Low friction linear slide

[rajesh.poddar]

I have an application for which I need a very low friction linear slide (less than 10g (.1N) of friction). For this application, accuracy/precision are not important at all, so there can be plenty of play between the carriage and the slide and lots of backlash. I quickly put together standard makerslide carriage with 2 fixed dual V wheels and 1 eccentric dual V wheel. However, the friction seems more on the order of 30-40g. I was wondering if people have suggestions for how to lower the friction.

With just 1 bearing pressed onto the dual V wheel, the wheel seems to roll quite smoothly, but with two bearings pressed in, the friction seems to become noticeably higher. Also, I noticed that the fit between the M5 screw that passes through the wheel and the precision shim washer is very tight - hence it requires quite a bit of force to push an M5 screw through a dual V wheel assembly. Often removing the M5 screw from the wheel manages to pop out one of the bearings. Does this seem true for other people as well or am I doing something wrong? Would getting different bearings help in lowering the friction?

Thanks
Rajesh

[butterfingers]

"Out of left field" suggestion:

http://www.instructables.com/id/Rolamit ... s-bearing/

[Zat German]

I think you might want to check the size of your bolts. I have a V wheel on a M5 bolt right here and it spins quite easily. I also have some Makerslide and one of the generic carriages and it moves quite easily as well. By adjusting the eccentric spacer, I can control how much play or how easily it moves on the Makerslide. With just a slight turn, the carriage moves by barely tilting the Makerslide but that creates a slight play between the wheels and the Makerslide(something which it sounds like wouldn't matter to you, but does for something like an ORD Bot).

[bdring]

The grease in most standard bearings will increase the "friction" past what you are looking for. With the dual bearing V wheels that adds up quickly.

If the load and speed is not an issue you might look at an open bearing with less lubrication. VXB sells a lot of bearings including ceramic ones.

You might even try pulling out the seals of a standard bearing and washing them out with a solvent.

[BenJackson]

The ultimate in low friction is http://en.wikipedia.org/wiki/Air_track

[rajesh.poddar]

Thank you everyone for the suggestions. My M5 bolts (from the shapeoko kit) were indeed slightly bigger than standard ones (perhaps a bad batch?) and hence things weren't fitting quite right. Following Bart's suggestion, I pulled out the seals in the standard bearings and washed them in soap water and removed all the grease and water using pressurized air. This made a substantial reduction on the friction.

For anybody who is interested, here are the results of my friction experiments. For all friction measurements I used a 3 wheel carriage that weighed 130g. I slid the carriage onto a horizontal piece of makerslide such that all 3 wheels were on the horizontal plane. Then I pulled the slide with weights via a pulley and a string. The minimum weight that caused the slide to move was noted as the 'friction'. I tightened the eccentric nut until all 3 wheels were rotating rather than sliding on the v rails. With this configuration there was still decent amounts of play between the rail and the carriage.

With 2 standard bearings per wheel X 3 wheels I measured a frictional force of about 60g. With just 1 bearing per wheel this went down to about 30g. After degreasing the standard bearings, the friction went down to 6g.

[dzach]

This is interesting, thank you for the data.

This data would give the following table for the friction coefficient, wouldn't it?:

Code: Select all

V wheel friction coefficient (static), μ = Ff/Fn
----------------------------------------------------
Double bearings:   60g/130g/3 = 0.154
Single bearings:   30g/130g/3 = 0.0769
Single bearing, no grease   = 6g/130g/3 = 0.0154


( Friction in Wikipedia )
EDIT: This should be considered under the given conditions of the measurements, as described above.

[r691175002]

The friction coefficient of the bearings is probably a lot more complicated than that.

A large portion of the force on each bearing is the preload holding them against the v-rail (so the total normal force will be greater than 130g and very difficult to measure). As well, the coefficient of friction in the bearings/v-wheels will vary based on whether the load is axial or radial.

[dzach]

The friction coefficient of the bearings is probably a lot more complicated than that.

Sure, but does this influence the measured data?

I think one should also take into consideration the friction introduced by the string and pulley used in the measurement apparatus, which should be measured separately and subtracted from the values above.