinions are rolling in grooves, meaning that any misalignment introduces friction on their sides. Eliminating this and using only the rolling surfaces for guidance would be better.

For the small (top) carriage:
I think you might need to add a third vertical-axis wheel on the opposite side from the motor to stop the carriage moving away from the motor-side.

For the long (bottom) carriage: it is indeed probably too long to be driven from one end only.
This could be solved in 3 ways: your already mentioned center-drive or double-ended drive; Or decrease the aspect ratio of the carriage (make it wider, at least at one end).

I expect the increased dead-space incurred by Solution 3 is undesirable, so here is an idea for a double-ended drive, with few extra (non-lasered) parts:

Replace the up-down guiding wheels + motor with a rack and 2 pinions per side.
Link one pinion to its counterpart on the other side of the carriage with a shaft. (This shaft can simply be a strip of laser-cut wood, with a square-peg-in-square-hole link to the pinion).
You will need a small bushing/bearing where this shaft goes through its guiding hole (think nylon standoffs, pen body, plastic straw, skate bearing, etc.), and maybe a small disc to make the shaft circular where it goes through the hole.
Then, either drive one of the other pinions with the motor, or (maybe better) drive one of the coupled pinions via an extra pinion on the motor.

You seem to already have 2+1 wheels for side-side guidance, they just need to not be in grooves; 1+1 might be sufficient since the coupled pinions already guarantee parallelism.

I realize this is not very easy to follow, so here are some pictures:

Also, the .skp file:

Feel free to grab and rehost these.

In this model, I have staggered the pinions relative to the rack, so they would not touch a flange on the outside of the rack (flange not modeled).
Obviously, the rack and pinions should have proper profiles, mine are just schematic.
Also, this model relies entirely on gravity to stop the carriages from lifting off their (t)racks, more wheels would be necessary to prevent this. Another possibility would be a very loose pin-in-groove approach, preventing the whole assembly from falling apart, but not touching at all in normal operation.

Best of luck,

What modeling software are you using?
Hi Jonathan, thanks for the comments!

Sorry for my brief description -- you guessed correctly, it's using a rack and pinion system to drive each of the axes, where the rack is sandwiched between the guide rails at the stepper-end of each axis. At the opposite end, in place of a rack, there's a rail that small idler wheels ran ride upon.

Since centre drive isn't an option, I think you're definitely right in that driving from both sides is likely the best bet -- the trick being how you accomplish that using only laser-cut parts. I'm not sure that cutting a shaft would work particularly well -- over the length of the axis there would probably be a good amount of torque on the shaft, and I don't think it'd be just as simple as cutting out a small (or even fairly wide) strip of material to use as a shaft. You could probably come up with a tounge-and-groove design for a shaft with spine supports?

I think the next thing I'll give a try is using a system of gears sandwiched between the layers of the axis to bring torque from a stepper located at the center of each axis to two points at each end. It's kind of a crazy idea, similar to the design of the dual-z axis, but with far more internal gears. This removes the need for a long shaft between each end, but the friction across that many gears may be too much for these little steppers. We'll have to give it a try and see how it works! :)
Hi Peter,

If the sides of the gears/wheels are not rubbing against the sides of their grooves (sides that you call guiding rails) there _should_ not be too much torque needed across the shaft.
This is just my feeling though, it would need to be tested of course.
If the strip-shaft is too flexible, it could be strengthened by gluing extra pieces on each side of it (basically give it a solid square section); This would be slightly smaller although a bit heavier than the equivalent-strength hollow-square or I-shaped section, but a lot simpler!

If you go with your multi-geared method, look out for these:
* Friction due to many gears (use fewer, bigger ones if possible)
* Friction on the sides of the gears: Make sure you use washers so the gears spin freely without touching the layers above and below (you might want to do the same on the Z axes drive also)
* I doubt bevel gears will be possible so you'll need to drop the racks on their side (not much of a problem)
* Extra gears means extra backlash (how much will depend on the precision of the laser-cutting and kerf compensation), but this can mostly be corrected for in software, or pretty much ignored as long as you are always going in the same direction.
You could also make anti-backlash gears, but that would probably be overkill.

Thinking about it, if you don't need much height above the carriages, a center-drive could be quite feasible, and possibly simpler than the other methods, using a rack above the carriages (down-facing would be easy, its support columns being part of the same layer).
The side-side idler guide wheels would need to guarantee the parallelism properly though.

Hia Peter i like your laser sintering, but im wondering what program are you using to make the g-code for the parts you will sinter? Is ReplicatorG enough?

Im aiming to make my own vertion of a metal 3D printer, i gott a cnc cotnrolled cplasma cutter now but some of the stuff i cut needs after work like turning. So i been looking to make a 3d printer instead to make parts that is good enough to go str8 from printer to costumer.

Best regards Toffe "The mad swedish inventor" (and farm owner in Zambia)
Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]