[Coco] [Color Computer] [coco] Coco CNC

Gene Heskett gene.heskett at verizon.net
Mon Feb 12 02:21:43 EST 2007


On Monday 12 February 2007 00:55, George's Coco Address wrote:
>Okay!
>
> I've been working on my Coco powered "Tiny CNC" machine this weekend.
>
> Whew!
>
> What a task! I discovered that the axis designations were all messed
> up.
>
> (You learn a LOT when working in a machine shop)
>
> All three were wrong. X, Y and Z were wrong. So I renamed them and in
> the process, I lost the driver(B09 driver) for the REAL Y axis.
>
> No problem, I thought. Just rename the X or Z axis driver. Well, it
> didn't work. After two days, I discover several problem with my
> hardware that proved that the port for the Y axis was wired
> incorrectly, the cable that connected it was also wired incorectly and
> the software to correct this error is now lost.
>
> No matter about the software. I repaired my mistakes with the hardware
> and now the software to drive the servos works correctly. I can swap
> the cables to each of the servos and it works correctly.
>
> Since I've been working at a machine shop, I was forced to buy some
>precision instruments such as a dial caliper. My old vernier caliper was
> not acceptable at work.
>
> Hmm. I really can't see any accuracy difference except for the fact
> that the dial is easier to read.
>
> Anyway..... Math always works......
>
>
> All three lead screws are 32 TPI. The three servos are 200 steps per
> inch. Doing the math, this comes to 6400 steps per inch, or precision
> to .00015625.

Humm, ignoring the fact that the screws aren't that great, I have for x & 
y, 16 tpi screws, and 200 step motors that are being microstepped at 8x, 
or 1600 steps/rev.  That is .0000390625" per step.  That's optical 
wavelength territory, and certainly wishfull thinking on a $300 
micromill.  z isn't quite that ridiculous as its a 10 tpi screw.  Typical 
backlash is 5 or 7 thou, correctable in the configuration with a lookup 
table that allows one to fully characterize the screws error profile.  
For what I do, its not worth the effort, nor do I have the means to 
measure it that accurately anyway.

I run emc2 HEAD from about April 2006, on a bare bones 1.4GHZ Athlon with 
kubuntu-6.06 installed on it, along with a real time kernel.

> Not bad, but not as good as what I work with at my job.
>
> I've learned that offsets and HOME are important. (I did have trouble
>figuring into this on my coco)
>
>FINDING HOME:
> Originally and even now, I use brute force to drive the axis into a
>mechanical stop. The stepping motor would stall there and hum until the
>software stopped driving it. At that time, I set the software to assume
> it was HOME. Actually, it works! However, it isn't elegant. So I'll add
> some micro switches to the sytem to fix this. Besides, I can move that
> switch to a more convenient place for each project.
> My coco takes a long time to step these motors to where they are
> supposed to be. Basic09 is a lot faster than RS Basic, but it doesn't
> hold a candle to ML. .....Someday, maybe!
>
> I envy you folks that can "whip up an ML program".
>
"Whipping" up an ml driver on the coco could be done, but I'm not sure how 
one would get around the os9 60 hz interrupt limitation for a stepping 
timebase.  With emc2 on the kubuntu (linux) box, I can step at 5000hz 
with nearly full torque, and 20khz if running free where it doesn't need 
full torque.

>  So far, I can move each of the three axis from home to the limit and
> back and my dial indicator reports a return of exactly zero. This
> implies that my stepping motors, power transistors, software and math
> are working correctly. The only problem is finding home. The micro
> switches will take care of this.
>
> Backlash on each of the three axis are different. The Y axis is only
> about two steps on the stepping motor. This is incredible!.

Yes it is!

> However, I 
> took extreme care to minimize backlash when building this thing. I
> won't go into detail on how I did this until later.
>
> I've learned at work, that extreme brute force and extreme mass of
>machinery is important to get the results necessary. Heavy metal is
> good! However, it's expensive. The machines that I work with cost a TON
> of money. We're talking 600 thousand dollars for a small one. My first
> 4k coco was $300 and took a couple of pay checks to pay for it. So, I
> won't expect to do what those monsters can do. I just want to make some
> small gears, sprockets and pulleys. Later, I want to do some 3D stufff,
> small things.

The ideal machine for the hobbiest today is I believe, the SAIG X3 as sold 
by Grizzly.  I've been drooling over that for several months, but I've 
got to get to a place where I can separate the wood from the metal, 
sawdust causes instant rust on fine metal surfaces even if its dripping 
with wd-40 or similar stuff.  In other words, my shop is wayyyy too 
small, or I have wayyyy too many toys. The glass is half empty...

> More later....
>
>
>
>George

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2007 by Maurice Eugene Heskett, all rights reserved.



More information about the Coco mailing list