Re: [CAD_CAM_EDM_DRO] Interface options, headless boxen, and remote control ...

Larry Edington wrote:

> The timing seems to be the single biggest problem.

Of course, but that's where an RTOS comes in. Maybe we're talking
apple-to-oranges here -- since I'm coming at it at a different
angle.

> With Windows doing the timing, it can only go so fast.

And even the two NT RT extension products I have evaluated only do
"soft" real-time. Both are also based on a license of the same
patent held by Finite State Machine Labs (fsmlabs), the same guys
behind RTLinux. Small world. ;-PPP

Again, I'm coming at it from the EMC system with an RTOS approach.
The standalone RTLinux kernel on a good 500MHz CPU can handle
interrupts in less than 10us, and usually has a guarantee process
response time under 100us. The NT+extensions as well as Linux+RTAI
approach is more soft-realtime, although the later is "ready and
usable" in the stock 2.4 kernel.

I have actual used none of the above, just VxWorks as a RTOS, but
I'm planning on changing that.

> With logic out on the card doing the timing after being handed
> a table of step data, you can go as fast as your external
> hardware can take the data and not burdon the PC host.

RTLinux is designed to be burdened. It is design to run as the
"host" OS, handling all the RT I/O and processes, then running any
"general" I/O and processes on a real Linux kernel that runs as a
non-realtime process.

-- TheBS

--
Bryan "TheBS" Smith mailto:b.j.smith@... chat:thebs413
Engineer Absolute Value Systems, Inc. http://www.linux-wlan.org
President SmithConcepts, Inc. http://www.SmithConcepts.com