12 Sep
2002
12 Sep
'02
1:19 p.m.
--- "Dwight K. Elvey" <dwightk.elvey(a)amd.com> wrote:
That's why they built this device I have - a BERT - Bit Error Rate
Tester(?) You hook two serial devices to it and crank the knobs
to inject errors. It can insert as few as 1 error per 1E06 bits
passed through, IIRC.
It's excellent for testing error recovery schemes, V.whatever as well
as those in the protocol stacks. We used it for testing our Bisync
products for proper packet retranmission behavior on noisy lines.
i.e. - to verify that we implemented the HASP protocol correctly
(since one of our advertised features was that our HASP engine did
all the error detection/retransmission, presenting clean blocks of
data to the PDP-11/VAX, reducing its overhead).
-ethan
__________________________________________________
Do you Yahoo!?
Yahoo! News - Today's headlines
http://news.yahoo.com
From:
"Jeffrey Sharp" <lists(a)subatomix.com>
100 feet at 115200 bps through three unshielded wires, wrapped around a
three-foot-tall Jacob's ladder apparatus enclosed in plexiglass. No bit
errors. We gave up and decided to build a device to simulate bit errors.
Hi Jeffrey
What makes you think that wrapped on the outside of a Jacob's
ladder has either the right coupling or the right frequency
spectrum to effect a RS-232 signal? Arbitrary test like this are
not real world test and only impress those that don't understand
what it takes to interfer with a particular signaling method.