Re: tensioned ballscrew redux

Doug,

I'll be VERY curious to hear how your "extra" support for the screw
works out... (I'm skeptical)

Let's see, Les suggests simple-simple end mounts (with bellevilles on
one end for expansion, no less!), and you've got two "moving" screw
supports... Is the screw direct drive? Or will a belts' sideways
force be an additional factor? What rapid traverse rate do you intend
to use?

Re: the supports
Is the outer diameter of the screw consistent? Will the "screwing
action" wear out your bushings in the supports? Or will your supports
wear out the screw? Either way, as the wear increases clearances,
will your supports still be doing their job? Have you considered the
localised heating at the supports due to friction?

Re: spinning the nut. I'd still choose this! Thousands are in
operation everyday. I've never heard the concerns Les expressed
before (but machines ARE getting faster, so maybe it's becoming more
prevalent; since so many machines DO have spinning nuts), and your
intended speed may obviate those concerns anyway...

I don't think the "nut spinning" transmission is any more difficult,
expensive or exacting than the endmounts plus traveling supports
you're entertaining...

Should be interesting,

Ballendo

P.S. I'd SURE use RIGID on at least ONE end. But then, I'd just spin
the nut, so I wouldn't need to...

--- In CAD_CAM_EDM_DRO@y..., "Les Watts" <leswatts@r...> wrote:
> Hey!
>
> I'm kinda waiting for the ups morning delivery right now
> so I can do a couple calculations here.
>
> Ok I will consider back to back AC or tapered roller bearing
> tensioning a long screw using the screw's own elastic properties.
This
> implies that EI for its support is much higher.
>
> First thing to note is the block-and tackle driven support blocks
> pretty much eliminate buckling problems. A great idea, wot. ;^)
>
> OK now temperature.
>
> I will use 12x10^-6 c-1 as the thermal expansion coefficient
> of steel.
>
> For 120" we get 1.44 thousandths (0.00144") for every degree of
delta t of
> the screw relative to a fixed support.
>
> Lets say it gets warm to the touch in use. Actually they often
> get quite warm. But let's say 10c. So that gives 14 thousandths of
expansion
> of the screw.
>
> Now let's consider the screw tensioning level. 5% max dynamic
preload is
> common in bearings and is usually refered
> to as light preload. For a bearing of about the right size for
> that screw it might be about 300 lb for angular and higher
> for tapered roller... don't have the TR data but let's say
> 500 lb. Too high and it will have a lot of drag. (I use 300).
>
> Now how much do we have to stretch that screw to get
> 500 lb tension? Well that is FL/EA. The effective area of the screw
with the
> lands is about 0.8 in^2 so we get 0.0025".
>
> Uh oh. There's a problem. A 500 lb preloaded screw
> will unload completely with only a 1.7 degree c temperature
> rise in the screw. Let me grab Slocum's Precision Machine
> Design off the shelf. That's my bible.
>
> Yes there is some on the subject.... he does not reccomend
> it. Says it's thermally unstable. Maybe I did the calcs without
> arithmetic errors.
>
> But tensioning the ballscrew is a good thing for the reasons you
mentioned.
> So do it with a preload spring. Something
> that can maintain the force over a larger displacement.
> I use stacked bellville washers.
>
> Per Slocum I would use simple-simple end fixity with two
> bearings preloaded to each other on both ends. One set
> is allowed to slide in it's housing a little.
>
> Buckling load is inverse squared to length and with the
> travelling support I don't think you need fixed end support.
> Slocum mentions the great wear and accuracy advantages of simple
support. He
> likes kinematic construction.
>
> Oh the other thing to check- torsional loads. I had no problems
> with 70"x1" screws but at some point it will be significant.
> I am referring to errors generated from twisting the screw.
>
> Let's see.. I need shear modulus G for steel 12x10^6 psi.
> Polar moment of area J for a 1" screw is about pi/32
> or 0.1 in^4.
>
> Let's have torque at 500 in oz. That is in the middle of the range
for
> machines of our size. converted to in lb that would
> be 31.25.
>
> So worst case twist isTL/JG or 0.003125 radians.
> That's 0.18 degrees. Not too bad. Seems a little small
> though. Check this. Torsional resonance could magnify it.
> A slender ratio of 120 is the outer limits of ballscrew use!
>
> Oh since I decided to post this to the list rather that private
> e-mail I want to mention rotating nuts.
>
> Why not just rotate the nut?
>
> I can say why I don't.
>
> Rotating the nut has much much lower inertia than rotating the
> screw. There are no rotating end bearings. Critical speed
> issues diminish.
>
> Here is my problem: Like many I often get parts surplus new.
> Many ballnuts can have ball scuffing and lube starvation issues if
rotated
> at high speed I'm told. Manufacturers make rotating nut certified
units
> (internal recirculation helps but is no guarantee) but that is not
what I
> use due to cost. I don't have
> the option to test a unit to failure to see how it does so I
> take the conservative route and use the units as specified
> by the manufacturer.
>
> Also- the rotating nut transmission is an expensive bit of
> precision engineered components- more costly than end bearings.
>
> OK time to run out and build up some little rs-485 differential line
> drivers. I am using TI SN75174. Looks good- has enable,
> thermal protection and 3-state outputs.
>
> Later
>
> Les
>
>
>
> Leslie Watts
> L M Watts Furniture
> Tiger, Georgia USA
> http://www.rabun.net/~leswatts/wattsfurniturewp.html
> ----- Original Message -----
> From: "Doug Harrison" <prototype@c...>
> To: <CAD_CAM_EDM_DRO@y...>
> Sent: Monday, February 18, 2002 9:19 PM
> Subject: [CAD_CAM_EDM_DRO] tensioned ballscrew
>
>
> > I have two ideas that need to be challenged. The screws on the X
axis of
> my
> > router are ten feet long. They are 1" diameter, .25 pitch from
Rockford.
> >
> > 1. We have come up with a way to provide intermediate support so
as to
> get
> > around critical frequency problems. An extra pair of blocks on
each
> linear
> > rail will carry support bushings. These blocks are connected to
each
> other
> > through a beam about five feet long. This beam is tied to the
gantry
> > through a 2:1 cable arrangement so that the support blocks move
at half
> the
> > speed of the gantry. Thus, the worst case support is when the
gantry is
> at
> > one end and one support block is in the middle of the screw.
> >
> > 2. Now for the end bearing blocks. I want to use a tapered
roller
> bearing
> > at each end of the screw and tension the screw between them.
This poses
> two
> > possible advantages. First, by tensioning the screw I can
further raise
> the
> > critical velocity (though I've been out of college too long to
remember
> how
> > to calculate just how much). Second, by varying the tension in
the screw
> I
> > should be able to adjust out any first order pitch error (though
not
> second
> > order or cyclic errors).
> >
> > Of course it would be best to have two bearings at each end of
the screw,
> so
> > I propose putting a regular ball bearing about two inches
outboard of each
> > tapered bearing in the bearing block. This would take up radial
loading
> and
> > thus improve end fixity of the screw.
> >
> > I welcome y'alls opinions.
> >
> > Doug
> >
> >
> > Addresses:
> > FAQ: http://www.ktmarketing.com/faq.html
> > FILES: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/files/
> >
> > Post messages: CAD_CAM_EDM_DRO@y...
> > Subscribe: CAD_CAM_EDM_DRO-subscribe@y...
> > Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@y...
> > List owner: CAD_CAM_EDM_DRO-owner@y..., wanliker@a...
> > Moderator: jmelson@a... timg@k... [Moderator]
> > URL to this page: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO
> > bill,
> > List Manager
> >
> >
> >
> > Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
> >
> >
> >