Re: Bad Date Stamps

["Bob Scott" <rlscott@xxxxxxxx>]

Steve,

Thanks for your input.

There are from what I have seen two scenarios of BMI Bad Date problems - one
which has affected all users and the second which randomly affects users.  I
have been plagued with the Bad Data problem that affects all users on a
couple of occasions but I have not been subject to the "random" problems.
The Bad Date problems all result in the same type of date -  in 1999 it was
10/27/99, in 2000 it was 10/27/2000  and now it is 10/27/2001.  The "random"
problems experienced could indeed be the result of a weak RS232 driver and
that should be checked out but the problems which have been experienced by
all users must be a system related problem.  The problem may lie in the BMI
datafeed or it might be in the TS4 software.  The BMI datafeed transmits two
types of Time/Date markers - one each 30 seconds which contains month, day,
hour, minute and second information and one every 5 minutes which adds two
digit year and day of the week information.  Since the "Bad Dates" repeat
from year to year I am assuming that a problem exists in the 30 second
marker and that is what I am concentrating on.  The ID field for the 30
second marker is a two byte value but there is also a two byte checksum that
could get mistaken for the ID if the ID were not properly decoded by the TS4
software.  This is what I suspect the problem is but I want to rule out any
anomalies on the BMI datafeed end.

~Bob

-----Original Message-----
From: Stephen Gibson <gibson3d@xxxxxxxxxxx>
To: omega-list@xxxxxxxxxx <omega-list@xxxxxxxxxx>
Date: Thursday, July 19, 2001 6:04 AM
Subject: RE: Bad Date Stamps


>Bob:
>
>Find someone with an oscilloscope and check the data coming
>out of the receiver. You want to see nice clean looking square
>waves. Adjust the timebase until they appear to stand still.
>
>The key pattern to look for is the vertical rise in the waveform.
>It must be 100% vertical. If you are seeing bits going up or down
>at the slightest angle then you have a problem. You may have fried
>the output RS232 driver chip in the box or you may just have too
>many computers loading the output. Look at the waveform with the
>machines connected too. You must see nice clean square waves or
>they can be very easily misinterpreted...resulting in bad data,
>bad dates and a bad day.
>
>Steve Gibson...
>
>
>
>