Re: Vexing EMC problem; SUCCESS !!!!

IT
WWWWWWWWWWOOOOOOOOOOOORRRRRRRRRRRRRKKKKKKKKKKKKKSSSSSSSSSSSSSSSS!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

WHOOOOOOOO HOOOOOOOOO !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

See comments below

--- In CAD_CAM_EDM_DRO@y..., Ray Henry <rehenry@u...> wrote:
> Reed (comments mixed in)
>
> I really hope that you and your Linux buddies take this not as a
rebuke but
> as a, "there is always more to learn."
>
> >    From: "reed_irion" <Reed_Irion@h...>
> > Subject: Vexing EMC problem
> <s>
> > I have managed to get the BDI up and going on my computer.  I
think I
> > have managed to get it mostly working (no small task..)  I find
the
> > most useful thing to have when doing BDI is a bunch of Linux
friends.
>
> We may be closer to an answer here than we might have imagined.
The bdi is
> just that and for most installations, ignorance may be an initial
advantage.
> I have found a number of folk who try very hard to interpose their
will into
> the install or initial operation process and only succeed in
messing things
> up.
>

I'll take guided learning from an expert any day over wading
aimlessly through mountains of documentation without a clue. I
started from square -1 with Linux. My friends vastly accelerated the
learning curve.




> > Anyhow on to my problem-
> >
> > The clock signals are not coming out of the correct pins on the
> > parallel port.  When I run the 3dtest.ngc example I can see x-axis
> > clock coming out on pin 13 and y-axis comes out 3 pins over.  I
cant
> > see z-axis clock coming out anywhere.  Nor do I see any direction
> > signals.
>
> What are you using to view these signals? IO_Show.tcl will do it
from the
> scripts menu in tkemc. If this is a recent version of that script,
you can
> toggle through the pin definitions and you will find one that lists
a0s for
> axis 0 step and such. The direction signal should show correctly
but steps
> will only show some fraction of the real number of signals.
>

I was using an "oscilloscope" to view the signals. I tried the
parallel diags you listed here as per your suggestion. They are
quite useful. I do however feel that the oscilloscope is a much
better verification of proper system function.


> > I tried it with both minimillio and bridgeportio.  When I look in
the
> > system information in KDE it has nothing assigned at 378.  I have
> > tried every bizzare permutation in the BIOS with the
ports.  System
> > is a Dell Optiplex GXPro 200MHz.  It has two serial (DB9) and one
> > parallel (DB25) connector.  One of my linux friends had me try all
> > kind of contortions to get it (linux) to see the port, but to no
> > avail.  Get a parallel port card ?  Get a different sytem ?


Turns out the problem was a SERIAL DB25 cable masquerading as a
PARALLEL cable. I was chagrined as that darn cable has screwed me
before. I marked it so it wont happen again


>
> NO! Linux only shows that it has a parport when Linux is using a
parport.
> When steppermod or any other EMC related module is using a parport
it isn't
> linux doing it. The EMC uses its own drivers here and they do not
show up as
> a part of the normal Linux device drivers in that screen.
>
> I am not trying to rag on your friends or to demean their
understanding of
> Linux at all but the solution here may be a six-pack when the
friends come
> over. Linux is a great, open-source project but some advice should
not be
> taken -- even when you consider yourself to be a newbie and others
to be
> expert.
>
> > Is there more info somewhere on the parallel port in EMC (as in
NOT
> > the generic IBM web page on the parallel port)  Like: what is the
> > pinout when using the auxilliary port with bridgeportio ?  What is
> > the pinout for a fourth (rotary) axis ?
>
> These are all shown in the labels to the parport pins as you toggle
through
> them in IO_Show. Bridgeportio only works correctly if you are
using the
> default bridgeport definitions in your ini file.


Yes I found it. I still wish the auxiliary port pinout was detailed
in the docs somewhere



>
> You might want to test your parport using IO_Exercise.tcl but you
will have
> to do this when the parport is not being run by steppermod. Start
a terminal
> and move it to /usr/local/emc if this is a normal BDI. Issue the
command
>
> tcl/scripts/IO_Exercise.tcl

For "newbies" I would suggest recommending using the menu driven
IO_exercise in the default TKEMC window; under "scripts". Until a
few days ago I had no idea what a "terminal" or a "script" was.


>
> and you should get a widget that looks as confusing as all-get-
out. Default
> here is 378 so if that's your parport address assigned during boot
you should
> be good to go. If not enter the correct address in both "show
address" and
> "set address" windows and press both "enter" buttons. If this is a
viable
> parport the "lights" should be green or some combination of green
and red.

All this worked like a charm.

>
> Click any pin or combination of pins and you will see the value of
that pin
> in decimal notation in the "value to set" widget. Press the enter
button
> beside that widget and it should change the real pins to that
value. If it
> doesn't, something is wrong with that pin on your parport. (Note
that some
> pins are reversed in the way that they work)
>
> Enter the value 379 in the view window and you can see the
condition of the
> input pins from the 378 parport. Enter 37A in both and you can
toggle the
> value of the extra output pins 0-4. Toggling pin 5 will change
them all.
>
> > So close I can taste it, yet so far...

Like I said: IT WORKS!!! I have a cheapo XY table rigged with two
wimpy step motors. Not great, but it responds to commands. I have
vectorCAM, CMD-260 drivers from Mr. Goldstein (VERY robust, a great
deal !) Some less wimpy motors and a Grizzly mill ready to party !!
I know what I'm doing this weekend ;)

>
> Yes you are!
>
> Ray



Thank you for your advice