[Skimmertalk] VHF RBN Update

[John Ackermann N8UR]

We got the "proof of concept" version of the Skimmer/Aggregator running 
at the N8ZM VHF contest site (EN80EE) this weekend.  It worked (and is 
still working) but needs some tuning, particularly to filter out busted 
calls -- I know what needs to be done, but I won't be able to get up to 
the site until next weekend to make the changes.

Just for reference, we're running 4 dongles into a Raspberry Pi 3 
running the latest version of N1GP's "rtl_hpsdr" software.  The four 
dongles support three bands, as 2M requires two receivers because the 
operating and beacon segments are too far apart.  This will also the 
case on some of the higher bands; we haven't figured out the final 
frequency settings yet.

The antennas are a single M2 halo on 6M, up about 30 feet, and stacked 
M2 halos on 2M and 70cm up about 20 feet.  Feedlines are 100+ foot runs 
of RG6.  The permanent installation will raise the antennas up higher 
and put preamps on the tower to make up for cable loss.  The 2M antenna 
goes into a two-way splitter that drives two dongles.

The Skimmer Server and Aggregator software are running on an HP laptop 
with i5 processor.  CPU usage hovers around 10% and though we'll be 
looking at wide swaths of spectrum, we're typically decoding only one or 
two signals per band, so I don't think horsepower will be a problem. 
(And the RPi3 seems to be loafing with four dongles plugged in.)

I was pleased that the system worked as well as it did this weekend, 
given how low the 2M/70cm antennas are, and that we had high power 
transmitters banging away on 50 through 1296 MHz.  The dongles were 
overloaded when any of the transmitters came up, but there was enough 
down time that spots still got through.  Using RBN data from 20 January 
and 21 January, we spotted 4 stations on 6M, 10 on 2M, and 2 on 70cm.

One thing I missed was setting the Aggregator filters to block grids and 
<1dB signals.  We also have a lot of blown calls from the nearby 6M 
beacon, so I need to put those into the bad calls list.  Of the 32 
callsigns recorded in the RBN database from N8ZN, 14 were busted and 18 
were real.  We'll get that improved.

I'll be experimenting at home this week to figure out optimum gain 
settings for the dongles.  This weekend we ran them at about 25 dB and I 
have no idea if that's anywhere near the right choice.

Overall, I'm very pleased with how this worked out, and we plan to (a) 
improve the N8ZM setup, and (b) encourage more VHFers to install similar 
setups.

73,
John