Re: [CAD_CAM_EDM_DRO] EMC Resolved (Somewhat) and more questions

reed_irion wrote:

>Thanks for the inputs.
>
>I discovered that the "hanging" on running EMC was due to the floppy drive not being
>present (since it runs off the parallel port of my old laptop). I took out the CD and put
>the floppy in and now it runs OK. I even hooked it up to my step motor test bed and it
>works OK. Getting this far is most of the batlle so I am a happy camper.
>
>As a note to the developers- I thought that the "zero axis following error" in the
>"simulation" mode was a bit disconcerting. I got it to go away by monkeying with the
>feed rate, but it seems like since this is a simulation it should just work (at least it
>should work out ot the box, before the .ini file is messed with). 2 cents
>
>Which brings to mind a few questions:
>
>1) There is no PCMCIA stuff on the BDI CD, hence I have no ethernet, which makes
>me feel as if I am in the dark ages of computing. Will installing (and running) card
>services goof up the EMC ?
>
>2) "Mini" wont run. It wants me to "install a symbolic link to an existing platform" ?
>Is this hard ?
>
>

No, just cryptic. There are directories under the EMC that hold
"platform" specific
executables and installable kernel modules. These are generally
something like
/usr/local/emc/plat/linux_2_2_18 and /usr/local/emc/plat/rtlinux3.0
You need to have links to these directories using generic names, such
as /usr/local/emc/plat/nonrealtime and /usr/local/emc/plat/realtime

Believe it or not, the reason this all needs to be done is because you
DON'T have
a network card installed! When the network software is running, it builds a
hostid name that is requested by the emc.run file, and is used to set up
these
links automatically. That bit of shell script breaks down if the network is
not running. So, the easiest fix is to get the appropriate drivers,
etc. loaded
to make your net interface card work. (This may not be so easy as it should
be, however.) Or, you can manually install the links. The commands
should be
something like :
cd /usr/local/emc/plat
ln -s linux_2_2_18 nonrealtime
ln -s rtlinux3.0 realtime

another fix is to edit the emc.ini file, and every place you see :
PLAT = nonrealtime
replace that with PLAT = /usr/local/emc/plat/nonrealtime
and every place you see :
PLAT = realtime
replace it with PLAT = /usr/local/emc/plat/rtlinux3.0 (or whatever the
directory
name actually is)

>3) Is there a way (like with an oscilloscope or something) that I can verify that the real
>time kernel is working properly ?
>
>

Yes, there is a test program for the RT patch that flips a data bit on
the parallel
port at a scheduled interval. A frequency counter or scope allows you to
check the accuracy and jitter of this timing.

>5) I have 350 oz. in steppers that I plan on gearing down 30:12. Will this be enough
>to drive my grizzly G1008 (8"x30" half size knee mill)
>
>

Probably, but in many cases steppers perform best when NOT geared down.
The torque falls off so rapidly when the speed rises that gearing down may
be counter productive.

>6) Is there an easy way to measure the "effort" of the step motors ? What are the
>guidelines for figuring out the optimum speed for the motors ?
>
>

A very simple baseline can be generated with a fish scale. Hook it to
the hand
crank and pull with the fish scale until the handle moves, and the screw
engages
the table and moves it, too. (ie. don't be fooled by backlash in the
screw.)
Meaure the radius of the crank handle, and you can convert to Oz-In or your
favorite torque units. This of course neglects cutting loads, so you
have to
add a fudge factor. Assume with good drivers that the motors put out very
roughly constant power, with the max speed being maybe 1000 - 1200 RPM.
This is with Gecko 201 or 210 drivers, others may not perform this well.
(Yes, you can drive the motor faster than 1200 RPM, but I'm throwing in
some fudge factor here, too.) This would mean that the motor's torque would
be 350 Oz-In at zero speed, and 175 Oz-In at 600 RPM. If your load
measurement
showed that 100 Oz-In would just move the table, then you have 75 Oz-In
left over for acceleration and cutting loads, at a speed where the motor was
turning 600 RPM. If you were to gear down 2:1, you would have to spin the
motor at 1200 RPM, where the torque is nearly zero. You don't win in this
situation!

>7) The Z-axis of my mill is sloppy, sloppy, sloppy and I dont think there is much I can
>do about it. I've been considering some kind of arrangement whereby I put the mill
>into "drill mode (with the lever)" and use some kind of backlash proof worm gear
>mechanism to drive the z-axis up and down. Has anyone else had (and solved) the
>same problem ?
>
>

Very difficult. I ripped all that rack and pinion stuff out, and put an
eBay ballscrew
in to move quill or head (I've done a Bridgeport and a "mini-mill".)
Even if you
had an anti-backlash worm drive on the pinion, you still have lots of
backlash
in the rack and pinion gearing, so it doesn't help.

Jon