[CQ-Contest] Skimmer - A long way to go?

David Kopacz david.kopacz at aspwebhosting.com
Wed Jun 11 12:38:48 EDT 2008


I thought I would report my initial results testing skimmer.

I am running Skimmer with a FlexRadio Systems 5000A receiver using a
48Khz audio bandwidth. I would prefer running 192Khz, but I haven't yet
dedicated a PC to the Flex and using a higher bandwidth is choppy on my
3.4Ghz Pentium IV machine with 2GB RAM, even when I am doing nothing
else with it.

Before I report my initial results with Skimmer, I would like to mention
that having never used an SDR radio and never seeing skimmer, this
wasn't a simple plug and play operation. There were lots of
prerequisites that needed to be done to get skimmer up and running with
the Flex 5000A, all of which were documented individually, but with no
step by step instructions. It is my desire to reduce this burden on
future installations and write up a prerequisite checklist to save
people hours of frustration.

Nonetheless, it is up and running and reporting stations. Please keep in
mind, that I have not used the system more than a few hours, so this
intial report is based upon my limited knowledge of an SDR and Skimmer.

My initial findings are skeptical. Skimmer is able to decode call signs
and even entire QSO's, as long as the copy is good and the sending
station sends good CW. It does not seem to handle mistakes well. For
example, when a seding station sends "KY1V de K1LZ GM DAVE HOW ARE YOU
DOING", depending on timing, QSB and noise, it may come out like KY1V DE
K1LZGM DAVE HARE YOU DING, which as you can see is not very effective.
My intitial thoughts are that tied with an AI dictionary, populated with
common CW shortcuts like TNX, OM, GUD, CU and such, Skimmer could be
trained to correct these issues.

I haven't read all of the doucmentation on Skimmer yet, but I find that
it reports lots of call signs when in ALL CALLS or RAW TEXT mode, but
very few in verified mode. Another issue I have discoveredm which may be
a setup issue, is when adjust the Flex main VFO frequency, the Skimmer
display field changes and calls move from one displayed location to
another. On my setup, this has the undesired effect of not only moving
the calls signs on the Skimmer display, but also to shift their
frquency. The net effect is that when I look at, or click on, a station
in the skimmer display, it displays the incoorect frequency and sends
the radio to the wrong frequncy.

This is certainly an issue that would prevent me from effectively using
Skimmer for daily DX'ing, let alone a contest situation. As stated, this
may be a setup issue, but from my initial perspective, it appears to be
a software problem. I should be able to tune the radio and have the
skimmer display move in synch with the Flex VFO. All of the reported
stations should remain on the same freuqnecy, but they do not.

Next I decided to hook this thing up to the reverse Skimmer network and
see what happens. When I fiorst did this, I was overjoyed as 4W6R and
AH6K immediately appeared on both my Skimmer screen and the reverse
beacon network. I thought, how cool, I could feed this data into my
logging program instead of a packet cluster and be able to quickly find
and work stations on the band (gave me quite a vision of the advantage
this would have for an SOAB A operator over an SOAB U operator).

Then disaster struck! What? Disaster, how so? To my embarrassment, as I
continued to listen to 4W6R work JA's in rapid succession (he couldn't
hear me as I didn't have my Acom 2000A hooked up, BUMMER), I was
horrfied as Skimmer started respotting him on the same frequency, 4W6R,
4W6K, 4W6A, 4WEAR and other incorrect variations. I was so embarrased to
be sending out these bad spots, I shut down Skimmer before it continued
to send out bad spots.

What went wrong? I have no idea. I guess some more reading and
experimentation is in order. For now, Skimmer will remain in
experimental mode at my home. If I can get it working correctly, the
whole system will be moved to Jamaica to test in a live contest (CQWW
perhaps?) at 6Y1V, most likely a M2 operation where it won't effect (or
infect) the coveted SOAB U category.

While I can see where Skimmer could be a very effective tool of the
future for finding and working new multipliers, it isn't ready for that
task now. I don't see this as a tool that everyone is just going to run
out and buy and use to start dominating contests. Perhaps Stan and
others are correct that the future may be dim and hold a grave end for
CW contesting, but I really don't see this technology going away.

In the early days of packet, I never used it as I didn't own any VHF
gear, but as the Internet evolved, I found myself connecting to telnet
servers to find those elusive DX stations and a pileup or two to try and
crack. Like most, I now regularly use the cluster, but I have to say
something dear was lost along the way...the thrill of tuning the band
and being the first to find that elusive DX station calling CQ, like
Africa in the afternoons on 15 meters. Packet has forever lost that
excitement as even tuning the bands today, we often find ourselves too
late to be the first when someone quickly fires out a spot and causes a
pileup.

I suspect it will get worse, as reverse Skimmers will find DX in an
instance and pileups will begin before we ever have a chance to tune the
band.

It is a shame, but its here.

Foir those of you tasked with the ungrateful job of making the rules for
future contesting, I wish you well and good luck.

I'll report more as I continue to experiment with Skimmer.

David ~ KY1V






More information about the CQ-Contest mailing list