DIY air bearings

[daedalus]

Hi,

Has anyone here ever experimented with producing their own air bearings?
From what i understand of them they are far more tolerant of uneven ways then rolling elements, and dont seem overly complicated to produce (the porous material type at least).

Does anyone here have experience working with them? am i missing something?

[unterhaus]

they crash and burn very expensively. The lab next to mine uses them for precision machining, and crashes are a fact of life. They have very expensive backup air supply system and driers. They are very laborious to make unless you have very expensive tools. I shared a lab with someone who used to make them. There is still polishing compound around the sink that we haven't managed to clean up over the last 10 years.

[phantomcow2]

Indeed, air bearings require a good amount of accuracy. You will find them on CMM machines, and there is a reason they are not cheap.

[DennisCNC]

There is a air bearing X Y slide on eBay right now.

[daedalus]

well i must admit i am very new to air bearings, and havent got much research done. My understanding is that orifice type air bearings are out of the question for home construction, due to the accuracy requirements of the channels.

My thoughts on construction were to build a casing for the bearing, fit the porous media, then have the surface of the bearing and media ground by a 3rd party company, as i dont have the tooling.

Does anyone know any suppliers for porous bronze, porous carbon, or any other suitable materials (preferably UK)?

Is this the sort of thing that i might be able to simulate accurately in cosmos floworks? turns out i have a licence where i work, have never really used it though.

My understanding of the principles of these bearings is that the effect of local surface discrepancies should be averaged out over the area of the bearing, so why is the surface of both the bearing and way so critical? Is this due to the air gap sizes normally used, and interference between the two surfaces during running?

Surely there is a tradeoff between surface finish and air gap size, allowing a less stiff bearing to be produced by increasing the air gap, and allowing a rougher surface finish?

My other question is regarding air quality, i am assuming that the requirements for particulate content of the air supply will require a standard air filter, followed by a coalescing then a desiccant filter to remove water content, is there any extra filtering required?

[unterhaus]

That sounds like enough filtering to me. The guy in my lab with the homemade air bearing used very little filtering. Of course, he would spend a week making a bearing and then run it for a day before it locked up, don't know if that was an issue of him not having enough filtering or the fact that he was load testing it. I don't know that you could consider his bearing high performance, but it ran fairly fast. Too bad he didn't publish his thesis online.

[greybeard]

Porvair in King's Lynn Norfolk might be worth contacting.
I don't know anything about them, except that they are there !
John

[NEATman]

Daedalus-
This is a topic that I work with everyday. I've designed all types of airbearings, and there is art and science to do it properly. The porous type seem to be the simplest type, but there are many tricks to get the proper flow out of the bearing. This website sells porous carbon and porous ceramic air bearings.
http://www.newwayairbearings.com/
Thier prices aren't too bad. These are the most modular components you can buy.
The surface needed to create a good air bearing requires precise grinding. To get the best performance , the air gap or flying height has to be less than .0005". The best performance (about 30 pounds load capacity per square inch of air pad at 60-80 psi inlet pressure) is achieved with a gap of .0003 to .0004". So, the flatness of the bearing carriage and the mating beam must be about .0001" over the area that the air pad covers. This is achievable with a properly set flycutter or a very large shell mill on an accurate mill, but usually this is done by grinding or lapping.

The material used in the graphite air bearings is a porous type of graphite used in die-sinking EDM work. I have produced these before, but it's tricky to get the flow correct. You need to use something like laquer to plug the pores if it is flowing too much air, or if there is not enough flow, there is no easy way to correct for that in a porous bearing. They are the simplest to get flat though, as you can literally lap them flat on a piece of sand paper on a granite surface plate. Just polish your way to a nice finish with finer and finer grits.

We typically produce orifice type bearings, and some designs can be simpler to produce and align than others. If you are trying to DIY, I would suggest that you make the "puck style" like the ones on the new way website. The spherical joint lets each pad self-align. You can use the stiffness chart on thier site to do some calculations for your application.

The filtering arrangement that you suggest is right on the money. If any fluid such as water or oil gets into an air bearing, it will clog the pores, and crash the bearing. Filters are cheap insurance.

Here is much more info:
http://www.nelsonair.com/NA_primer.htm

Here's a 2 year old brochure of some of the machines that we make:
http://www.neat.com/Precision-System...ns-06-2004.pdf
All of the large systems are air bearing gantry style machines (the Gen 5, 6 & 7 machines). The bases are granite, and they are all supported by passive air isolators. These are all linear motor driven, with non-contacting liear encoders, so the only connection is the cables. These can be very high speed systems, some over 3 meters per second. The accuracies over some of these large travels are also incredible, usually a just few microns of position error over 3 meters of travel.

What type of machine are you planning on building?

NEATman

[daedalus]

Thanks for the tips, those machines you build look amazing. I have had a good look through newwayairbearings.com, and its answered some of my questions. The puck type bearings look like the best way to go.

I have a few projects which would benefit from air bearings, but at this stage im just looking to make a small prototype to get some experience with such bearings. I work with robots that are used to simulate virtual objects, so you can reach out and pick up computer generated objects, feeling their mass and inertia. One of the problems is the friction losses in the robots bearings can be felt by the user, air bearings would greatly improve the experience. A far less practical use would be fitting them to my milling machine project, but it might be worth doing 'just because you can'.

[NEATman]

Daedalus-
Your work sounds very interesting. What type of bearings are you looking at? I have 7 3" new way pucks that were slightly damaged, but could still be used for a prototype. Do you have a website that shows what you are trying to make? Does it need to have all 6 degrees of freedom? We make custom air bearings for all types of markets, and if there could be production for this system where I work may be interested. We have designed multi-axis systems, and even systems with non-contacting vertical counterbalances to negate the effects of gravity. Plus, this project just sounds too cool, and I'm currious.

Feel free to PM me if you are interested.

NEATman