I attended a talk on the LabVIEW RT system recently. We considered
using it for generating/logging car signals - pulse trains, analogue
I/O, CAN, fault simulation and control of GPIB instrumentation. For our
system, it's important that the timing is repeatable and we can take a
snapshot all the I/O at the same time.
Please remember that these opinions are my own, not my company's - and
that I may be totally mis-informed about the NI RT product...
The drawbacks with the LabVIEW RT system (the way I see it) are:
Limited I/O - The RT hardware carries just one daughterboard, and these
are limited to 16 channels at up to 250kHz. This particular board has 8
DIO, 2 AOut and 2 24bit timers. The only way to expand I/O (at the
moment) is to use the SCXI product which I understand bit-bashes serial
I/O using the DIO (ie = slow). SCXI gear is expensive. Apparently
there is the feature of synchronising the acquisition with other NI
boards using NI's RTSI bus, but it wasn't clear how the RT board would
read from these boards. It doesn't allow us to interface to the NI CAN
or GPIB products. Our requirements needed 6 counter/timers and 8
analogue outputs - not easily done using the RT/SCXI Hardware.
Volatile program storage - The RT board has to be reloaded with the
firmware everytime the system is cold-started. At the moment this has
to be done via LabVIEW.
No watchdog timer - Maybe there is, but the NI technical guy didn't know
about it.
Other Information:
Communication - Shared memory and TCP/IP.
OS - Pharlap ETS.
If anyone is interested at looking at the technical information, it's
available at http://www.natinst.com/labviewrt
In all, LabVIEW RT was expensive and wouldn't do it for us. Instead,
we're planning to use rtl for all the hard work. A LabVIEW PC running
WinNT will generate the simulation data and graph the results,
communicating with the rtl box via TCP/IP. This way saved us around
$20,000, gives us every feature on our wishlist and is expandable using
off-the-shelf hardware that we can source from any number of
manufacturers. If all goes to plan, the rtl system will give us 5x the
sampling/reconstruction rate of commercial systems for 20% of the cost
(including development time).
Having said that, I've now committed the company to rtl and put my foot
in it - I'll be doing the development. So you can bet I'll have some
(mostly dumb) questions... look out! :).
Stuart Warren
BTRA
"David J. Christini" wrote:
>
> Other than the obvious price hurdle (hardware=$2500-$3500;
> software=$3500-$5000), and the stomach-turning need to use Windows, can
> anyone give me their thoughts on what the other disadvantages are of the
> new LabVIEW RT system? (they've put out a real-time version of LabVIEW
> that utilizes special boards that have an on-board 486 processor to run
> the real-time tasks: http://www.natinst.com/labviewrt/).
>
> I'm not considering it for myself, but in the interest of RT-Linux
> advocacy and fairness, I want to be able to recommend to people why they
> may or may not find the LabVIEW-RT system suitable for a given
> application.
>
> Thanks!
> Dave
>
> --- [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/
--- [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/
