Re: linear bearing based on ground flat stock

re: high speed.

You find plain/plane bearings more often than roller bearings in
high speed linear apps (talking linear motion).

The issue with them is that they require clearance and expansion
allowance. But used within those two constraints they are reliable
and well capable of HIGH speed.

Not to mention their ability to deal with short repetitive motion
which is bad for ball bearings.

Ballendo

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "Graham Stabler" <eexgs@...>
wrote:

>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, "ballendo" <ballendo@>

wrote:

> > Consider swarf in the equation...
>
> The three rails are in the "roof" of the machine, swarf will be

foam

> dust for this machine. Main purpose of the machine is to

say "look it

> really does work" and then get a simple parallel kinematics
> programming system done for the g-rex by the powers that be :)
>
> > A roller bearing on round rails will be far easier to use
> > successfully, IMO. However it seems you might want the added
> > constraint to torque rotation that you'd lose using a single

round

> > rail?`
>
> Yes I would need to use a pair which puts me off a little but I
> haven't ruled it out, manufacture is easy enough anyway and it

could

> be very light, the aluminium carriage with the two bevelled edges
> could be pocketed out substantially.
>
> > Oops. On re-read I see cheap AND light.
> >
> > The Zoltar plain bearing idea works well--and uses a flat bar as
> > you've suggested--and if your materials are chosen carefully, it

can

> > be used high speed. Very easy to mfr. as well. We use a version

of

> > it in our smallest gantry routers.
>
> Not convinced about the high speed part I really do mean high

speed.

> I showed an expert on parallel kinematic machines one of my 2-axis
> machines running with these belt drives, he thought they were

linear

> motors. For that I was using some tiny 9mm wide thomson rails,

just a

> single carriage on them but I only have two and they are a tad

short.

>
> Graham
>