Re: Looking for Green Letters (Again)

["DPM" <dpm-at-enteract.com>]

Dear Matt, Ray and List,

Here are the answers to your queries:

> ; Name of motion control program
>  EMCMOT =              stg8mod.o
Matt>You've got the 8 axis card?

Yes, we have the STG model 2 8 axis card.

> ; Base address of Servo To Go board
> STG_BASE_ADDRESS   =    0x240
Matt>Double check this...

This is confirmed, 240 is the address.

> [TRAJ]
> AXES =                  2
Matt>This should probably be set to 3

We tried setting this to 3, and then in the [AXIS 2] section
made the settings identical to the first axis. We only have
2 axes anyhow, so this makes me curious.. Why would it need
to have 3 specified?

> HOME_OFFSET =                   0.0
Matt>This says that you'll home right on the index pulse.

I had gathered that this meant the axis travels no further
than the position at which it tripped the home switch?

> HOME_SWITCH_POLARITY =          0
Matt>This is what you'll probably need to change.

We Always try any changes we make (while debugging this home
switch problem) with both polarities, just to be sure.

So far with your suggestions here we changed the number of
AXES to 3 and tried homing with both polarities... No dice.


Now Ray's comments:

Ray>My first quick look says that you need to use the bridgeporttask and
Ray>bridgeportio for the stg board.  Minimill doesn't try to look at any
Ray>extra I/O beyond what stepper or freq mod wants from the parallel port.
I
Ray>suspect that stg8 does demand those extra signals in order to work
properly.

At this point we have only one switch hooked up to pin 47 and ground. When
we
switched over to bridgeport (by changing only the TASK and EMCIO .ini
variables
to their corresponding bridgeport values. First thing we noticed upon
running
is that we were not able to get out of ESTOP.

>From the linuxcnc.org site, under e-stop problems we find this familiar list
of
things that cause one unable to turn e-stop off:

1)input from amp fault
2)input from +/- hardware limit switch
3)+/- software limit
4)input from estop sense

To test number 1, (since we have no amp fault inputs anyway) I just tried
toggling
the FAULT_POLARITY on every axis. The original polarity was the only one
that did
not cause amp faults to show up in the console window, so I assume that this
was
not the problem.
2 and 3 would have caused the letters in the display to turn red, and they
were yellow still, so I ruled these two causes out.
4th, Once again we have no estop input, so I just tried toggling the estop
sense
polarity, and once again we have no luck.

I didn't think just switching over to bridgeportio/task would've caused our
situation
to progress negatively, as it did. What could have gone wrong?


Ray>What I'd suggest is that you switch to tkemc for your gui and when you
get
Ray>it running you can use the script under the menu button named
IO_Exercise to
Ray>look at the locations on the stg card.  You can do that by entering the
Ray>addresses. (only 8 pins at a time)

I tried the tkemc interface and the io_exercise script (I always knew I
wasn't
exercising enough... {joke}). I set the view address to 640 (base address
0x240 plus
0x400 for the PORTA offset) and was able to flip bit 0 back and forth from
red to
green (with the hardware switch). From all the documentation I have read
this seems
just dandy.  I am able to easily flip the switch via the hardware switch, so
I didn't
try changing the bit with the script. Is there an advantage to manipulating
the portA
register with the script rather than with the hardware switch?

Ray>We really should write a whole Tcl/Tk script for analysis of the stg
card with addresses
Ray>and EMC names for the signals on each pin.

We downloaded some software from www.servotogo.com that was a console
diagnostic
application complete with source. It compiled under windows as well as
linux. This worked
pretty well for us, and let us be sure our switch was wired correctly (at
least as far
as the STG card was concerned). We realized the need for this software
before we even had
it, and it was sufficient, but the EMC names would've saved me some browsing
through stg2.c
for the home_0 bitmask for confirmation. The software was at
http://www.servotogo.com/dosprog.zip, or the linux binary
http://www.servotogo.com/lstg.zip
both right off the download page.



I'm eager to learn more about EMC, and I truly appreciate all the help this
list has
been to us.

-Jason (and Joe)