Skimmer Server & CQWW: one year later
- What a difference a year makes!
First an apology. Of all the nights for Verizon to turn off the fiber optic
line to the house for two hours, Saturday night was the wrong night. Of
course they changed the IP address so users who were using a numerical
address instead of cw.wz7i.com were unable to reconnect. I apologize.
I think I have found a computer system appropriate for running SkimServer.
I used an Intel Core 2 Quad Q8300 and at times it ran over 2200 decoders
without getting above 80% CPU utilization. I never saw less than 100%
decoding. XP Pro was the OS and the Skimmer Server version was 126.96.36.199.
I was unaware of the CWSegment lines in the SkimSrv.ini file. But Saturday
morning I looked at a bandmap and saw a hole between 7.035 and 7.045 and
realized something was wrong. I modified the default CWSegment lines and
broadened the segments to accommodate contesters outside the strict CW
portions of the bands. Alex suggests having two or three versions of the
CWSegment line and then �commenting out� the undesired lines so one can
rapidly modify the file. My current version of this area of the file is
below but I can�t guarantee it is optimal. I am still learning. I would be
very interested in suggestions from contesters about appropriate band
;DX CW Segments
;Contest CW Segments
Of course, there must not be any �carriage returns� in these lines. If you
bring up your SkimSrv.ini file in Notepad and turn off Word Wrap, the
command should be one long line. The above contest segments line assumes
you will run 192 kHz bandwidth.
With the telnet spot filtering set to �Aggressive�, and having the CWSegment
line set inappropriately for the first 15 hours or so, the system generated
63,000 spots from a HyGain HyTower Junior vertical. This is not as many as
some of the systems.
I want to express my appreciation to those cluster owners who made these
spots available to contesters without them having to connect directly to my
computer. I don�t have a good feel for how much telnet connections cost in
cpu time or network traffic � I wouldn�t think it very much. However I
averaged 10 connections and it seemed to have an impact, at least on the
Hopefully during this contest I corrected a couple of stupid errors and next
contest I will be better prepared. Things went much better than last year!
I welcome suggestions that would improve the data. I hope folk had fun
[Non-text portions of this message have been removed]