frictionless bearing

[willib]

need a frictionless bearing

the bearing has to support 144g and spin frictionlessly

i was thinking of using a ring magnet and a cylinder magnet, as the bearing

 the ring magnet would be fixed or otherwise supported and use magnetic repulsion between the ring and the cylinder?

use a string connected to  the cylinder , with the string going throgh the hole in the ring?

i have tried all kinds of roller bearings , and all have too much friction

so with the load (144g) and the cylinder magnet (36g) the total weight the bearin has to support is 180g

a cage would have to be built around the ring and cylinder to keep if from flopping over when there is no weight on the string

what size ring would anyone recommend?

and does a ring magnet have enough repulsive abilities to support the needed weight?

any other ideas are welcome

[maker of toys]

are there size constraints?  will the application be stationary while the bearing is in use? how high wiill the rotational speed be? is the load balanced around the axis of rotation?(or could you arrange for balance by adding a little more mass?)

 you could use a fluid bearing. . . the old lighthouses used to float the lens assy on mercury in a narrow trough.  things were so frictionless that a 10 lb falling wieght could operate the rotating assy for something like an hour per foot of fall. granted, we're talking maybe 5 rpm, but the frensel lenss often added up to 4-5000 pounds of glass and brass.  Most of the drag came from the windlass gearing and the occasional 'guidence bump' of the floating assembly against the side of the trough.  I believe that something less than a gallon of mercury was required for a 6 ft diameter bearing, because the trough was made to a very close clearance on the 'float.' Good leveling is important with a close-clearance fluid bearing . . . fortunately, such a device has its own built-in precision bulls-eye level. <G>  

(note, I'm not recommending mercury persay . . .just something with a low vapor pressure.  Like some of the newer 0w-20 motor oils, baby oil, or maybe low odor kerosene.  if increased size and maintainance is allowable, water or alcohol would work, too;  the lighter the fluid, the higher the allowable rotation speed.)

a simple one can be made with a couple of nesting soup-bowls or deep saucers. . . add just enough fluid to one to allow the other to float by a millimeter or so. of the bearings I outline here, the fluid bearing (being inherantly lubricated) will far outlast the others.

another possibility for such a light mass is to use a single needle/jewel type bearing. these are used in the toys that you may have seen that look like a light bulb with a little anemometer looking thing in them that spins when exposed to light.  how's that for frictionless?

a simple one can be made by grinding a sharp point on a hard metal rod (I've used TIG  electrode and hardened drill rod) and a punch mark made on the mating part with a single strike on a sharp punch.  Alternative sources of 'jewel bearings' include Small Parts Inc, http://www.smallparts.com who will sell you a selection of nice ruby/ruby bearings for really disgusting prices.  (but theirs'll last much longer than even a TIG rod)  A jewel bearing will run 'lossless' at signifcantly higher rotational speed than a fluid bearing, and will tolerate a slight 'off level' condition better.  Balance, or course, is crucial, and the mass needs to be below the bearing for static stability.

 of course, you can combine the jewel bearing with the magnet. grind a piece of drill-rod to a nice, sharp, well-tapered point, and put a disc or cylinder magnet above same.   choose the magnet just strong strong enough support the total load for best performance. . .  you can fine tune things  by adding or subtracting mass from the drill rod near its base.  this bearing will wear the fastest, but will tolerate the most mis-alignment and unbalance.  you can even put another magnet/point combination at the bottom for better alignment and more tolerance of imbalance/ higher rotational speed. balancing weight and attraction forces is the trick, but this mode gives a little flexibility in positioning your suspended mass.

if nothing else, the above should get your brain fired up. . .

-Dan

[willib]

the load is suspended below the bearing and spins at 30 RPM

yes the load is balanced

 the bearing will be stationary , or part of it

i'm not sure the fluid bearing will spin fast enough?

and i'm not sure i could make the channel needed anyway

take a piece of sewing thread twist it in your fingers while keeping 144g tension on the thread pulling down , the bearing needs to spin free enough to release the tension of the thread

while supporting the wieght of the load.

magnetic bearings are tricky little buggers , because you need to attract and repell the bearing at the same time, otherwise your bearing sticks to one side and is not much use

one thing that had promiss was a ring magnet(microwave magnet) held in place by several 1.5"long x 1/2"dia neos, if i could keep it centered it would support the weight for sure,but i couldnt keep it centered

i have not tried this yet , but what about using three cylinder neos ,north pointing up and another cylinder north pointing down?

could they trap the upper 'bearing magnet'?

[maker of toys]

30 RPM on a water bearing is entirely achievable, given attention to balance, clearances and alignment. 300 rpm i'm not so sure of.

you're a creative person;  I've seen your other work. I'm sure you can beat functional fixedness if you think about it a bit.  here's a couple of freebees:  think nesting soup cans (steel ones) with axial tubes well sealed. modified funnels might be better bet, but the cans are free once you choke down the soup.  <G>

good mechanical cleaning  (scotchbrite) and liberal application of plumbing flux will enable you to solder steel with ordinary solder. take care not to heat the joint too much, but use a torch. an electric iron just won't have the cajones to make a good join in that much sheet metal.  or use epoxy. . .  

most hobby stores will have nesting brass tubing which would work for the center tubes.  so long as the center tubes are as a small a diameter as possible, reasonably centered (just so that the outer walls don't drag) and square with the bottoms you should be fine.

or, if you'd prefer a magnet, have another think about the magnetic needle idea;  ballasted properly, it will last many hours and has a desireable static stability characteristic. the key is to ballast the thing right up to the point where the upper magnet can barely hold the load; that way the bearing force is only a couple of grams or so.  when the thing does wear out, it's the work of moments to sharpen the needle/s and flip the magnet/s over. with some thought, some sort of wear shield could be provided for the magnet face. . . delrin or some such?

(my favorite lazy method of sharpening a needle out of drill rod:  cut a chunk of rod, file it roughly to shape and add any special geometry you want (holes, threads, snapring grooves, etc,) harden it, chuck it in a drill motor, fire up the bench grinder, and use the drill motor at a high speed to rotate the rod opposite the stone rotation while taking very light cuts.  finish with oiled emery paper on glass.  LIGHT CUTS and steady hands!)

I keep thinking Halbach array, but things need to be moving a bit before one will start to levitate.  maybe some combination of Halback array and static repulsion. . . I have vauge visions of a cone-shaped dingus with the halbach array around the base and the static repulsion section at the apex.  I'll think on it some and get back to you.  If it works at all, it'll be pricy in magnets. . .

I've rejected pumped bearings (like air bearings) as too power intense. . .

-Dan

[willib]

that Halbach array looks promising , do they stick together in the standard configuration? or are they held in place?

also i was thinking of coils and varible PWM to levitate a magnet array.

the amount of power required to do this isnt important , i have lots of recycled auto batteries  

[maker of toys]

I've never built a Halbach array into a rotating assembly; I'm sure it can be done, (the turbopumps at work use them) but I haven't got any guidence to give. the magnets for the linear ones MOST DEFINITELY need to be held in the correct alignment, which can be tricky to arrange if you've never tried it before.  see:  http://www.matchrockets.com/ether/halbach.html for a brief discussion of the idea.

 to build a circular array, I'd think you'd need to shape the magents a bit (wedge shape?) and there are some constraints on the size of the array because you need to build in complete modules.   Smaller magnets would give less granularity, and it seems like using several arrays arranged 'out of phase' (for lack of a better term) might lower the lift-off speed some more, but I don't think you'd ever get the lift-off low enough for the speeds you're contemplating.  I know the numbers didn't look promising last time I thought about something like this.  

(I 've thought of building an axial-flux driven mag-lev turntable, but I don't think I'll ever actually do it. . . I don't have enough vinyl to make the effort pay off.)

without knowing what you're building, I keep coming back to the magnetic needle bearing as the best compromise between construction effort, drag reduction and useful life.

-Dan

[willib]

Dan,can you be more specific on the construction of the magnetic needle bearing ?

say i use phenalic(sp)as the bearing seat.

and sharpen a drill bit to a point , how am i going to support the lower half ?

which way does the needle point?

i'm having a hard time visualizing this?

[maker of toys]

right.  look for a drawing sometime late tomorrow (PST) (sorry for the delay.  Playing in the shop.)

 beware:  the linen reinforced grades of phenolic plastic tend toward higher friction.  it does wear well, though.  do you have any UHMW polyethylene?  it's a good compromise between lubricity and wearing properties.  or perhaps better in this instance, teflon?

in the mean time, for a single magnet,  magnet goes on top, with a pole exposed on the downward face.  needle point in contact with exposed magnet pole (or wear shoe, if provided).  Load at bottom of needle, or suspended lower as desired.  (I remember something about thread?) magnet chosen to be just powerful enough to hold the loaded needle in contact with itself. greater magnet strength = higher friction forces and faster wear.

two magnet version, proceed as above, then add a second magnet below load, with another needle pointed at it.  space this magenet such that there is no actual contact between needle and magnet. . .  you're only providing a little additional stability to the system. . . damping harmonics, that sort of thing.  you want the load supported by the upper pole. the lower magnet should be substantially weaker than the upper to avoid detaching the load from the upper magnet.  (having the load supported on the lower magnet will work, but the supported structure must be more rigid, wear is rapid and friction is higher.)

if your structure is sufficiently rigid and the load sufficiently light, you can do a similar thing to the two magnet version with the axis horizontal.  useful for static balancing of light rotating assemblies  (like small turbocharger wheels, model airplane ducted fans, small props, etc.)  There are even better ways to balance turbine wheels, but this method gets them pretty close pretty fast.

one of these days I've got to see if the diamagnetic effect can be used for a light-duty bearing (fractions of a newton vertical load. . .)  Item 1,489(b) on my to-do list. . .  why do items seem to add themselves to the list faster than I can do them?

-Dan

[willib]

Dan if this is going to work it has to have zero friction, but i suppose that if it rubs a little then has zero friction ( between rubs) it might work

i once used a stiffer string made if nylon multiple single strands, the twisting friction built up very quickly , the sewing thread worked much better

picture if you will a motor spinning horizontally

this motor consists of a two pole rotor( four 1.5" x .5"dia  mags end to end) hanging from a string

when i pulse the motor it spins

said motor builds up twisting friction in the thread ( after  minuits), to the point where it will not work , hence the need for a frictionless bearing

[maker of toys]

well, I didn't draw this as a motor with magnets, etc., but this sort of bearing should allow the thread in your application to unwind with very little drag. . .  you could probably even go back to the monofilament if you needed/wanted to.

I hope this gets the basic idea across. . .

because of size constraints on the image, it's hard to see that there is a 1/4" airgap between the bottom magnet and its needle. . . that number is just made from whole cloth, of course. . .  use what works for you.

I also didn't provide wear shoes for the poles in this version. . . so add seasoning to taste.

I can e-mail you a larger image with pole shoes if you like.

-Dan

[willib]

i apologize for not responding sooner

the upper magnet holds the weight , hmmm

if it held all the weight with just a keeper needle in the bottom , i could pu thr string through the lower Keeper needle?

[maker of toys]

No worries on the time lag.  I'm not always Johnny Prompt myself. <G>

there's no obvious reason you couldn't run the thread through either (or both) needle/s

you really don't HAVE to have the lower needle/magnet setup; it just makes things track a little more accurately.

Have fun!

-Dan