[MrBean]

I've been trying to come up with a design for a cheap linear motion axis design using skate bearings and belt drive. Ground rods are not too expensive and work well if supported. I can get 1800mm belts for ~£10. The pictures only show one side of what could possibly make one axis.

All are welcome to comment and improve on the idea. Hopefully we can come up with a good design between us.

My basic requirements were that it be:

1. adjustable
2. cheap
3. rigid
4. Leave room for rod supports

The basic idea uses bearings mounted on small dovetail slides that can be adjusted and then locked using some kind of bolt and gib arrangement. The dovetails would probably require the use of a mill. The models are not super accurate and don't show all bolts etc...Just wanted to get the idea of the carriage design across.

So jump in and let me know what you think. Good, bad, don't bother, whatever.

Regards Terry.....

[Torsten]

You may like to have a close look at the one pictured here.

http://www.cnczone.com/forums/showthread.php?t=8108

A good design in my humble opinion.
notice the lateral ajustment screw on the right side only.
It allows for a slight adjustment of the bearing pressure this
is really all that is needed the rail represents a fixed distance
and you only need maybe a little oversize mounting hole giving
a few microns for ajustment.
A further improvement of this design may be to get rid of the nuts
used for spacing out the bearings and making some dedicated
spacers that take better advantage of the bearing shoulder surface
and provide a wider support on the other side.
Good Luck.

[chuckknigh]

I like the idea...it looks like it could be mass produced with relative ease.

Reminds me a bit of this arrangement, though.
http://www.linearmotion.com/pb.html

Don't you think there's a similarity?

-- Chuck Knight

[Hobbiest]

It does remind me of that links idea. They both remind me of this one though
http://www.buildyouridea.com/cnc/hblb/hblb.html. If mass produced it could be worth it. Dave at buildyouridea.com told me that his design, or at least his method was not economical enough.

[MrBean]

I guess any skate bearing design that allows for rod supports is going to look pretty much like any other.

I could make one up and see how it goes.

I have a CNC router that cuts well in aluminium.
I also have a manual mill to cut the dovetails. I'd make the two matching dovetail parts as two long sections, then cut pieces off at the required length.

One of the main things from my point of view, is that all the holes required are perpendicular to the part they are in. (no angled holes).

The bearings being on dovetail slides should allow a single carriage to fit more than one rod size.

I don't think I'd want to mass produce them. It's really for my own use, but If anyone else likes the idea they are welcome to use it. Thought I'd share it anyway.

I guess it's no better/worse than any other of this type of design, but it suits my limited machining capability, and should be fairly sturdy.

Regards Terry.....

[ynneb]

This is an adaptation from mvauhan's crazy idea post.

[mocnc]

have you or anyone tried this linear bearing using the u-bolt and angle irons? How would you get the nuts to hold the bearings in place?? Is the u-bolt threaded all the way?? Do they make all the way threaded u-bolts??

[ynneb]

I dont know if anyone has made this b4. I dont know if the nuts would be able to spin on the curve either. It was more of a concept drawing that could evolve with input and experimentation. Possibly packers should be used instead of nuts. Possibly the angle iron peices should be turned the other way so as to reduce the length of the Ubolt.

Who is willing to give it a go and try to make such a system ? I would be interested in your feed back.

[mocnc]

please bare with me.....but what is a packers? Well at least I don't think I know what they are? Where can we find them?

[ynneb]

Not to be confused with an alpaca, a packer is a bit of material that is inserted into a place to hold two other bits of materials apart. In this instance a stack of washers could be used as packers.

[nuplowboy]

You can also find packers in Green Bay, WI....

Instead of nuts, you could cross drill the u-bolt close and use cotter pins or roll pins, and a shim pack to take up any slop.

[MrBean]

Well, based on the feedback so far. It looks like the answer is "don't bother".
Most poeple seem interested in pointing me to other designs. While I agree it looks similar to other designs. It differs in the fact that it uses dovetail slides for adjustment and should fit more than one rod size. Maybe I'll just abandon the idea, at least for now.

Thanks for the responses.

Regards Terry.....

[sol]

Well actually...
I think it is a good idea in that it is adjustable, solid, self contained and not dependent on whatever it is connected to for its stability.
The drawback I see is that the top bearing needs to be adjustable as well, or the side bearings have to be able to be dropped to accommodate larger pipe. Of course this could be accomplished simply by adding shims (washers) to the side bearing's mounting/axle bolts.

Another plus for your design is that it allows the side bearings to be slid out far enough so that they could completely clear the track if/when total axis removal is necessary.

One of the features you wanted was perpendicular hole drilling which, though it may not appear so, is there in my units that Torsten linked to (post #2). I came up with that design for the same reasons you listed, but its strong point (very adaptable) is also its week point (lots of slop ‘till finally mounted to its supporting axis). I have found it extremely difficult to create a bearing truck that rides perfectly straight since the slightest misalignment of ANY of the bearings will pull it off course; on test-runs the assembly spirals down the track/pipe. Again that is why I settled for a decidedly basic unit with adjustment in virtually every dimension.

Your idea if carefully built resolves this but of course requires more metal and time to fabricate. For a turnkey unit which is already trued and that could be easily installed I think that yours is good solution.

Build ‘em! Don't let us dampen your enthusiasm!

[Hobbiest]

I think its a good idea Bean. Does seem like a lot of work though. While I would not consider myself to be a lazy person, I do believe in economy of motion, and energy, especially when it is my own! To repeat Sol, do it! If you have the machinery, and the time, try it out! Some of my best designs have always been the ones people tried to lead me away from, in fact, all of my hot rods have been "that'll never work" projects. It sure surprises the heck out of people when you finally show up, revving that obnoxious engine!

Cheers

[ynneb]

Yes Mr bean I also think it is a good idea, it is a nice looking setup and would be very strong too. I say go for it. Make up a set and take some pictures of the results. Give us some firsthand feedback of how they perform. If they are easy to build and cheap then you will have others who will follow you.
I hope you dont feel my earlier contribution was distracting to your great thread.

Oh and nice drawings too, I wish I could draw like you.

[coolman]

The design uses belts. Belts are probably cheaper than acme screws/ball scews and accomodates long axis (e.g. 6+ feet). The question is how accurate can belts be? I understand there will be some flex in the belts, correct? Are there belts that reduce or even eleminate the flex? How about overlapping a second belt over the first one to create a 2-ply affect. Will this work?

[jeffserv]

OK, I saw a man build a unit with Pipe for the rod, and then use skate bearings for the rollers. Then used pipe flanges on the end to support each end with an elbow on each end. This was low cost and for wood carvings and engraving gave good results... would be accurate to within .015" when all dial in I suppose. If you are not trying to machine down to a micron you could look into this idea. If longer runs use heavy sch. 80 pipe instead of sch 40.

I loved the idea for it's simplicity.

[Mcgyver]

the need for accuracy with the belts is eliminated if its a linear encoder closed loop system, but still I wonder if stretching would be a problem.

is your idea to go with welded fabrication, lost foam casting (i really like the build your idea cast approach) or bolt together?

also, those renderings look good - what cad app did you use?

[ger21]

The design uses belts. Belts are probably cheaper than acme screws/ball scews and accomodates long axis (e.g. 6+ feet). The question is how accurate can belts be? I understand there will be some flex in the belts, correct? Are there belts that reduce or even eleminate the flex? How about overlapping a second belt over the first one to create a 2-ply affect. Will this work?

Good quality belts may not be cheaper than decent acme, depending on size. GT2 runs about $30 meter. go to http://www.sdp-si.com to read all you'd ever want to know about belts. The majority of stretching will occur very quickly after installation. After that, good quality belts shouldn't really stretch much at all.

[ynneb]

but still I wonder if stretching would be a problem.

I run my machine over 4 Metres with steel impregnated belts. It seems fine.
I guess it depends on what you are cutting and how accurate you need your items to be. I have had no accuracy issues with the things that I cut. But I guess if want to start machining high precision components other systems would be much better.
All I know is if I cut a square out of timber, 2m x 2m and measure it afterwards it is just that. No micrometer stuff, just a tape measure.