[Fourlanders] Contest Computer RFI

[Brian McCarthy]

Some facts:
1. The link from "great white" to the smaller trailer was run at 10MHz.
2. Two different hubs were tried in the smaller trailer. The second hub 
tried was of all metal construction.
3. The amplifiers for both 50 and 222 MHz where tied to trailer frame 
ground by short braided line (about 18 inches.)
4. The measured resistance from neutral/ground to trailer frame + skin 
in the great white measured < or = 1 ohm.
5. While it seemed there was a possible correllation between 2m antenna 
trouble and network problems, the network was problematic through to the 
end of the contest.
6. Windows firewall was turned off on all the PC's for the contest.
7. We were using static IP addresses and had set network DDE and other 
settings as recommended by the Writelog documentation.

I know there are more significant details that could be used to fill in 
an "is vs. is not" chart. What did I miss?

I have three of the five PC's that were connected to the contest network 
here in my basement. In the next couple of weeks I will start a test bed 
where I will run 3-5 PC's with Writelog in a manner like what we do on 
the mountain. I will connect my K2 and may ask to borrow one or two 
other radios that can be RS-232 connected. I would like to run the main 
part of the test for days, weeks or even a couple of months. That will 
be easy with my hardware, but borrowed radios may be the challenge.

I have been playing with thoughts of how to RF-proof 10/100-Base-T 
cabling. I liked the theoretical advantage of the piece of CAT-6 that we 
used between the trailers. In all of the documentation I have been able 
to find, it looks like Ron is correct in observinng that only 2-pair are 
required for 10-Base-T and maybe also 100-Base-T.

Here is a thought that I had for a, hopefully, RF-proof cable and 
termination:
1. A piece of shielded CAT-6.
2. Shield connected to a ground rod outside the trailers.
3. If only 2-pair are required and 4-pair are in the cable, ground one 
pair to one end only at each end.
4. Put toroids (type 43?) on the data carrying leads right before they 
terminate to the hub in each trailer.
5. Use metal case hubs, or install the hubs into metal boxes bonded to 
trailer frame/skin with minimum lead lengths.
6. Discard the wall warts by running the hubs from the 12VDC used in the 
adjacent station (may require internal mod's to hub).

Bob, K4SZ, have you ever looked at whether you are able to measure any 
signnificant interference between 2304 MHZ and WiFi hardware? What kinds 
of testing would we need to do before we would trust WiFi as a 
connection between the two trailers? And just in case anyone is curious, 
the WiFi access point that I was able to "connect" to (really associate 
with) was 11.1 miles away and was an optical path. I believe I had a 
rather solid 1Mbps link, but no IP address so no internet or otherwise 
was possible.

I don't think that RF into the ethernet cables is our only problem.

Some other thoughts that were discussed on the mountain include trying 
to quantify the amount of data bandwidth occupied by 4+ stations with 
RS-232 connected radios. We should run the network with a DHCP server 
instead of static IP. There were some other ideas, but those escape me 
at the moment.

We never want to be on the mountain without some way of connecting to 
the DX Clusters. Neal made a number of decisions about where to point 
the 6m antennas at significant moments in the contest due to the 
information available on the cluster. I spotted hearing one of the 
larger northeast stations on 222 MHz and ended up working that station 
and quite a few others. I believe that spot helped alert the northeast 
that we were hearing them and that we should be able to work them. Next 
week I expect to take delivery on a new datacard like the one we used 
during this past contest. While I am not getting it just for contesting, 
it helped in the emotional part of the justification to be able to use 
the card for contesting.

I need to get back to a few other things, but I hope this was useful to 
all. I like the discussion and I am willing to listen to ideas and 
investigate what it will take to have rock solid networking and internet 
connectivity for the contests.

Cheers,
Brian
NX9O



Rogers, Ron wrote:

>Another thing to think about in reducing the amount of localized RF in
>or around the equipment is the fact that we DON'T GROUND ANYTHING and
>everything is just "floating".
>
>Yes, it is difficult to find a decent way to ground the trailers, the
>Ethernet hubs being used, etc. But it is worth brainstorming to see what
>we could do. The high speed, high impedance differential line-driver,
>line-receiver amplifier circuits used in those hubs are simply packaged
>in their plastic cabinets with no shielding (I suspect). In any case, I
>know there is no frame grounding and simply have 2 wire "wall warts"
>powering them which means not even a having frame ground brought out to
>an "Earth Ground" pin. 
>
>We old-time repeater builders designing equipment for mountaintops with
>lots of 500 watt pagers and 2-way systems have learned by our share of
>"hard knocks" how important some of these small details are.
>
>Also, the suggestion of switching Ethernet gears down to 10 Mbs may have
>merit and should be tried.
>
>I'll also be shopping around for some more of the clamp-on ferrite cable
>"decouplers" we use on some of our NCR equipment. I've had to install
>some of the same type on my RS-232 4800 baud cables at the Sawnee
>Mountain repeaters site to maintain reliable low speed RS-232 comm
>between the multiple repeater controllers and the "modem gateway"
>computer. We run Win Laplink on that computer to provide remote console
>programming capability of the repeaters at the QTHs of the repeater
>technical support people.     
>
>Ron
>WW8RR
>
>
>_______________________________________________
>Fourlanders mailing list
>Fourlanders at contesting.com
>http://lists.contesting.com/mailman/listinfo/fourlanders
>
>  
>