Got a rough version of the print-bed assembly together.

3D Printing
, , , ,
1

Got a rough version of the print-bed assembly together. Again, trying dubious things, to see if they might work.

Core question - what is the simplest fabrication that will work well?

Printed plain-bearing sliders out of PETG (green from Inland) and PLA (yellow). Went through a couple iterations with narrower bearing surfaces, but they did not move freely.

Not fond of the PETG. For fast-print/functional parts, I get better result with PLA. The PETG parts had to be printed slower, seem to have more friction, and are not as strong as PLA. Might be a limitation of my printer, as I settled on the highest print temperatures allowed by my printer (255C).

Also my yellow parts are Matterhacker’s PRO PLA. Given my (limited) experience with other colored PLAs, starting to think that if your interest is functional parts, that “natural” PLA prints faster, and is stronger. Ordered a couple more spools of 3D Solutech “natural” PLA, as I got really nice prints from the last spool of their filament.

All the sliders were soaked before testing with a generic Silicone/Teflon spray from the local hardware store.

Of interest, the supports with the two separated plain-bearing surfaces will slide freely down the rails, just with their own weight. Static friction does not cause the bearings to bind, when load is applied out to about a 30 degree angle from the bearing surfaces.

Once the prototype print-bed is in place (like the high-tech plywood?), how does this behave?

Specifically, can I expect to drive this with one worm screw rod, or do we need more?

Pushing the bed with an off-center force (as though one worm screw), the bed rocks enough to bind and stutter. That is pretty much a non-starter.

Pushing the bed from two sides (as though two screws) moves smoothly, without binding.

Bit bothered by the fact the bed can rock a bit, on the axis between the two screws. That is probably not an issue as the bed moves slowly and steadily down in Z.

As an alternative, considered using four screws, one at each corner. That calls for a ~2m closed loop belt … which is not really a retail option.

(Wrote to the Gates folk, as I would really like to use one of their Kevlar belts. Not expecting a positive reply. :slight_smile: They do not seem to be in this market.)

Dosed the rails in a PTFE mold release spray (a recommendation from my father), and … that does in fact does reduce static friction by quite a lot.

Emulated the single-screw (pushing from the center of one side), the bed moves smoothly.

Right.

So … choices are:

Single screw, plain PLA bearings … simplest fabrication, simplest BOM, could rock a bit but probably not. Assumes PTFE mold release spray applied to rails.

Dual screws, plain PLA bearings …simple fabrication / BOM. Could rock a bit on one axis, but probably not. Could skip the PTFE spray, but why?

Quad screws, with strictly constrained movement, but a non-starter unless the Gates folk show interest.

Yeh. This is fun. :slight_smile:

a9136cfb8d5a43c870fcf0ebc3300957.jpeg
a9136cfb8d5a43c870fcf0ebc3300957.jpeg4048×3036 2.93 MB

f6f956f532bd626cb070d04db21c6ace.jpeg
f6f956f532bd626cb070d04db21c6ace.jpeg4048×3036 2.9 MB

70d333147392698f9751fd0dd8dcc947.jpeg
70d333147392698f9751fd0dd8dcc947.jpeg4048×3036 3.2 MB

2

if you are using printable slides a better option is to use either a ptfe rod or plate between the bracket and the frame. This will give a smoother movement and last longer.

3

@ekaggrat_singh_kalsi In terms of static friction, even better is to use Delrin bearing surfaces. My father (a retired aerospace engineer) already ordered and received Delrin rod and has sufficient tools to machine into shape.

Delrin beats Teflon, generally.

So if I am going to add inserts to my printed parts, they are likely to be Delrin. But … what is the best I can do with simple PLA?

What can be done with simplest fabrication?

4

plain pla is sticky and even if you get it smooth by rubbing it against the frame , it wont last . i have designed printers with pla slides and they eventually fall apart , not worth the trouble.

5

@ekaggrat_singh_kalsi Yeh, I am skeptical about the plain-PLA bearings. But if they serve (perhaps with a simple spray), might be enough. Might be that the PTFE spray reduces the wear enough to make this viable in the longer-term.

In any case, trying experiments.

6

Fun to watch your process!
Best of luck,
Brook

7

4 screws is possible, you would need 4 endless belts from the screws to 4 double pulleys in the centre, and use another endless belt to connect them all with the steppermotor.

8

@Brook_Drumm Thanks. To be clear, I do not know you or your reputation in the community. But looking at your designs, I see solid, beautiful, conservative always-works design. By comparison, I am a guy running around with my hair on fire (an analogy not easily applied to you).

And I will keep to the same theme, as yet. :slight_smile:

9

@Rien_Stouten ​
You need only one belt which is running around all 4 screws. See the Quadrap channel. They have done this.

10

Your build is smaller than my current corexy build , 1mx1mx1m. I started with 4 z rods and 4 nema 23’s and while it moved smooth I could not microsecond past 32 while my x and y are 256. So I actually redesigned everything and went with 3 z rods and 3 nema 23’s and i am able to move the z to 256 and 2x the speed I could get other 4. I am also defiantly against printed sliders especially PLA. Defiantly the best long term investment is go steel or aluminum.

11

@Sven_Eric_Nielsen Yes, but as @Preston_Bannister said, he has trouble obtaining a belt 2 metres long…

12

Splice the belt… get this kit holds up better than anything out there http://myobelts.com

13

I think that if you are going to be using slides on the aluminum extrusions instead of ball bearings of some sort, it may be advantageous to make the bed out of something that is more rigid and consistent in length over time(resistant to warping) than wood. Maybe build the bed frame out of aluminum? on top of that, you can attach a smaller piece of wood to act as a base for the rest of the assembly, however I think that using wood for such a high tolerance and load-bearing place would be a mistake.

14

@Preston_Bannister ​ you could make the slides in the z-axis enclosing the extrusions completely and make them a bit wider. You can print nylon inlays then, sliding in the groove of the bearings so the bed stays straight.

15

Expect to go with just a single screw to drive Z. The bed has no perceptible play in XY, and the force needed to move in Z is small. Just do not see rocking when pushed steadily down from a center-side. And I have headroom in the use of better bearing material, if needed.

The theme is simplest possible fabrication.

I really like the notion of driving Z from all four corners. If I could do it simply, I would. Simplest would be four screw-rods, four printed gears, and a belt. (Printed some test gears.) Modify the bottom-corner brace to offer a bearing surface. Modify the corner-sliders to pass-through the screw, and anchor the screw-rider.

Got a nice note back from the Gates folk. Pretty much out of luck on getting a closed loop belt of that length.

@Rien_Stouten I did think about multiple-belt lashups, for a while. Certainly possible, but adds complexity. Given the single-screw use looks passable, hard to justify adding that much complexity.

Also thought about torsion rods and bevel or worm gears … but again, this could not be of described as “simple”. :slight_smile:

Could use a bicycle chain (or similar) as those are easy to splice to length, but they are not really meant to be used sideways (gravity sag), so … maybe not.

@Matt_Hallett I did look at splicing a belt of the size needed. While possible, this is a bit funky. Doing something like this once is a clever hack. Asking other folk to do it is near to sadistic. :slight_smile:

@Sahil_Jain The hunk of plywood currently serving as the bed is just for prototyping, so I can evaluate fit and motion. Could use wood, if properly treated and sealed, but have not made that choice, as yet.

@Daniel_Stauffer The current PLA sliders work surprisingly well. The sliders use only the inner sides of the rails, of purpose, as the diagonal braces are attached to the outer sides. This was one of the things I wanted to try, before I started building. So far … seems to work out.

On to the next experiment…

16

You could go with two belts both coming into a central motor and pulley. Or three. Should work fine. You can use a couple tensioner pulleys to keep things tight.

Keep hacking!
Brook

17

@Brook_Drumm I agree. And not. :slight_smile:

Could be made to work. Not the simplest.

Saw one of your older videos a couple days ago. Entirely and completely unsurprised that Microsoft reported a lot of your bots still surviving in the wild. My impression is that you are an extremely conservative builder, and I like that a lot.

That said, I want to push design to the ragged edge, for the current state of our art. Once the edge is found, I can backstep a bit, and offer very conservative designs.

So for now, I want to push to the ragged edge. :slight_smile:

18

Are you looking for the simplest or cheapest? I sure like the sound of “ragged edge” :wink: but what is it exactly?
Best of luck
Brook

19

@Brook_Drumm Short answer, I am trying to build a hotrod. As much performance as I can get at modest cost. Simplifying the design and fabrication as much as possible, and trying dubious things along the way. But not cutting corners that lose accuracy and performance.

Or trying not to … we shall see. :slight_smile:

20

My father sent a sketch. I like the approach. Ordered Kevlar cord. Figuring out how to mount the screw, motor, cable guides, and key the nut (on a rail).

So might be able to solidly drive all corners of the print-bed in Z, with reasonably simple fabrication and assembly. Maybe. Will see.

Now I need to figure out how to fabricate gears, to attach to the stepper and worm screw…

cecb74a4f9339e32d44768f65c170588.jpg
cecb74a4f9339e32d44768f65c170588.jpg1275×1650 139 KB