RE: PCI I/O cards

[Kurtis Martineau <klmartineau-at-shaw.ca>]

Thanks for all the advice regarding the above issue.

My EMC application at this time is not a milling machine, but rather a
BurgMaster Houdaille drilling table with a turette. The X Y Z axis limits
are 144", 48", and 12" respectively. The turette has 8 limit switches (one
for each tool). Each axis has a brake.

I am collaborating with someone local who does C programming and has
controls experience.
We will be using a second monitor and keyboard for the remote operator
station instead of keeping the existing remote pendant with its manual
pushbuttons, selector switchs, and indicators.
We will now be able to use the STG cards 32 I/O without any other PCI cards.
My original reasoning for using the second I/O card was the following:
	1. I didn't want to modify what is already done for the STG card.
	2. I didn't want to use the parallel port.
	3. I wanted to move the existing parallel port logic to a second I/O card
and then add the logic needed to keep an existing remote pendant.
	    (The total I/O count would have been more than the 32 on the STG card)
We've now decided to enclose a 15" NEC flat screen monitor in a box with a
window for a second display. I am mounting a relay controlled switch-box for
a second keyboard. A touch screen would be ideal here and it may become an
option in the future.

I have spent much time deciding how to uprgrade this machine - and three
others in the future.  I have spent alot of time trying to bring myself
up-to-speed with the EMC project and I don't expect to get everything right
without a mistake or two along the way.  I have found a lot of good
information on this list; although it does take time to seperate out the
important and relevant.

I thank all those who contribute time and expertise to the questions posted
here. (Now if more of the information could be put into an up-to-date and
complete handbook, the world would be a better place.)


Kurtis

> -----Original Message-----
> From: emc-at-nist.gov [emc-at-nist.gov]On Behalf Of Ray
> Sent: Sunday, December 09, 2001 7:02 PM
> To: Multiple recipients of list
> Subject: Re: PCI I/O cards
>
>
>
>
> Kurtis
>
> I didn't see your original post of this but have a few comments.  First
> the software issues and then the hardware.
>
> Fred's comments re the stg are that you can use the extra STG I/O that
> isn't used for axis stuff for anything that you wish.  The hook here is
> that you will have to hack the bridgeport C code to do that.
>
> There is a Tcl/Tk parallel to the bridgeportio named tkio.  You can read
> about creating a PLC and about using Tcl/Tk  with the iosh interface in
> the linuxcnc.org/handbook.  You can run tkio by simply replacing
> bridgeportio with it on the approariate line in your ini file.
>
> You do not need any external proprietary device drivers.  The interface
> iosh which is called by tkio and several of the tcl scripts can
> be used to
> directly read and set any hardware that has a specific address.  This
> direct access to ports is typical of the EMC.   An example of this
> approach is emc/tcl/scripts/IO_Exercise.tcl.  If you have a free parallel
> port, you can use this script to pull solenoids and read switches as a
> stand alone or running alongside the EMC.  If the Sealevel card
> has or can
> be configured to have a defined location you can directly access that
> location using the available emc software.  I don't have a clue how you
> would do this across the PCI bus.
>
> There was some talk a while back about using a common parallel port with
> some external logic chips.  The parports eight out could be expanded to
> many descrete signals and the five in could be combined from many
> descrete
> signals.  This could be accomplished with a few logic chips and a power
> supply.
>
> I have the advantage of knowing a little tcl but what I would do in your
> case is write it using Tcl.  It is a rather easy language to learn at the
> level that you would need for your PLC project.  Doing this would require
> some processor power but it is a quick code to test.  You could have your
> code starting to pull servos and read switches in a few minutes.
> (Immediate gratification is good here!)  As the code and logic grows and
> evolves to include full machine control and safety and is running the
> machine to your satisfaction, you could migrate, or hire, or con someone
> into migrating that code to C to produce a final system.
>
> This code would eventually become of value to other emc users if
> you built
> it as a set of blocks, ie tool logic, pallet logic, operator panel logic
> that could be configured from the ini file or an aux ini and combined to
> produce a working machine.
>
> I'd be able to share my ideas and work with you a bit on the tcl coding
> here.
>
> Ray
>
> On Saturday 08 December 2001 10:29 pm, Don wrote:
> > The STG card has 32 bits of digital input.  Isn't that enough?  Of
> > course, we're not PCI yet.
> >
> > Don McLane
> > Servo To Go, Inc.
> >
> > At 11:46 AM 11/30/01 -0500, you wrote:
> > >This is a multi-part message in MIME format.
> > >
> > >--------------InterScan_NT_MIME_Boundary
> > >Content-type: multipart/alternative;
> > >  boundary="Boundary_(ID_KUzl4HCuPVZJ1AAgI0loPg)"
> > >
> > >--Boundary_(ID_KUzl4HCuPVZJ1AAgI0loPg)
> > >Content-type: text/plain; charset=iso-8859-1
> > >Content-transfer-encoding: 7BIT
> > >
> > >Calling Linux programmers,
> > >
> > >My EMC application requires more I/O than is presently configured to
> > > the parallel port with the Bridgeport "plc".
> > >I am hoping to move the parallel port stuff
> > > (lube,spindle,coolant,estop) to a Sealevel PCI Bus 48 TTL Digital I/O
> > > Interface as well as incorporate the remaining I/O.
> > >(8 tool limit switches, remote pendant functions, x and y axis brakes)
> > >The software that came with the Sealevel card has two steps:
> > >step 1.  =  install drivers.  This worked and prompted installation
> > >complete.
> > >step 2.  =  load drivers.  This uses the mknod command to create device
> > >special files.  Running the load drivers resulted in :  mknod: wrong
> > > number of arguments
> > >The manufacturer of the card (Sealevel) assures me that the program
> > > works on Linux.
> > >How can I read and write to this PCI card the way I do with the Servo
> > > to Go card.
> > >
> > >The next step will be to include STG analog inputs of spindle speed and
> > >spindle load on the Tkemex Gui.  Ray says the GUI stuff is the easy
> > > part.
> > >
> > >Kurtis
> > >
> > >
> > >
> > >--Boundary_(ID_KUzl4HCuPVZJ1AAgI0loPg)
> > >Content-type: text/html; charset=iso-8859-1
> > >Content-transfer-encoding: 7BIT
> > >
> > ><!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
> > >Calling Linux programmers,
> > >
> > >My EMC application requires more I/O than is presently configured to
> > > the parallel port with the Bridgeport "plc".
> > >I am hoping to move the parallel port stuff
> > > (lube,spindle,coolant,estop) to a Sealevel PCI Bus 48 TTL Digital I/O
> > > Interface as well as incorporate the remaining I/O.
> > >(8 tool limit switches, remote pendant functions, x and y axis brakes)
> > >The software that came with the Sealevel card has two steps:
> > >step 1.  =  install drivers.  This worked and prompted installation
> > > complete. step 2.  =  load drivers.  This uses the mknod command to
> > > create device special files.  Running the load drivers resulted in :
> > > mknod: wrong number of arguments
> > >The manufacturer of the card (Sealevel) assures me that the program
> > > works on Linux.
> > >How can I read and write to this PCI card the way I do with the Servo
> > > to Go card.
> > >
> > >The next step will be to include STG analog inputs of spindle speed and
> > >spindle load on the Tkemex Gui.  Ray says the GUI stuff is the easy
> > > part.
> > >
> > >Kurtis
> > >
> > >
> > >
> > >
> > >--------------InterScan_NT_MIME_Boundary--
> > >--Boundary_(ID_KUzl4HCuPVZJ1AAgI0loPg)
> > >
> > >
> > >--Boundary_(ID_KUzl4HCuPVZJ1AAgI0loPg)--
> > >--------------InterScan_NT_MIME_Boundary--
>
>