Re: second computer for DRO readout

["alex" <telecomt-at-tht.net>]

Paul,
thank you for pointing out deficiencies of my idea.
The normal pinging time in a local network is close to half of a
millisecond -
may be too slow for 1 millisecond update, as you indicated.
What I don't like in an ISA card is the fact that ISA bus virtually
disappeared by now.
This  network distributed architecture can help to go around of this
limitation.
I wonder how much precision will be lost for a slower update times -
may be it work for a hobbyist?
Best regards, Alex

----- Original Message -----
From: Paul <paul.corner-at-tesco.net>
To: Multiple recipients of list <emc-at-nist.gov>
Sent: Friday, December 20, 2002 4:11 PM
Subject: Re: second computer for DRO readout


>
>
> Hi Alex
>
> EMC normally reads the encoder position at one millisecond intervals
before
> running the internal trajectory calculations. To add the DRO feedback from
a
> slow non-realtime source such as a TCP/IP network would negate the reasons
> for running a realtime operating system.
> Whilst there are low latency "realtime" network drivers available, these
> still do not have the response times required - The time involved in
getting
> everything to work reliably would be better spent on wire wrapping a DRO
> board.
>
> Regards, Paul.
>
>
> On Friday 20 Dec 2002 4:34 pm, alex wrote:
> > Can the DRO input be done into another machine, a DOS machine ,
> > for example? I think this was already accomplished, and DOS machine can
> > capture encoder data into the parallel port.
> > This DOS machine can be networked to the EMC Linux machine over the
> > Ethernet (WATTCP). Then data in a DOS machine can be written into the
> > socket and Linux Emc can read them from the socket at the Ethernet rate
(
> > no RT required), similar its is done now for the ISA card, and used for
a
> > closed loop control of the steppers.