| reply to Anon_tertech
Re: Yet Another Modem Statistics tool Hi TerTech,
It may take a while until you see this but i though it would
be better to post my findings to your attention. The result
isn't what i was expecting but i must confess i've played a
lot with your numbers... In any case, i'd wish to bring your
attention on the date stamp of March 26 3:00:19 (3:01:29)
and that of March 27 23:00:29 (23:01:39): you'll recognize
a pattern in the date/time bug i've found, i presume...
Among the few things i've done, i tried to reject the effect
of disconnections on the SNR Margin curves in order to get
a better idea of what these are like... Clipping has been my
favourite trick for some time now so it was applied here too
as an attempt to ignore "sparkles" which would contaminate
the curves with single events otherwise. Last but not least,
i used a logarithmic function in some creative way and was
so pleased with this i finally included that element for you to
consider as a potential addition in case of future upgrades:
it allowed me to fit low and high value patterns together on
a very same graphic (there were so much more FEC errors
than the CRC/HEC ones i just couldn't use the same scale).
Looking at the graphic i notice your line was relatively OKay
between 1h30 - 4h00, 10h00 - 15h00 and 22h00 - 23h30... |
%20.GIF "46858 bytes")
