Re: hardware driver

[Jon Elson <jmelson-at-artsci.wustl.edu>]

"Terry L. Ridder" wrote:

> hello;
>
> i am not sure i totally understand what you are
> asking.
>
> 1. take eric's code and integrate it into emc?
> 2. take eric's code, dac code, and encoder code, and integrate
>    them?
> 3. take eric's code, dac code, and encoder code, and integrate
>    into emc?
> 4. none of the above.

This is all from memory, so some of the file names will be off a bit,
but, EMC has generic routines that perform functions that are
broadly defined so that they should operate nearly any hardware
that drives servo amplifiers.  There is a routine to read encoder
counts and produce 32-bit position values, and a routine that
produces +/- 10 V DAC voltages, and some others that read
and set digital I/O bits.  These are surrounded by a 'wrapper'
to make the linkage from specific procedure names to generic
procedure names, so that specific driver routines can be easily
substituted into EMC.

I wrote some routines in Turbo Pascal a few months ago to drive
the new hardware.  Eric volunteered to convert this to C and integrate
it into EMC.  He got as far as converting the stg_dac, stg_encoder
and similar routines so that they compile and link to the diags
program that was originally used to test the STG card and its
driver.  I took this work and made the few small changes to get it
actually working with the hardware.  That is not the problem at this
point, I can manage the hardware/software debugging.

But, the 'wrapper' and all the makefiles that allow specific drivers
to link into EMC at run time (through the .ini file and the run script)
are just a little more complicated than I am able to cope with, without
much written to describe the myriad interactions and the flow
of control through these many layers.  Also, the makefiles need to
be adjusted to allow the building of these additional new driver
routines.

That is what I need the most help with at this time.  I think the actual
C code itself is pretty well in hand (but I could be wrong on that).

Jon