Distance between WiFi access points

There are two WiFi access points (802.11b/g/n, 2.4GHz, SOHO type); one operating on channel 1, the other on channel 12. Considering typical selectivity and out of band emission, what should be minimum distance between the access points so they won't jam each other?

Vladimir Vassilevsky DSP and Mixed Signal Designs

formatting link

Reply to
Vladimir Vassilevsky
Loading thread data ...

I've had them sitting within a foot without problems (channel 12, really? Mine only go 1 to 11 unless they are the ones with the higher number 5 ghz channels.)

Unless there is a particular need to do that, I usually avoid it, but that's engineering caution, not any actual "jamming" going on.

--
Cats, coffee, chocolate...vices to live by 
Please don't feed the trolls. Killfile and ignore them so they will go away.
Reply to
Ecnerwal

Are you sure it's Ch 12 and not Ch 11? In the USA, there are only 11 channels.

For 802.11b/g, channels 1 and 12 are essentially non-overlapping so any distance where the RF doesn't produce overload (blocking) of the receiver is adequate. Even 40 Mhz 802.11n is ok, because that channels again do not overlap. I've stacked routers with built in antennas on top of each other and they still manage to function normally. However, a safe distance would be about 3 ft using horizontally spaced omnidirectional vertical antenna. Directional antennas will need to be on a case by case basis. If you want me to grind better numbers, please disclose the tx power output, chipset, and antenna description.

However, I can offer a rough guess(tm) based on commodity 802.11g hardware:

- Tx +15dBm into a 2dBi rubber omni antenna.

- Rx starts to overloads at about -20dBm anywhere in the 2.4GHz band (because there's no RF front end channel filter).

- Rx antenna is another 2dBi rubber omni.

Grinding numbers: Tx Tx ant Path loss Rx ant Rx +15dbm + 2dB - XX dB +2dB + -20dBm (approx) Plugging into: and guessing various values for the distance, until I get a -20dBm Rx signal level, the distance is approximately 0.0005 miles or 2.6ft. (Note that Rx sensitivity value in the calculator is irrelevant).

Note that overload (blocking) is NOT the same as co-channel or adjacent channel interference. With blocking, the receiver front end rectifies and amplifies the RF until something downstream saturates and distorts the modulation sufficiently to make the receiver appear to be comatose or deaf to desired signals. It need not be spread spectrum and can be an unmodulated carrier. This blocking RF signal can be anywhere in the receiver front end bandpass. Considering the rather wide bandwidth of commodity wi-fi ceramic filters, anywhere between 2.4000 and 2.4835 will suffice. If you have a 2.4GHz RF source, you can simulate the problem by attaching an antenna to the generator, and play jammer to a wireless link running some manner of performance benchmarking software, such a iPerf or Jperf. As the blocking signal level increases, you'll see a rather abrupt drop in thruput.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

Transmit channel produces substantial spill into other cnannels, as well as wideband white noise. That feeds directly into receiver and could degrade noise figure. The selectiveness of the receiver is limited also. I don't have the numbers however I guess the receiver attenuation between 1 and 11 is 20dB or so.

They seem to work in ideal tabletop situation; with substantial reduction in throughput.

What would be the distance that the degradation of NF wouldn't be noticeable?

It is known fact that powerful (~20dBm) WiFi extenders disrupt other WiFi networks in the area.

VLV

Reply to
Vladimir Vassilevsky

Are you sure it's Channel 12? In the US, we only have Channels 1 thru

11 (but are allowed lower power on ch 12 and 13).

The xmit spectra for direct sequence spread spectrum looks like a sin(x)/(x) waveform on a spectrum analyzer. It does on forever and never really hits zero. What makes it hit zero are ceramic bandpass filters found in some (not all) consumer wireless devices, and requirements that the transmitter conform to the FCC spectral mask: Ch 1 is centered on 2412 MHz. Ch 12 is centered on 2467 MHz. The difference in frequencies is 55MHz. Half that is 22.5MHz. According to the 802.11g spectral mask, the noise should be about

-22dB down. By the time it gets to the center of the other channel, the noise should be -40dB down.

One of the nasty problem with direct sequence spread spectrum is that it doesn't take much of a signal to degrade the BER (bit error rate) and thruput. My guess(tm) is that anything strong than about -6dB below the desired signal is going to appear as some form of degradation, either in BER, association speed, or thruput. Using my

-6dB guess, that means that the noise from the other xmitter can be

-34dB down before causing problems. Using the same equipment, calculations, and guesses as in the previous example calculation, and throwing in a receiver sensitivity of -68dBm at 54Mbits/sec from: I get a distance of 0.0025 miles or 13ft, at a fade margin of about

34dB.

With 802.11n, if your radios have adaptive beamforming, you can pickup some additional isolation. The radio will recognize the emissions from the other radio as noise, and insert a null in the antenna pattern in that direction. Effectiveness of such a scheme varies significantly with reflections and implementation.

It depends on what you're trying to accomplish by stacking routers. In general, it's a bad idea. However, the typical indoor wireless environment is so bad, that most wireless routers spend their life operating at the low end of the speed range due to perceived interference, where any further reduction is unlikely.

According to my above guesswork, about 13ft for 20MHz modulation between channels 1 and 12. Seems a bit high. From my experience, between ch 1 and 11, about 8-10 ft. It should be easy enough to test. Start a file transfer between one access point and a laptop and watch the transfer speed. Throw another access point and client laptop into the mix, and move the access points together until something changes.

+20dBm (100 milliwatts) is not very powerful by todays standards and is not a problem. The common AT&T branded 2wire 2701HG-B wireless router belches +26dBm (400mw). There are illegal devices available online that produce +30dBm (1 watt) and more. The problem is that these devices produce an "alligator" which is an animal with a big mouth and small ears. A +30dBm (1 watt) access point can be heard over a much further distance than the +15dBm laptop with which it's communicating. The laptop can easily hear the access point, but the access point cannot hear the laptop. Beyond the range of the laptop, there's a rather large area where nobody can reliably connect using a +15dBm transmitter, where the +30dBm access point is polluting the area preventing other users, of other access points, from using their systems.

Also, if you're dealing with an overload (blocking or desensitization) problem caused by the proximity to a high power wi-fi xmitter, such excessive power levels will require additional separation.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

That should be 27.5MHz.

and that should be -30dB down.

I never could do math in my head while talking on the phone and filling out the online warranty return forms for yet another failed hard disk drive.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

On those channels so far apart that the sideband tails do not overlap signals much at all you can operate them stacked one on top of the other subject to overheating. Try the experiment to be certain but I would guess that even 1m apart would be enough that they didn't diminish effective throughput to a measureable extent. You will obviously have to load both networks hard and simultaneously.

formatting link

As a rule of thumb you need to be at least 5 channels apart from any other strong WiFi signals which means that since round here almost everything BT defaults to channel 1 that 6 or 11 are good choices.

--
Regards, 
Martin Brown
Reply to
Martin Brown

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.