Michael Barabanov wrote:
>
> Some preliminary code for this have been in RTLinux for quite some time.
> I plan to have support for this in the next version of RTLinux.
>
> What are your timing requirements, e.g. what maximum clock jumps over
> some fixed period can your application tolerate?
>
> Michael.
>
We plan to operate an interrupt-based synchronous control system, with
the
interrupts derived from our GPS clock. We are designing the system to
run with
two external clock signals: 1 PPS and 1 kHz, so we need to know GPS time
to 1 ms.
The GPS clock time is distributed by a dedicated NTP server (Datum
TymServe 2100)
to all machines. The simplest scheme would be to use the NTP clock to
"time
stamp" the 1 PPS events. This way each machine would resync once a
second.
We will also keep track of any lost 1 msec clock pulses as well.
My main concern is that all the machines (we plan to use at least three
separate computers, and this number may increase if we convert the rest
of our aging control system to RTL) are synchronised to a common time.
We
can probably recover from small 1-2 msec glitches, as long as the errors
don't
accumulate.
FYI I am very much an RTL newbie. We are fortunate in collaborating
with
a group with similar interests who use RTL 1.0 for their system. I am
working
through their code and trying to adapt it for our system.
W J Tango
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
