[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-rt
Subject:    RE: [rtl] Timing/response time jitter question
From:       "Estabridis, Janet P" <EstabridisJP () navair ! navy ! mil>
Date:       1999-09-27 13:10:56
[Download RAW message or body]

I would also suggest trying to find an A/D converter with a FIFO so that you
can cut down on the IRQ's.  You have to send a group of converts over the
e-net anyway.  That way when you enter your IRQ you read a group of converts
and you aren't constantly getting interrupted.

I am doing pretty much the same thing you are describing. Also, we found
that it took about 40 msec to send e-net packets on our machine.  We set up
a real-time timer task to send a packet to see how fast we could go before
we started missing the timer tasks.  So, for our conversion rates we really
can't keep up with sending the converts in real-time, but we will be getting
the data where it needs to go with a delay that is acceptable for our
application.

Janet Estabridis
Electrical Engineer
NAWC Code 473E00D
China Lake, CA 93555
estabridisjp@navair.navy.mil
(760) 939-2896  FAX (760) 939 -3075


> -----Original Message-----
> From:	Tomasz Motylewski [SMTP:motyl@stan.chemie.unibas.ch]
> Sent:	Saturday, September 25, 1999 9:31 AM
> To:	Reef Morse
> Cc:	rtl@rtlinux.org
> Subject:	Re: [rtl] Timing/response time jitter question
> 
> On Sat, 25 Sep 1999, Reef Morse wrote:
> 
> > what is the minimum interrupt period that can be handled.  I'd like to
> be around
> > 25 microseconds, AND have time left over for handling all the other
> tasks.
> > The system will be acquiring data from an A/D in real time, and
> communicating
> > that data over the net to another computer.  Will the communication
> timing foul
> > up the A/D, and will the A/D found up communication?
> 
> Will be tough, but if you do not have too much data to transmit, could
> work.
> Keep in mind that every ISA read/write of byte or word takes at least 
> 1.25 us. So 25 us is much too litle time to send a packet over ISA network
> card. Assuming with every IRQ you will just read 2 bytes from A/D and that
> network does not need to be hard real time (one packet per several A/D
> interrupts) it could work. Definitely, the priority of A/D ISR needs to be
> higher than network. So your routine sending packet will be interrupted a
> few
> times by A/D interrupt - but this should be no problem.
> 
> Expect 2-5 us delay on RT interrupts. If you manage to do your part of ISR
> in 5-10 us it should work leaving 5-10 us for the Linux kernel and user
> processes.
> 
> --
> Tomek Motylewski
> 
> --- [rtl] ---
> To unsubscribe:
> echo "unsubscribe rtl" | mail majordomo@rtlinux.cs.nmt.edu OR
> echo "unsubscribe rtl <Your_email>" | mail majordomo@rtlinux.cs.nmt.edu
> ----
> For more information on Real-Time Linux see:
> http://www.rtlinux.org/~rtlinux/
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail majordomo@rtlinux.cs.nmt.edu OR
echo "unsubscribe rtl <Your_email>" | mail majordomo@rtlinux.cs.nmt.edu
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/

[prev in list] [next in list] [prev in thread] [next in thread]