Live from Hollywood... District!
Noon (and before): ScottMcClung shows up with a white van packed to the gills with hard gear - ladders, more electrical cords than Home Depot stocks, and a 40' tower broken down into 10' sections. He also brings an important ingredient to move troops - Pizza!
Noon - 1pm: Nate, About 5 people show up, and quickly gravitate to the back yard -
Well, I got to put a lot of cables together, and found that having the right tools does make it a lot easier. Hopefully a lot of people will be able to setup open nodes now that they have the cables they need.
Towards the end I went on a little walking tour of the Hollywood area with a 15dBi SuperPass patch panel pallet wrapped to my notebook. We were testing the range of Scott's 8dBi SuperPass Omni with a HyperLinkTech 1watt amp on his 40 foot tower. All I can say was that the range was GREAT. We got very usable signals through LOTS of trees and buildings. I would estimate that we got about 1/3km away at ground level. Other people that went driving with a patch and a yagi said they got past Lloyd Center which is about 3km away... (CaseysMappingSoftware works great for figuring out how far you got!) Hopefully everybody will fill in more information from their notes.
I am hoping for the next PlayDay that I will have Linux properly installed on my Toshiba 430CDT (P-120 with 48Megs). If Linux will support all the hardware (including the dock) it has 4 PCMCIA slots, two stacked, and two on opposite sides of the dock.. So you can have 1 ethernet and three wireless cards in without any problems, and the cards are far enough apart that you wouldn't get a lot of interference from each other. (I have been told all wireless cards should be 2-3 inches apart or they will cause interference even though they are on different channels.)
I arrived around 12:30pm to dump stuff, got back about 1pm. Started by setting up a table and putting my stuff on it just to get it out of the way. Linked up the the main network and verified that I could get out.
After the tower was assembled (hopefully ScottMcClung, others will write more about that) with an ethernet-converter on one onmi and a 1W amp on the other, I started to set up my AlphaMultia at the base, with two Lucent Silver cards. First one linked up to the main network just fine. Second was attached to the amplifier and configured as Net 3 (see PlayDayNetworkNaming). The 100Mbps NIC in the machine was attached to the cat5 from the tower attached to the ethernet converter, and was configured. First problem we ran into was that the Alpha for some reason doesn't want to set the channel of the card in ad-hoc mode. The idea was to set the various networks to different channels so they wouldn't overlap and conflict. Well, that wasn't too easy to do with iwconfig hanging. So we ended up moving the main wireless network to channel 11 (and determined the very useful fact that every node moved with the AP, seamlessly), and when we set up the RG-1000 off the ethernet port of the Multia (replacing the ethernet converter) we set that to 6. I'll have to try to figure out what's wrong with the driver and get that working properly.
Another issue was that I need to learn how to do iptables. We did get the Multia set up as a masquerading machine for the two subnets (net2 and net3) hanging off it after a while, but it took a little while. I never did get dhcp up and running I don't think. More preparation is necessary on the central routing machine. I think that we even need to set up a machine or other config that does not much *but* act as the main router/traffic-cop for the PlayDay. Something with 100Mbps to the host network that masq's into 10.x.x.x, with different interface options. A Multia might be optimal, with some more software, but I also want to use a Multia as a cell gateway, so I'll need to get another functional one...
Later three of us drove around for an hour with a couple laptops and antennae seeing how far we could get from the tower and still link up. We got a lot of strange looks driving around with what looked like a green PVC bazooka sticking out of the sun-roof aiming in seemingly random directions. I had to hack up a script to do ping surveys while we drove, and came up with something that worked rather well, considering. Any packet would almost immediately give a beep, which would meld into a constant buzz if the connectivity was good.
Something I need to work on is a program that can do all the stuff we need to do surveys and such. I'd like to make it a modular Perl script that uses as much external stuff as possible to avoid having to write too much code. It should, among other things: allow multiple network settings to be saved; watch the air for all nodes/networks (a la NetStumbler); show ping times, bandwidth, SNR, etc.; link to a GPS and store permanent maps of ping times/bw/SNR against location; whatever else people want it to do. I'll set up a page for it soon.
I didn't get to test out self-routing, or madhoc, which we'll have to do at the next meeting (three laptops needed, only one has to be Linux).
I arrived about 12:15. Being the new kid on the block, I kind of hung back and observed. My father and I were able to learn quite a bit. Once the tower went up, the one antenna worked great (this was the one w/ the 1 watt amp stuck on it. The other one did not work so well. What we were thinking originally was that this had something to do with being in a dead zone under the tower. To prove or disprove this we took my laptop (with a fresh copy of NetStumbler)---hooked up to a directional antenna (3 guys with a bazooka and a laptop (and a police car that rolled on by) Erik found it was a problem in some of the hardware after this....My father and I had to leave before testing distance, but what I did find, was on a plain 'old vanilla orinico card, sitting within a window van, was that netstumbler lost the connection off the amp'ed antenna right around 56th and Sandy. (Not to sure how far that would be--haven't looked it up on a map yet....
I arrived around 1:15. Unpacked some equiipment and promptly got on the net thru Dan's AirPort. After some beer drinking and general "wows" at other peoples cool equipment I helped along with others in getting on of my Lucent Gold cards running in Sam Churchills notpad machine (not sure of the name), which after some driver troubles seemed to work pretty darn well. Scott and I took his 15dB i SuperPass patch panel out with netstumbler running on my notebook and got fabulous signal out to about 440' from the tower. I would have gone further but notebooks get heavy after awhile. The signal was actually stronger 1 1/2 blocks away then right next to the AP at the house! I watched as MichaelCodanti showed us the wonders of having the right tools when it comes to making cables, and promptly bought an 8dBi omni and lucent pig tail from him. Overall this was a great event, lots of sucsess with the tower cant wait to do it again!