Re: [CAD_CAM_EDM_DRO] Re: Timing Pulley Accruacy and Servo shaft runout

Polaraligned wrote:

>
>
>
>>5 thou ? on a new motor ? I shudder to think that is even possible !
>>
>>I would think that is scrap from the trash bin of the QC department.
>>or it fell on the floor from the top of the rack.
>>
>>Dave
>>
>>
>
>
>Well Dave, I retested the motors. The worst was .005" the best was
>.0002". The best was a new Ametek motor.
>Many were .002" runout (I tested 5 Baldors)and these new Baldor motors
>came to me in a sealed factory box. The motor was shrunk wrapped to a
>plywood bottom.
>Boxes are in fine shape, no damage here.
>

In that case, I suspect Baldor has taken a shortcut in their process.
Most likely
no final grind after heat treating the shafts, or possibly something
going wrong
in the process of assembling the armature/rotor onto the shaft that is
bending
them. The Ametek is better than one might expect, anything up to about
.001"
would normally be considered acceptable in most cases.

> The Ametek that tested
>so nice has a short 1.5" shaft that appears well machined.
>The Baldor servos have longer 2.25" shafts that don't have nearly
>as nice a finish as the Ametek servos. But the runout is there
>regardless of the surface finish. The .002" motors are the ones
>I plan on using on my machine and I am struggling to see if that
>runout can cause any significant error.
>

It certainly will not.

> I would think that as
>the belt gets tighter and looser the teeth engage a bit deeper, or
>a bit shallower and that error would be a function of the tooth and
>how deep it is going into the pulley.
>

No, the tooth engagement doesn't affect accuracy much. And, the .001"
eccentricity
will have so tiny an effect on your system accuracy as to be only
measurable in a
metrology lab (6.33 x 10^-5 ")

> I would imagine that a spring
>loaded tensioner would eliminate the runout as a source of error.
>
>

A spring-loaded tensioner won't work, as the tensioner will get lifted
up when
the motor accelerates in one direction, and possibly overload the side it is
on when the motor accelerates the other way. Do not use that. A fixed
tensioner,
maybe with an eccentric post, is what you want.

How it would eliminate runout as a source of error, I have no idea.
Specifically
because the belt and pulley have teeth, then tooth engagement is
cancelled out.
But, the eccentricity is still there, no matter WHAT you do about belt
tension.
And, so, smooth rotation of the motor will cause a faster/slower
rotation of the
leadscrew.

Why are you quibbling over ten to the minus 5 inch errors when you have
so many
other LARGE errors still in the system? You'd be AMAZED at all the many
sources
of errors that add up in these machines! Thermal errors,
non-orthogonality, bowing
of leadscrews, axial thrust bearings, flexing of the machine itself, and
on and on.

Jon