Printable Linear Bearing

Mostly, and maybe. This is my first draft of a design for a mostly-if-not-completely printable bearing. I can make no absolute claims of printability, functionality, or sanity. The design is in 3 or 5 parts depending on how you look at it (I think the images are fairly self explanatory), plus a number of 6mm diameter balls. 6mm was chosen so that if someone is willing to make my day and try printing this, they can set their machine to print the big parts and pick up five hundred 6mm diameter balls at the local sporting goods, toy, or superstore as opposed to trying to get their machine to overcome a number of hurdles to produce precision 6mm spheres. You may not expect to find them there, and if you ask the workers for 6mm polymer bearing balls they won't know what you're talking about, because the manufacturer mislabeled them as Airsoft BBs. If anyone uses Solidworks and wants .sdlprt files, just ask.

Print and assemble the parts in the most obvious possible manner (the middle and bottom parts are oriented properly both for printing and assembly, the top is oriented for assembly I believe (if you can't fix that (I don't know if the current software can or not) but want to try this doo-hicky then tell me and I'll force it to flip somehow). The retaining ridge of the middle part should be placed so that the ridge is down onto the matching bottom part. You should add enough 6mm balls to each side so that they can all be in the loop without touching, so they will roll and push against each other. The channel in the carriage has a small clearance for the balls designed into it, the interface between the rail and carriage does not. It is intended that the whole unit should act as a very stiff spring, flexing very slightly to accommodate the balls and thereby stay in a regular position. For permanent assembly, epoxy or glue. I'm not sure if it will turn out to work or not, but the retaining segment may conceivably be flexible enough to allow balls to be added post-glue. If not, it should still do its job and be flexible enough if clipped at some point along its length.

The rail should be printed vertically ideally, with the indent side down. There is clearance built into it (the indent is longer than the out tab opposite) so that the overhang need not be flat to work. I don't have the time or energy (it's almost 3:00 AM) to do it now, but I'll make and post a shorter piece tomorrow.

P.S. Dimensions indicated in the Rail Assembly.jpg file are in mm.

Hi Corwin,

... what's your preferred material?

Until we haven't laser-sintering with steel- or ceramic-powder running, you won't receive acceptable surface-values for ball-bearings.

Or is it maybe better suited for CNC-milling?

Viktor

I think a force pushing down on the top would cause the balls to move inwards, deforming the thin arms on your carriage-middle. Would a solid plane rather than two flimsy arms be better so the inwards force is transmitted through rigid plastic?

otherwise, looks like a neat printable linear bearing. I like the modular rail approach- perhaps a cross-hole for a retaining dowel or something to keep it together?

Looks nice! So print it and see how well it works.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

@Victor - I agree that the balls aren't particularly printable, which is why I recommend airsoft BBs as an unprinted element. If you're talking about the carriage and the rail, I'm not sure I agree, though the design is such that a CNC mill definitely could produce the parts perfectly (I'd definiitely want to use screws instead of the posts, though, the current design is very wasteful of material in a subtractive machining process).

However, the surface geometry on the rail and the carriage ball-channels doesn't have to be perfect. It was designed to produce semi-circular cross-sections tangent to the ball, but it would work just as well with triangles or squares. As long as the spherical bearing ball can only contact each surface at one point and a certain amount of force is maintained on between the carriage and the rail, none of the surfaces except the bearing's need to be particularly smooth. It'll run best if they are, but perfection should not be required (again, in theory).

@Triffid_Hunter - Not sure what you mean. When the balls are engaged with the carriage and the rail, the thin retaining arms don't (or at least shouldn't) take any force at all. They just keep the balls in the carriage when it's not on the rail and guide them into the backside of the loop so they recirculate properly (again, in theory). The whole mechanism for putting it together is definitely first draft. It would be fairly simple to add holes that go through the center of the loops for screws or plastic rods besides (or instead of) the ones out on the edges that are currently in the design. A cross-wise hole (probably teardropped) through the rail for securing it without adhesives is definitely a possibility..

@Forrest Higgs = No reprap in my dorm room regretfully, I'm working on being able to get it printed but if I do it'll be on a stratasys machine, which wouldn't prove much for replication purposes (with acid disolvable support structure you could print the whole carriage as one piece. And also be a tad expensive for me personally.



@Everyone: [www.thingiverse.com] Based on this, the stls may be broken! This is definitely a problem.

Corwin Wrote:
-------------------------------------------------------
> @Victor - I agree that the balls aren't
> particularly printable, which is why I recommend
> airsoft BBs as an unprinted element. ...

Hi Corwin,

My thrust bearing experiments with MDF and hardboard were very successful using precision 5mm balls from McMaster (My son tested it out by standing and spinning on it. I had not planned on performing this kind of test, BTW.) You might want to see if this or similar would be a common commodity world wide, rather than assuming a lower precision bb pellets, etc.

Search for "stainless steel ball" or "steel ball" on mcmaster.com; the part number I used was 9292K39, at $7.66 per 100. Ideally, this would include orders for all the other parts you could get from Mcmaster at the same time to save on shipping (I ended up paying $4.50 in shipping an handling, as I ordered this single item.)

Corwin Wrote:
-------------------------------------------------------
>
> @Forrest Higgs = No reprap in my dorm room
> regretfully, I'm working on being able to get it
> printed but if I do it'll be on a stratasys
> machine, which wouldn't prove much for replication
> purposes (with acid disolvable support structure
> you could print the whole carriage as one piece.
> And also be a tad expensive for me personally.
>
Aw hell, maybe I'll print it this weekend.

Edited 1 time(s). Last edit at 01/29/2010 04:06PM by Forrest Higgs.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

@Beagle Fury:
You've got a good point. What the balls actually are isn't really important.

However, when you start looking online for parts, the price on precision airsoft pellets is an order of magnitude less than the steel balls. A very, very quick google turns up this: [www.amazon.com] as a good example - $11.71 for 5000 means each bb is $0.002342 per BB, versus 7.66/100 = $.0766 per ball in your example. If it turns out that plastic BBs can't do it, then steel or another material would be just fine. But if the BBs can do it (and I think you underestimate the precision of high-grade airsoft ammo), then $11.71 could buy the BBs for several entire printer setups using them in bearings, and that's probably not the lowest price on good BBs out there. Again, I chose the BBs because I know that they are a common commodity throughout the US and could be picked up locally in the states. They're also lightweight, which helps on shipping costs when you start talking about sending things out to the third world to help jumpstart manufacturing (which is supposed to be an eventual goal), and you can get biodegradeable BBs as well.

@Forrest Higgs: That would make my month!

Edited 1 time(s). Last edit at 01/29/2010 04:07PM by Corwin.

... draw the backfeeding curve nearer to the rail, so the balls can better entry - actually the will stall instead of slipping into the feeding ...

Embedding steel-rods in the rail and in the carriage will enhance the smoothness and lower the friction dramatically

Viktor

@Viktor: Closer how? Do you mean a smaller radius curve on the loop? or should the backside be curved rather than straight?

Embedding steel rods in the rail and carriage certainly would enhance the smoothness, lower the friction, and almost certainly greatly increase the rigidity. Also, there would be steel rods in it. As far as I'm aware, a steel rod is rather diffcult to extrude at the same temperature as ABS, though.

> ... the price on precision airsoft pellets is an order
> of magnitude less ... (and I think you
> underestimate the precision of high-grade airsoft
> ammo)...

Sounds good. I don't understimate the precision, but I doubt they are round to 2 microns . Pricewise, you do have a point; if the plastic bearings work out, great; I'm not sure how much percentage wise it saves, but every little bit adds up. In my case, I wanted the extra strength (I doubt my son standing and spinning on plastic bearing would have left me as happy? Hard to say without trying it), and 100 balls allow me to build bearings for 3 RepolaRap machines, so about $2.50 per machine if you ignore the shipping cost.

If you're willing to go with lower precision imperial parts, you can get 1/4" steel balls accurate to 25 microns for $3.11 per 250, or about a penny per bearing. Still not as cheap as plastic, but if hardness or durability becomes an issue, and you designed your bearing carriers to accept 6.5mm (how much tolerance do you need? Would it self correct if you used bearings off by half a mm?), they should just be drop in replacements for the "upgrade premium" machine bearings

Hi Corwin,

... the halfpipe-entries into the feeding should be closer to the rail, so the balls are feeded properly when they left the rail.

And for the steel-rods: - leave canales in the right positions in the rail and in the carriage, then insert pieces of steel-rod ...

Viktor

@Beagle Fury: I don't know if they are round to 2 microns, probably not. And you're right, they would probably break if someone stood on a bearing (mostly because they're hollow, not because they're plastic XD). At $2.50 per machine, you're right that the price difference may not be significant enough to justify the weaker plastic. We'll have to see how that goes.

@Viktor: I was joking about extruding the steel. My real point was that the goal here is to reduce the number of non-printed materials and their cost by as much as possible. If thin steel rods prove necessary, however, I definitely will take your advice on it and design a new rail and carriage with grooves for them.

I'm still not sure what you mean about the pipes being closer to the rail. Currently the spacing is such that a 6mm ball will be exactly tangent to the carriage and the rail at two points each. The internal loop of the carriage has .1mm clearance to let the balls roll smoothly. Do you mean that the carriage should have less clearance for the balls?

Corwin...

I downloaded those STLs. When I brought them up in Netfabb they were TINY, fractions of a mm in size. It looked like the scaling was off by maybe 3 magnitudes. I took them into Art of Illusion and the same thing happened. I scaled them up by 3 magnitudes and the rail looks like it is 117 mm {100 mm nominal} x 17 mm x 20 mm.

Those are very clean STLs. What did you design this parts set with?

A couple of observations.

1) Just making inserts to extend your rail isn't going to work unless you tack it down to a flat surface. You can, for example, glue it to glass. Otherwise you need to design in some sort of groove into the bottom side of the rail and then design a mounting flange that slides onto the groove. Using something like a Picatinny Rail

[en.wikipedia.org]

on the bottom side of your linear slide would work pretty well. Notice that the Picatinny uses a 45 degree profile rather than the rounded surfaces that you use at the top of your linear rail. It doesn't have to accomodate ball bearings so the extra complexity of the top of your rail isn't necessary.

2) You connect the top and bottom halves of your ball bearing slide using just pins. The dimensions of those are too small as features to print successfully. the extruder head will spend too much time hovering over them and they will just melt. I would suggest using holes instead and connecting the upper and lower halves of the slide with either M3 or #4 bolts.

3) You've made no provision for attaching anything to the top of your linear slide. It's okay to print the design without that, but when you get around to actually doing a final design realisation you're going to want to redo that.

As it stands I can't print it, but if you'll scale it in mm instead of that weird fractional measurement that your STLs are currently in and swap out those plastic pins for bolt holes, I'll have a go at printing it. Size the bolt holes for #4 bolts. I have those in my stores. M3s in that length are hard to come by in the US.

The suggestions for rails and attachment flanges are not necessary at this point. I'm assuming that you are using standard bb's for bearings?

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

Weird. I don't know why they came out like that. I made binary and ASCII stl files this time in case that's the issue. I had to zip the ASCII file for the top, it was larger than 614.4 KB

For the new version here, the posts have been replaced by #4 boltholes with counterbores. I also extended the sides into wings to get more material around those holes and allow a third hole on each side for more compression and alignment. An arrangement like this is what I'm leaning towards for a mounting setup, though obviously this is not in any way final.

The process for these is Solidworks 2009 student edition for modeling, then solidworks eDrawings to convert to .stl. They took out saving as .stl natively for the student edition so I have to take that extra step, which may be the issue. The modeling was all done in mm units originally, so I have no idea why they would come out as anything but mm in the stl (but then, I know basically nothing about .stl files).

The intended bearing balls are standard 6mm diameter plastic Airsoft BBs, yes.

Thanks for the feedback and the prototyping!

Yeah, looking at your ascii stl you can see the tiny dimensions on those vertices. I don't know anything about solidworks to know why that's happening to you.

At a guess, it looks like it is scaling your drawings to metres instead of millimeters.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

Interesting. If there are options to make it not do that, I don't know where they are. Can it be fixed by scaling the stl or does it still come out wrong? I brought up the ASCII file in notepad. It looks like everything is at e-2. That means everything that's supposed to be 1mm is turning out to be .01 mm. If you can scale it by 100 times, might that fix it?
Regardless, I'll start digging into this and see if I can figure out a solution.

Edit:
HEY! There it is. You're exactly right, the stl files exported with eDrawings always puts the units in meters. Every value is the value in meters. Everything that should be 1mm is listed as .001, so if it's scaled by 10^3, it should be the right size.

Edited 1 time(s). Last edit at 01/30/2010 01:53AM by Corwin.

That's easy enough to fix.

Don't know why you put those big flanges on your linear platform for the bolts. You could have, I think, just knocked out those pins and replaced them with bolt holes.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

Corwin Wrote:
-------------------------------------------------------
> If
> anyone uses Solidworks and wants .sdlprt files,
> just ask.
I use BRL-CAD rather than Solidworks, but I'd still like to have a look at the files, if you can export them to collada or X3D/VRML I could have a go at getting them imported into BRL-CAD.

@Forrest: I took another look at the amount of material next to the holes and decided to add some more, and while I was at it I thought I'd give it enough to add a third hole for security and to prototype mounting.

@anton: I'll see if I can figure a way to do that. Unfortunately the student edition of solidworks has crippled format functionality, it only lets you save as solidworks files.

Corwin Wrote:
-------------------------------------------------------
> @anton: I'll see if I can figure a way to do that.
> Unfortunately the student edition of solidworks
> has crippled format functionality, it only lets
> you save as solidworks files.


Any kind of text format, retaining the definitions of the solids is will be great, that way I should be able to write a simple importer.

I tried to recreate your bearing in BRL-CAD based on the stl files, here are the results as attachments.

I'm pretty certain I have similar dimensions to you, and I'm not a mechanical genious, and I could definately have made a mistake, but the 6mm diameter sphere in yellow, depicted in Assembled.png seems awfully small to me.

Is my sizing wrong? or should the ball be that small compared to the overall assembly?

@anton: that looks very good! Yes, the individual bearing balls are quite small compared to the whole setup. The design is based loosely on my observations of a commercial recirculating ball bearing and rail system, which I used in my attempted Repstrap last year (it's languishing back home at the CNC plotter stage, no extruder).

Here is a nice picture of a very very similar bearing (mine were also NSK, but not this model) [www.nskamericas.com] to give you an idea of what that looks like.

For a few extra reference points on comparing our models: the circles I used to make the inner loop are 2.5mm radius, and the length of the center, straight section of the carriage is 30mm long. That particular measurement is one of the most arbitrary, by the way. Theoretically, the longer the bearing is the more stable it will be (ie resisting torque and staying straight), while making it shorter would reduce the friction and make for smoother easier motion. What length is actually good, I don't know.

i have only just spotted this post, and i like the idea of a printable linear bearing. i am currently in the process of re-designing all the mendal carrages to accept Standard 8mm bore linear bearings. Theres a guy importing them from china on ebay for £22 for 10 off! very very good price (not sure of the qulity tho..lol). Does anyone have a Mendal running with linear bearings of any sort? if so what sort of issues have you encountered?
Cheers Steve

I've seen the airsoft bb idea recommended here multiple times, and although I think it's inspired, my contribution to it would be to switch to the steel BB's used for BB guns vs. airsoft.

These bearings are at least as accurate, are of smaller size, are virtually the same price, and are capable of much higher loads.

Since the surface finish necessary to print bearing quality plastic balls is a long way off: we might as well use the superior vitamin, as using plastic bearings for the advantage of a standard size doesn't have any advantage in the long distant future when this becomes possible, as the new printed bearings would be flexible to custom specifications anyway.

This may seem shortsighted in a Y2K bug sort of way, however I don't think the additional time it will take to develop a surface finish sufficient for ball bearings of the reduced size vs airsoft size justifies the load constraints.

On the otherhand: the market value of airsoft bbs may be significantly different to bb gun pellets in a country other than america, so in that case: it may make sense.

In any case: I don't think that a choice is really necessary, as the voluntary nature of contributions means that people can use whichever they would like. I will be using bb gun bbs, and I thank the person who came up with the idea of airsoft bbs for the inspiration.

sweaving Wrote:
-------------------------------------------------------
> i have only just spotted this post, and i like the
> idea of a printable linear bearing. i am currently
> in the process of re-designing all the mendal
> carrages to accept Standard 8mm bore linear
> bearings. Theres a guy importing them from china
> on ebay for £22 for 10 off! very very good price
> (not sure of the qulity tho..lol). Does anyone
> have a Mendal running with linear bearings of any
> sort? if so what sort of issues have you
> encountered?
> Cheers Steve


Hi Steve

Give it try and post it as a “high vitamin count” Mendel which was my working title when I saw the same thing. Too many other things going on to give it a try my self.

What you could do is place the bearings in triangle pattern for x and y and then use one on each side for the Z so you would use eight of the ten and have two spares. The bed and extruder mount could be attached by zip tying around the bearings.

For the apex's see the post about drilling out plastic blocks.

Actually I think the RapMan Darwin uses such bearings.

JohnnyCooper Wrote:
-------------------------------------------------------
> I've seen the airsoft bb idea recommended here
> multiple times, and although I think it's
> inspired, my contribution to it would be to switch
> to the steel BB's used for BB guns vs. airsoft.
>
> These bearings are at least as accurate, are of
> smaller size, are virtually the same price, and
> are capable of much higher loads.
>
> Since the surface finish necessary to print
> bearing quality plastic balls is a long way off:
> we might as well use the superior vitamin, as
> using plastic bearings for the advantage of a
> standard size doesn't have any advantage in the
> long distant future when this becomes possible, as
> the new printed bearings would be flexible to
> custom specifications anyway.
>
> This may seem shortsighted in a Y2K bug sort of
> way, however I don't think the additional time it
> will take to develop a surface finish sufficient
> for ball bearings of the reduced size vs airsoft
> size justifies the load constraints.
>
> On the otherhand: the market value of airsoft bbs
> may be significantly different to bb gun pellets
> in a country other than america, so in that case:
> it may make sense.
>
> In any case: I don't think that a choice is really
> necessary, as the voluntary nature of
> contributions means that people can use whichever
> they would like. I will be using bb gun bbs, and I
> thank the person who came up with the idea of
> airsoft bbs for the inspiration.

The only reason I chose airsoft BBs for this is that I wasn't sure if we could print a properly sized and shaped channel for the steel BBs.

Good reason! Didn't think of that.

... here i have some photos of a slider-rule type linear bearing, where the rails are 500mm long and the slide can go 500mm with excellent smooth running ...

This type of linear bearings should be easy to print ... with bigger balls then even them too





Edited 2 time(s). Last edit at 08/07/2010 03:10PM by VDX.

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]