Re: Lead screw/Guide rail install/pre-load (machine design thoughts)

Ron,

snips/thoughts/answers interspersed below:

>I'm doing what is essentially a 3 axis router: 48" x 36" x 18" of
>cutter travel. It'll have the z-y plane fixed in x, with a
>travelling x table moving below the y-z plane. So it's a little
>goofy in that the part moves only in x, while the router moves in y-
>z.

This is a very common "bridge" design. Not goofy at all! It IS a good
choice since you want 18" of Z axis. With this design the Y and Z
axis is stiffer than a gantry style.

The disadvantage is that you need TWICE the space to operate(2 times
the X axis travel length), and the material costs will be higher.
Also the mass of the "moving table" load will change, depending on
what work is loaded on the machine. This means your drive system will
need to be set up to handle the expected material load. And supports
for long work will need to accomodate this "movement".

>Because I want to maintain some decent accuracy, I'll be using a
>linear strip encoder to close the feedback loop for the long axis.
>This should allow me to simply drive the table to the correct
>position and not be concerned with the accuracy of the lead screws.

This is fine, but way overkill for your "stated" application. Rolled
thread leadscrews(ball and acme) typically have an .003/foot
accuracy. And it is generally non-cumulative. Your MATERIAL(pvc) is
going to "move" much more than your machine accuracy!

Unless you already have 'em(the linear encoders), think about none or
rotary style,IMO. EMC now has leadscrew compensation (just like the
big boys) so you would be able to "dial-in" YOUR screws. If you use
EMC.

You won't need the RIGID mounts we have been talking about for this
machine. You didn't mention travel speed, but the industry std. 5/8-
.200 ballscrews will be plenty rigid for these travels at nearly any
speed you might realistically want to attain. I'd suggest a FIXED-
SIMPLE for all axes.(either angular-contact or my "cheap" radial
bearing version)

Be sure to use a GOOD Y and Z linear bearing system. Y, because it's
the mounting surface for the Z; and Z because you are dealing with a
LOT of leverage when you extend out 18 inches in Z!!!

Ron, Let me ask you a question. Do you REALLY NEED 18 inches of Z
travel???

Very often, the principle I spoke of (earlier post) comes into play
while we are deciding what size machine to make. "I'm not sure, so
I'll make it big enough for anything!..." This sort of thinking can
lead to failure as the project "snowballs" into problems. Or
paralysis by analysis.

If you are using 10 inch PVC tubing, you NEED 10 inches of travel, at
least, right? So why not 18??

On the other hand: Maybe you need at least 5 inches of Z travel
(probably 6 or 8 inches), and a BRIDGE that adjusts up and down to
accomodate the different material thicknesses!

The resulting machine will be MUCH stiffer, cost MUCH less, hold
better accuracy, and still do everything you expect.

Keep in mind that most of the tooling you will probably use is about
four inches long (or less, router bits for plastic max out at about
three inches, unless a specialty product) and having 18 inches of
travel will not really let you cut any deeper as the bit/machine
spindle interference limits the cutting operation.

Now I realize I don't have ALL the info. And I may be wrong. But...

I've watched this process enough times to know that someone on this
list will benefit from what I've written.

Ballendo

P.S. While I'm in this "protect us all from ourselves" rant:

The other thing that seems to come up a lot is the "overstated
accuracy need".(we've been talking about the "overstated structural
need")

Keep in mind that machine shops (job shops) typically spec work at
plus or minus .005(location). And they have PRECISION TOOLS! Cast
Iron. Expensive. Temperature controlled ballscrews!(I'm NOT saying
they can't do better. I'm saying they realize the "real world" needs.
Which are GENERALLY much less stringent than first thought.)

If you read a trade publication like MACHINE DESIGN, or METALWORKING
NEWS long enough, you'll eventually see an article about accuracy of
machine tools. And when you start to realise how much the MATERIAL
is "moving", and even the machine parts, you start to think
differently. And then you'll look at this lower-end market and
realise that the part is STILL made by the OPERATOR(or programmer, in
our cnc case). And that a skilled OPERATOR can make excellent parts
from a "poorer" machine. And that you have to throw a LOT of money at
the problem to get to the point where the machine ability REALLY
overtakes the operator ability(remember, I'm talking ACCURACY here).

You'll see that repeatability is much easier to achieve. And a
repeatable machine can be "programmed into accuracy". And the cost in
time, money, and sanity, is much less. <ballendo rant mode off>