PDA

View Full Version : OT sliding bushing without stiction



ikdor
02-16-2011, 08:28 AM
I need some advice from the mechanical inclined people here, as I only understand things that spark.

We're supporting a university project which is run by electrical engineers, who have even less knowledge of mechanical stuff. The system itself is comparable to an electric version of a hydraulic cylinder. The 40mm steel rod is riding in a glasfiber/epoxy bushing with a PTFE coating, but is suffering from stiction, messing up the control strategy. They added transformer oil to lubricate it and alleviate the stiction, as that is what they have laying around.....

I don't have any numbers yet for the surface loading of the bushing and surface roughness of the rod. But it's a turned finish and probably rougher than the 3um Ra recommended by the bushing data sheet.

Apart from adding a dithering signal to the position to solve this, are there better suitable materials for the bushing? I'd think something like PTFE loaded acetal or plain PTFE, but this is way out of my comfort zone.

Would polishing the rod make the stiction better or worse?

These are technology demonstrators and are not intended for mass production, so it doesn't need to be a best cost solution.
The current size of the bushing is about 40mm X 300mm.

Any ideas?
Igor

gda
02-16-2011, 08:34 AM
I've heard that one solution is to always have the shaft rotating, that makes the friction alway past static into kinetic - Not sure if the complications of this merits it.

Or try an air bearing like they have on end mill grinders.

ikdor
02-16-2011, 08:40 AM
The rod itself is a linear motor, and thus wired. So no rotating.

The system also doesn't have pressurised air, and power consumption will skyrocket when adding a compressor.

Thanks for your thoughts,
Igor

Rustybolt
02-16-2011, 08:54 AM
Yes, polishing the rod will cause it to slide easier.(double entendre)
But if you look at an actual hydraulic cylinder rod you'll notice that it is chromed and highly polished. That is the finnish you're looking for.
On any close fit bearing, like a rod bearing, any imperfections on the rod will cause problems with the intended action of getting the rod to slide smoothly.
I would reccomend getting a piece of polished hyraulic rod.

ikdor
02-16-2011, 09:01 AM
The rod is actually a stack of steel discs with aluminium shims to isolate the magnetic phases. This will make it hard to use a hydraulic rod.
They tried coating it earlier with something, but the baking of the coating burned the isolation on the coils....

Would chrome plating also adhere well to the aluminium shims, or would it break off there and make a big mess of things?

Igor

Boucher
02-16-2011, 09:16 AM
There is a graphite impregnated teflon that could be used to make the bushings that is significantly slicker than other bushings.

Peter.
02-16-2011, 09:24 AM
The rod is actually a stack of steel discs with aluminium shims to isolate the magnetic phases. This will make it hard to use a hydraulic rod.
They tried coating it earlier with something, but the baking of the coating burned the isolation on the coils....

Would chrome plating also adhere well to the aluminium shims, or would it break off there and make a big mess of things?

Igor

Why not shrink a very thin liner over the rod?

dian
02-16-2011, 09:37 AM
there are bushings called fiberglide. they are ptfe with some fibres on steel and advertise a friction coeficient of as low as 0.02 and 210 n/mm. but since you have a ptfe coating already?

gvasale
02-16-2011, 09:45 AM
have you looked at Thomson Ball Bushings? http://www.thomsonlinear.com/website/com/eng/products/linear_guides/linear_ball_bushing_bearings.php

RussZHC
02-16-2011, 10:11 AM
The rod is actually a stack of steel discs with aluminium shims to isolate the magnetic phases.


"This" I feel is the issue. This series of alternating (?, not stated but inferred) discs is in effect a series of edges (as opposed to what a solid hydraulic rod would be, for example, a "single" smooth surface) and would, logically cause a lot of stiction.
How accurate are those discs to each of the same type but also to the other type?
The only way off the top of my head I can see to get them all as smooth as possible would be to set them up in the order they are being used on a "dummy" central shaft, some how apply a clamping force on the ends (clamping the stack of discs tightly) and then turning them to the finished size you need...as there will likely be some very very small amount of taper you would need to reassemble the final version in the same disc order (so number them first ?)

ikdor
02-16-2011, 10:34 AM
The stack of alternating discs is clamped and then turned on the lathe as one unit, making it reasonably smooth. But nowhere near a chrome plated hydraulic rod. The stack could be polished afterwards.
I wonder what it looks like though when it heats up, as the aluminium shims would expand more and might protrude slightly between the steel discs. I'll do a back of the envelope on that.

Shrinking a liner over it sounds interesting. Radial space is at a minimum though, so it would be really exciting getting a thin liner over the rod without binding halfway down......
Now that I think about it; as the assembly is not rotating, a slit liner would even be possible. No exciting shrinking actions then.

The ball bushings will take up too much space, they will displace magnets around the bushing, reducing the stroke.

Thanks for the tips on the other materials, I'll have a look at them tonight.

Igor

strokersix
02-16-2011, 12:29 PM
Any way you could use a ball bushing rolling on a guide rod inside the stack? Clearance on the outside of the stack instead of your current sliding contact.

Can't quite visualize your geometry so maybe not helpful.

becksmachine
02-16-2011, 01:09 PM
There are oils that are formulated to help reduce or eliminate stiction, some of them work. ;)

I have had good luck with a Kendall product made for use in limited slip differentials. Started using it on a Lodge & Shipley lathe I have with reground, hard ways. The carriage would move smooth as silk in increments of 1/4"- 6', but it was nearly impossible to move it .001", even by bumping. Started using that oil and even .0001 became possible. Can't remember the name of it, if you were interested I would get that for you?

Dave

Rustybolt
02-16-2011, 01:47 PM
Assemble the rod and have it centerless ground. It would be better if all the parts can't rotate individually. Like all glued together.

ikdor
02-16-2011, 04:16 PM
Dave, I'm definitely interested, if they can't use it because of compatibility issues it would still be a good bit of trivia to know. (especially now that I have all the parts in house for my submicron poor mans DRO)

With regards to the centreless grinding, that might be a good one. The discs won't rotate as the stack is permanently clamped with a few tonnes.

I'll shoot a video of the application next time I'm there ;)

Thanks,
Igor

rohart
02-16-2011, 05:02 PM
Can yoiu describe the application in a little more detail ?

Is the rod (stack of discs) horizontal or vertical ? If vertical, can the design be changed for vertical operation ?

Does it carry weight ?

Are the ends linked to other components, and if so, at what angles ?

Basically, I'm after reducing the forces between the rod and the sleeve.

If you don't want internal ball bushings, and you were prepared to go to a hexagonal cross-section, three ball bearings mounted externally at 120 degrees make a good bearing.

noah katz
02-16-2011, 05:09 PM
What is the rod velocity and how long does it have to run?

If it would last, oil or grease would be the way to go, as besides reducing or eliminating metal to metal contact, it would add damping - your control system will love it :)

Black_Moons
02-16-2011, 05:32 PM
Id give way oil a shot. Its designed to eliminate stiction, And its VERY thick and tacky so it does not fall off vertical surfaces.

becksmachine
02-16-2011, 06:13 PM
Dave, I'm definitely interested, if they can't use it because of compatibility issues it would still be a good bit of trivia to know. (especially now that I have all the parts in house for my submicron poor mans DRO)
Thanks,
Igor

Ok, so I looked on the bottle for a specific name, it says "Kendall Special Limited Slip Axle Gear Lube". Go figure. ;) Part #527-7737 in quart bottles. Doing a Google search also returns hits for "limited slip additive", that might work as well.

The stuff I have is 80W-90, don't know if it is available in other viscosities.

Dave

ckelloug
02-16-2011, 06:27 PM
Take a look at www.igus.com and see if anything there might help.

ikdor
02-16-2011, 06:34 PM
Sorry for being secretive about the application, I don't know yet whether divulging the application will get me into trouble :rolleyes:

The design is vertical, not carrying weight, but could be subjected to a few hundred pounds of side load. Engagement of the bushing and the rod is always the full 300mm. It's mainly for demo's, so 100 hours should be fine. We could always relube.
The wayoil does seem like a good idea as well.

Thanks Dave, I'll go and order a bottle.

Igor

Weston Bye
02-16-2011, 09:33 PM
I would run the stack inside a thin brass tube... unless the eddy currents that would occur within the brass would degrade the performance. Where I work, we build solenoids and where cost permits, we use brass tubes for the armatures to slide in. This provides the smoothest, quietest operation.

As for the aforementioned eddy currents, it is an interesting experiment to drop a magnet (neodymium works best) down a long brass tube, then drop a similar size and weight of unmagnetized metal down the same tube. The speed of travel for the magnet is remarkably slower, as the motion of the magnet induces a localized loop of current in the tube, forming a temporary electromagnet that reacts with the permanent magnet, slowing it down. The energy in the motion of the magnet is ultimately converted into heat along the length of the tube and dissipated away.

bollie7
02-16-2011, 09:57 PM
why not try polishing the rod and the bore and giving both a good coat of a dry film lub or spray silicon. If that doesn't help it really hasn't cost much and if it does work its a cheap fix.

bollie7

randyc
02-16-2011, 10:29 PM
I like your dithering idea, if it is possible. Regarding materials, the best mechanisms that I've personally used to minimize initial friction were graphite-on-graphite. Yes, polishing the interface surface will help.

Good luck !

Black_Moons
02-16-2011, 10:51 PM
As for the aforementioned eddy currents, it is an interesting experiment to drop a magnet (neodymium works best) down a long brass tube, then drop a similar size and weight of unmagnetized metal down the same tube.

Thats nothing, Find a solid chunk of aluminum. tilt at say 70 degrees and let a big neo magnet like from a hard drive slide down the side, It will creep.

If you secure the aluminum and try and pass the magnet over by hand, you can definately feel speed dependant resistance to movement.

Iv seen demos of magnets beween two large chunks of aluminum that looked like a rock falling in molasses.

J Tiers
02-16-2011, 11:21 PM
Applications?

he says motor, and that might be correct, depending on the actual detail setup it might even be an induction motor. There are some other types that might be possible if I think about it some. But it might also be a linear generator. Most any electric motor can be a generator, PM, DC, induction AC, synchronous AC.......

Putting Netherlands and linear generator together gets only one thing coming out of "watson" for that.... wave power generator. That would be about nine times more intelligent than the use of racks and pinions or other schemes for converting linear motion to rotary.

But "stiction" might not play a great role in that type application, so this may more likely be a motor, for some totally other thing. Since it has a control loop (the generator would also) it may be a servo positioner.

ulav8r
02-16-2011, 11:59 PM
Could the bearings be mounted separately and not have anything touching the motor rods? No matter how smooth a finish you get on the disc stack, it will change with temperature changes and the "stiction" problem willl still be there.

darryl
02-17-2011, 01:54 AM
As I'm reading all this, I keep thinking about slitting a chrome plated tubing in half lengthwise, then closing that around the stack of discs. Epoxy might be your friend in this case. There are several examples of chrome plated tubing in plumbing departments- usually brass.

It might be possible to do this 'inside out'. If you can find teflon tubing, you could slit that and wrap it around the stack of discs. It would then ride in a metal bushing.

J Tiers
02-17-2011, 02:00 AM
I think ANY enclosing conductive material is going to be a non-starter.... it should provide eddy current losses that the EEs will not like.

jkopel
02-17-2011, 02:12 AM
Why not reverse the order of things and put the rod inside a thin tube of PTFE? Then you could use a polished bushing (and proper lubricant).

J Tiers
02-17-2011, 09:48 AM
Maybe the eddy current losses are not so bad...... re-reading, I find that the rod is wired and has coils. Depends if the magnetic field changes, either from external induction into the coils or from variations as a static field established by the coils passes external poles.

Vertical, motor, but not 'carrying weight"... several hundred lb side force..... wave powered generator is looking more likely.

ikdor
02-17-2011, 10:52 AM
@Weston
The brass tube could be a nice solution, I'll keep that in mind. The friction is a bit higher but it should be stronger than a PTFE liner.

@bollie
Giving it a spray of dry lube is what I also recommended. However they're making a completely new one, so there is an opportunity to improve the design.

@J Tiers
Very creative, but it's not for waves. I'll hint them the application though as it might be a good subsidy sponge project to follow up :D

@ulav8r
The application is very space limited, so unfortunately no bearings outside the stack.

Putting some ptfe/chrome tube onto the stack and have that ride in a smooth bore might be a good solution. Unfortunately the application is also very limited in radial direction, so each mm added will reduce the rated power. This because the stack diameter is reduced and the magnetic gap is increased.
This made the glass fiber / ptfe tube a nice solution as there was only one part between the stack and the magnets, doubling as both bearing and bore.
Now if they only put more ptfe in there..... I think they just dipped it and called it good.

Igor