Linksys WRT54G antenna system

I don't know about the stock linksys firmware, but you can set the antenna to use for transmit and recieve (either single or diversity) with the dd-wrt ver.23 drivers.

Hi, Diversity antenna system has three different ones. You can combine them all to improve reliability/integrity of signal.

1. Polarization diversity. You use both horiz. and vert. polarized signal.
2. Space diversity. You use two antennas spaced in the multiple of wavelength.
3. Frequency diversity. You use two different frequency. Then you combine the received signal to produce more reliable output. So you need at least two antennas, two TX/RX and two different signal feeders(coax, waveguide, etc.) Tony, VE6CGX

Hi, Also TX power is adjustable. Tony

| > I'm also curious if there is any practical way to switch the system to a | > single mode antenna as my feeling is the diversity as provided by Linksys is | > more marketing smoke and mirrors than an effective/functional option. | >

| >

| >

| Hi, | Diversity antenna system has three different ones. You can combine them | all to improve reliability/integrity of signal. | 1. Polarization diversity. You use both horiz. and vert. polarized signal. | 2. Space diversity. You use two antennas spaced in the multiple of | wavelength. | 3. Frequency diversity. You use two different frequency. | Then you combine the received signal to produce more reliable output. | So you need at least two antennas, two TX/RX and two different signal | feeders(coax, waveguide, etc.) | Tony, VE6CGX

Good over view but it's a bit more complex than that as the design of the diversity switching/combination circuitry system itself can produce significant difference in the resultant performance which is why I wanted to get a look/see at how linksys was doing the deed.

I no longer have access to the lab so I can't run the necessary black box test ergo I need to see what's inside and how they do what they claim they do.

Hi, In old days, I worked on military long range comm. system based on troposcatter link. Diversity will make the signal usable while disabling it causes link to go down. Now there is something called MIMO on routers. I don't exactly know what it does. Tony

Whilst the article is about testing MIMO it will give you some idea.

Bob

Tony Hwang hath wroth:

Actually, there are about 10 different types of diversity reception. If you include the DSP based systems such as MIMO, the list doubles in size. I belive the question is how does the Broadcom chipset in the WRT54G do diversity.

Not normally used by the WRT54G. You can point the antennas in all manner of odd directions to improve the odds that an antenna will hear a signal, but the antennas are normally both positioned vertically.

The distance between the WRT54G antennas is 14cm or a bit more than one wavelength at 2.4GHz (12.5cm). Not even close to multiple wavelengths.

The WRT54G only works on one channel at a time. However, you could argue that OFDM works this way as it uses multiple carrier frequencies to improve resistance to multipath. However, since OFDM can be done with one antenna, that's not the diversity answer.

Combine? You can't just add them together with an RF combiner because the odds are that the two signals will cancel rather than reinforce. This will also create the rather dreaded inter-symbol interference that is caused by a delayed or reflected signal arriving to trash an incident signal. Since the WRT54G contains a diversity SWITCH, that switches back and forth between antennas, and only one antenna is on at a time, so that's not the diversity answer.

So, how does the WRT54G (Broadcom chipset) do diversity? As far as I can tell, from reverse engineering, the algorithm used is rather crude. With a lack of signal, the WRT54G scans (switches) between the two antennas waiting for valid data. When it hears and decodes enough to extract a source MAC address, it stores the antenna number (L or R) along with the MAC address. From that point on, it will use the stored antenna to communicate with the stored MAC address. However, if the WRT54G decodes errors, garbage, interference, or trash from that MAC address, it will assume that the selection of antenna is no longer optimimum and will try the other antenna. This will continue until it receives a valid packet which defines the current "best" antenna. In other words, it uses the antenna from which it received the last good packet.

Disclaimer: I deduced the aformentioned from watching the waveforms on the diversity switch and extracting what I could from the chipset data. There's a chance that this could be wrong or incomplete.

"Jeff Liebermann" | | So, how does the WRT54G (Broadcom chipset) do diversity? As far as I | can tell, from reverse engineering, the algorithm used is rather | crude. With a lack of signal, the WRT54G scans (switches) between the | two antennas waiting for valid data. When it hears and decodes enough | to extract a source MAC address, it stores the antenna number (L or R) | along with the MAC address. From that point on, it will use the | stored antenna to communicate with the stored MAC address. However, | if the WRT54G decodes errors, garbage, interference, or trash from | that MAC address, it will assume that the selection of antenna is no | longer optimum and will try the other antenna. This will continue | until it receives a valid packet which defines the current "best" | antenna. In other words, it uses the antenna from which it received | the last good packet. | | Disclaimer: I deduced the aforementioned from watching the waveforms | on the diversity switch and extracting what I could from the chipset | data. There's a chance that this could be wrong or incomplete. |

Bingo! As I expected.

If that's the case the diversity is a PoS. The same system was use in the original US cell phone systems (Chicago) by E.F. Johnson and is based on the detection of 'rice clicks' (long story but a theory developed by a Japanese PhD who's name was in fact RICE) in the received signal.

In theory (marketing's theory) this improved performance by 3db. In practice -- less than worthless for voice, data was a nightmare.

Now to find out how to disable the diversity and work on a single port.

"NotMe" hath wroth:

Not if your major purpose is to minimize the effects of multipath and reflections. My limited testing shows that it has a substantial advantage in a highly reflective environment (i.e. typical office). Sorry, no numbers. It doesn't do anything useful for interference reduction or when using two radically different antennas.

I vaguely recall part of that story. I was running an Johnson/Comco land mobile dealer at the time. The big thing in land mobile was receiver voting systems to make low power handhelds useful. The idea was to plaster the city with receivers, bring all the audio to one place via lease lines or microwave, and decide on the fly which receiver has the best S/N ratio. This works fairly well for VHF/UHF voice and miserably for data. The problem was that GE and Motorola had the patents on the easy ways to measure S/N on the fly, leaving E.F. Johnson with whatever technology they could find. Counting noise hits was what they contrived. I got to try it on one of their repeaters in the LA area. It barely worked.

When cellular appeared, the first AT&T test system was in Chicago using E.F. Johnson mobiles. I'm not sure who supplied the base stations but I think it was also E.F. Johnson. I vaguely recall reading something about a voting system being used. What happened with the original system before it was politicked by Motorola?

Just about all the voting system worked badly with data. There was no dead air time to switch and they always induced phase glitches and clock resyncs. However, the WRT54G does NOT work like a voting system and is fast enough to switch between packets.

There doesn't seem to be any way to do that with the stock Linksys firmware.

alternative firmware, such as DD-WRT, allows setting the router to either antenna.

Depends on your application! I have a WRT54G working as a WDS repeater. Left antenna is connected to an external dish, pointing at the base AP. Right antenna is broadcasting in my house. It was explained to me here that that shouldn't work, so I was really surprised that it did. Given Jeff's explanation above, I'm not surprised that it does.