"metric" ballscrews? , and other misc. was Re: Pulse Gen

Alan wrote:
<snip> and a few pairs of what I think are 8tpi. x 3/8". I think

>they may possibly be metric, as the shaft end was not a simple
>English dimension, and some of the 3/16" (?!) hardware associated >
>with them turned out to be metric!

Alan, Keep in mind that MOST bearings are metric, so the leadscrew
ends would be sized to fit. Also that a lot of "motion control stuff
comes from Germany, Italy, and other European countries, accounting
for the hardware.

8tpi is .125; 3mm is .118, which is 8.46tpi. I'd guess its 8tpi
imperial!

>200 IPM! In my current mindset, 20 IPM is fast!

And the signmakers, plastic fabricators and woodworkers would like to
see 600ipm rapids! This is one problem with CncPRO.
It places "artificial" limits on the top speed attainable. It seems
to change its' "opinion" of "top speed" based on order of entry of
parameters; or something! It's very frustrating to KNOW the machine,
computer, and drives can go faster; others have said CncPRO will put
out 20K steps/sec, but because of something internal to the program,
it "decides" on a ridiculously low top speed... Doug???

>What steps/in should I be shooting for? And what's realistic?

5tpi,400spr(1/2 step)or 5tpi,200spr(fullstep) 2:1 ratio, is/was an
industry standard; so 2000 spi (.0005 increment) will cover a lot of
older machines.

You want your SPI to be DOUBLE your smallest expected accuracy need,
as a minimum. So the drive setup in the paragraph above should be
good for a plus or minus .001 system. Four times smallest expected
increment will suffice for MOST systems. In reality, as long as all
parts of the system can handle it, more is better.

A fun exercise is to draw some circles, lines, and arbitrary shapes
on graph paper. Now using ONLY the "points" where the lines cross,
duplicate the shapes you just drew as nearly as possible. This IS
what the CNC control is doing! You may find that a "circle" is an
octogon! Or even a square! When you then realise that each square is
one "step", or "least increment", you start to see the CNC thing a
little differently, IMO. Try a VERY shallow taper! or "mill" a 9/32
hole with a 1/4 inch endmill...

The real limit to this has been the programming side of things. Until
recently, 16 bit long integers were the best compromise of size/speed.
Since 2 to the 16th power is 65536, with a 2000 spi system, you ran
out of travel at about 32 inches of axis travel. The new "magic
number is 2^32 or about 4.3 billion. If we start in the "middle" of
this number to allow for positive and negative travel, we have about
2.14 billion "steps" available.

So, take your desired travel and multiply by your desired spi, and
you'll see what sort of calcs the computer is going to need to do.
Next look at your desired top speed "times" your spi, divide by 60
and you'll see how many steps/second you need from the pulse train.

Find the "balancing point" between the two equations (based on your
system and programming needs/capabilities) and you'll have your
answer to how many steps per inch "should" I have...

Hope this helps. Keep in mind it's just one persons' thoughts. As in
most CNC issues, there are many paths to fulfillment...

Ballendo