re:Re: Pre load specification (leadscrew bearings)

Ron,

snips/answers contained below

>Does one specify a numeric value for the pre-load when specifying
>the bearing? Or, does one simply specify "pre-loaded matched set",
>or similar, and expect to get a bearing set of a standard pre-load
>value?

Unless you have a BIG budget, you pick from "standard available pre-
loads"(in the bore size and overall load capacity you need). Or you
specify zero clearance and do the calcs and shim/machine to what you
need.

>If one is to consider the bearing pre-load in the design, how is
>this done? With known design loads on the bearing installation, how
>is the preload incorporated into the analysis? Is there a
>quantified methodology to relate this desired preload to the torque
>of the lead screw retention nut? Something more exact than "a
>couple of grunts" on the nut.

What you are looking for is the smallest pre-load which is not
exceeded by your machine dynamics. The leadscrew is "pushing"
and "pulling" on the bearings as the machine does its' thing. The
effect of this is to turn the spherical bearings into ovals. When
they turn into ovals(between the bearing races) they "take up" less
space.(the pre-load is reduced. Or the zero clearance becomes non-
zero)So if you have a bearing pre-loaded to 30 lbs; then any force
less than(or equal to) thirty pounds will still be zero clearance. As
you exceed thirty pounds of force, the zero clearance is no longer
guaranteed, and you may have backlash when the direction of force
changes.

I'm really trying to keep this from becoming a bunch of engineering
calcs, for two reasons:

1) When/If you get to that point you really should be talking to the
bearing/leadscrew people.

2) It makes it seem harder than it is. If you really want to explore
the "quantified methodology" of this, it's more work than it needs to
be. A coupla grunts, and watching the results, works!

The nut puts the leadscrew in tension, in effect creating a pre-load
on the inner races of the bearings. As long as the forces applied (by
the machine in operation) do not exceed the "stretch-ability" of the
leadscrew between the bearings, the point is moot.(doesn't matter) In
practice, there IS some calc-defined "torque value" that achieves
this "squeeze" on the bearings. But a coupla grunts usually achieves
what is necessary with no ill effects.(watch the results; temperature
has been mentioned as an indicator) And relax!

If you're designing a machine for a wafer foundry (micron and sub-
micron positioning), then you maybe need to go all this way.
Otherwise ya don't, IMO.

Hope this helps.

Ballendo

P.S. Yes, I know. The ball races deform also, but I'm TRYING to keep
this simple.