802.11 WebLog

26-May-2002

You would not believe some of the things that I have seen in the ‘Wireless Community Networks’ out and about. People who you assume to be at least slightly professional, but cannot even afford a basic crimp tool for an ‘N’ type connector. The Tool is worth all of US$20, so they use a vice. That is not the way to get things to work correctly.

From my logs I am seeing a lot of people are interested in PRINGLES antennas. I am still attempting to work out the maths for them. I am sure that they are not the best antenna to build for the price, and I just need to do the maths on them. I just need to remember all my microwave enginnering from uni.

There was an interesting discussion a few months back on the NOCAT.NET mailing list on the maximum range of 802.11, subject to power being available. That was never resolved, but the standard does sound as if it can be stretched in a whole lot of ways that the designers never intended, so long as you dont mind loosing throughput.

Now for a piece of advice if you are testing 802.11 equipment over long distances. It works better if you use infrastructure mode rather than ad-hoc. It will be easier to get a lock meaning that your allignment is much easier.

4-February-2002

Frequency Source – How? Amplifier and feedback loop. Is Minicircuits better to build an oscillator.

31-January-2002

Question: How do you test 802.11 Antennas at 2.4 GHz without test equipment?

Answer: 2.4 GHz frequency source feeding the antenna under test. Use an old Galaxy TV Antenna and Receiver to test the antenna at a distance. The fun part is to work out where 2.4 GHz is tuned.

The signal source should be easy and cheap to build – less than $10.

29-January-2002

Started work on the Simple Radio Routing Protocol (SRRP) which is a protocol designed for wireless networks found in 802.11, where large bandwidth is available, the links are bi-directional, and there is little formal structure in numbering or anything else.

The work is based on some of the ideas of RSPF, extended with some ideas of my own. At the moment it is all on paper, with some of the ideas simply transposed to WORD in case I loose the writing pad. The protocol is based on a tree spanning algorithm, with each node knowing only being told about nodes it cannot hear by the nodes that it can.

I am beginning to see that there can be stability issues, and also the possability (or probability) of routing loops. I am sure that there is some way to ensure that this does not happen often, but the protocol needs some thinking about. My idea is that every wireless node knows about its neighbours. Therefore every neighbour knows about its neighbours, etc. Based on this fact it should be possible to build a picture of the network.

It should also be possible to build a method of update so that the changes are propagated from node to node, with minor problem if one of the updates goes missing, provided that the missing update does not cause a circular route. This may be as simple as internode handshaking before creating a possible circular route. This might be an ‘Understand before route update’, change outgoing routes before informing other nodes to route through it.

Some type of PING and quick kill is also needed to switch over to alternate links if a link goes down. Since we are talking an application level protocol something like ping is needed.

Check out my BOOKS page with links to AMAZON.COM. Purchases through this link will help pay for my research into wireless networking (Or more correctly will provide books for my research).

27-January-2002

Started writing a paper for the DCC on 802.11 and Ham Radio… Why it is important for us to embrace it, and where ham radio can use 802.11. Much more work needed. Also started looking at RSPF for 802.11. Put RSPF standard up on WWW site.

27-January-2002

I have been working on getting my Galaxy (Conifer) antenna up and running, or more correctly documented. There are now some photos on the appropriate WWW site page for the GALAXY antenna.

25-January-2002

Downloaded a copy of the Telecommunications Act from here . It is interesting reading for those building a wireless community network. I am still working out what the implications are. I have a feeling that the only way to get arround this is to get a designation from the Minister for Communications that this we do not need a carrier licenses.

I have been thinking about the wireless routing issue. There was a comment on Sydney Wireless about routing in mesh. Something to look at…

I am thinking that RSPF is the way to go – but it needs some work.

24-January-2002

I just tried the hack described below for mounting the Galaxy dipole… Seems as if it should work. Cutting with the Angle Grinder was quick. The only thing was that it was a bit smelly. Not something to be done in the house (Is angle grinding ever?)… Then I removed the PCB’s… Nice and quick.

The threaded tube was easy too… I used a 6mm drill bit to remove the core… The procedure was something like drill in, arround, and then out… And repeat… Eventually the drill bit was too short, so I started pushing the centre dielectric from the other end and kept drilling… This lead to a hollow tube. Worked well.. Now all I need is some LMR-200

23-January-2002

Found a better list of 2.4 GHz losses for coax today so I updated my page with losses and cable sizes. What surprised me was how good RG-58 was, compared to RG174 and LMR-100A.

Got offered some Cushcraft 2.4-2.5 GHz Yagi’s… Rather nice too. I will have to make an offer if I can work out how much they are worth. I think they are ex-Olympics.

Galaxy Antenna… The downconverter board in the Galaxy unit was designed so that it provided mechanical stability to the plastic work holding the dipole. Without the downconverter there is little mechanical stability. The solution? I am still working on that. I know that the downconverter thread is no good as it is terminated in an F connector, known for being cheap, and also being 75 ohms.

I personally like the idea of using RG-213 coax from the dipole, and using some type of tube. Still working on this one. Another idea I am having us using something like LMR-200.

GOT IT The answer is to use the existing downconverter assembly, or at least the last 100mm of it… Just measure 100mm from the end of the aluminium assembly furthest away from the dipole. Now using your angle grinder cut right through everything. YES. Everything. Now remove the upper and lower shielding.

Remove the remiaining circuit boards – side cutters can be used with plyers to unscrew the two bolts. You will need to de-solder one of the boards from the F-Connector device. Then place the F connector device and Aluminium in a vice. Now comes the fun part… Programmers should not attempt this without adult supervision (and adult supervision is not another programmer)…

Get an electric drill with a 5mm drill bit, and drill down the guts of the F connector unit… You might need to remove it from the Aluminium first, but drill right through it. You will smell some melting nylon or something. That is OK. You will then be left with a hollow tube. Make sure you can pass your LMR-200 through it. If not, make the hole bigger.

You can now put the Aluminium and the F connector into the plastic housing and connect them to the cast aluminium on the end. Then take the nut that was on the F connector and tighten it. Everything is now mechanically secure!

Connect the LMR-200 to the dipole as per directions. Remember, keep the LMR-200 short. You might want to mount an N socket on the antenna so that things are really short…

Tomorrow I will actually try this… :-)…

22-January-2002

I just checked the calculations on moving a Galaxy antenna to 802.11. Galaxy was allocated frequency in 2.3-2.4 GHz for their service. 802.11 is centered between 2.412 and 2.462 MHz. This creates an error of about 1mm in the length of a 1/4 wave dipole – nothing really. Not enough to worry about IMHO… If I was woried I think I would be more interested in the antenna dipole than the phasing stub.

I still love the idea of putting a USB wireless adapter in the focus of the antenna…

Seems that Andrew Communications purchased Conifer (makers of the Australis / Galaxy Antennas) back in 1999. I have been unable to find any information on the Conifer T5120 antenna. Found this while looking… Seems to be the only specs I could find. There is also a 75-50 ohm converter on the site.

20-January-2002

Richard Talbot came to visit today to play antennas. The two antennas we built are shown here, but neither worked very well… In fact the antennas worked worse than if we had no antenna at all… It was that bad.

From Richard
Well, me and a mate [hey, that’e me… ] sat down on the weekend to build some omni’s. We used the Collinear design posted on the Melbwireless site here. Building was reasonably straight forward once we worked out the more than cryptic documentation and sorted through all the measurements.
A few points of note…

A 6mm screw with 16 threads per inch didn’t exist at our hardware shop. Even if it had, the pitch would be wrong. There is just no way one could coil a 3/64 brass rod around the scree without the coils touching.
The design looks seriously strange to us. There is nothing like it in the usual sources, ARRL experimenters guide and the like. With no maths provided it is almost impossible to verify the theory.
We cut all the parts accurately and measured with verniers. (NOTE: We used a Stanley Knife to cut the parts… Quite easy really)

With soldering complete and attached to a SMA connector, we set about to test it. Initial testing was done with a Compaq WL100 and pigtail. We were talking to a WAP11 upstairs. We couldn’t see any appreciable results so we figured it would be easier to test if we walked up the street till we nearly ran out of signal and plugged it in. Once in position 50m up the street we plugged it in and to our horror saw the SN ratio go down (- ive scale) about 15db. The card then alternated up and down every second or so on diversity. The internal antenna presumably doing better than our “5db” omni.

We rechecked all our measurements, construction and integrity of the wiring. Still no go! Just to make sure we weren’t dreaming we plugged in one of the Galaxy antennas I have previously modified from the same place. After moving the antenna around a bit we notice a 14-15 db gain on the internal antenna. Considering we had 5m of RG 213 on the end, one of those terrible Lucent connectors and the Galaxy antenna isn’t really designed for ISM frequencies, we thought this was pretty good. The score thus far Galaxy 1, homebrew omni’s 0.

——————————————————————————–

We then set to building another design. This time a 26 element omni. The maths we could understand and the design involved simply bending the brass… so this time we thought we might be more successful. We didn’t actually use the full 26 elements, more like 7. We felt this antenna couldn’t be any worse than the last in any case. Anyway, back we walked to our original testing place and no joy on this attempt either although we didn’t drop quite so many db’s. The score now Galaxy 2, homebrew omni’s 0.

Having got this far we were a bit discouraged, but we decided to press on and build a Pringles antenna. This looks promising, although we haven’t finished it yet. The design is a well known ARRL one, used by hams for quite some time. I’ll keep you posted.

I’d also be interested in hearing from anyone that has managed to successfully build an omni with a verifiable gain….
Cheers Richard.