Difference between revisions of "Mesh/Spectrum"

From Sudo Room
Jump to navigation Jump to search
 
(2 intermediate revisions by the same user not shown)
Line 28: Line 28:
*157 - 5785 MHz (5765-5805)
*157 - 5785 MHz (5765-5805)
*161 - 5805 MHz (5785-5825)
*161 - 5805 MHz (5785-5825)
:Beware that [https://dev.openwrt.org/ticket/10136 you cannot use channel 165 as part of a HT40+/- channel]


Maximum power output for these channels at 20 or 40 mhz is [http://louise.hallikainen.org/FCC/FccRules/2011/15/407/ 1 watt], with the following caveats:
Maximum power output for these channels at 20 or 40 mhz is [http://louise.hallikainen.org/FCC/FccRules/2011/15/407/ 1 watt], with the following caveats:
Line 149: Line 150:
== 5 GHz ==
== 5 GHz ==


Here are some reasonable settings for a point to point or point to multi-point 5 GHz rooftop node:
Here are some (possibly?) reasonable settings for a point to point or point to multi-point 5 GHz rooftop node:


<pre>
<pre>
Line 156: Line 157:
htmode HT40
htmode HT40
chanbw 20
chanbw 20
txpower  
txpower 20
noscan 1 (?)
</pre>
</pre>
These OpenWRT settings are documented [http://wiki.openwrt.org/doc/uci/wireless#common.options1 here].
I'm not actually sure what the noscan feature does or whether it's a good idea. There's a warning that it will violate regulatory requirements, but I'm not sure if it's related to DFS or if it's something else. - ([[User:Juul|Juul]] ([[User talk:Juul|talk]]))


The ath9k driver does not yet have support for a 802.11n-only mode (also known as greenfield). The closest is 802.11n+a. Apparently you can prevent 802.11a nodes from associating somehow. [http://comments.gmane.org/gmane.linux.drivers.ath9k.devel/10735 source].
The ath9k driver does not yet have support for a 802.11n-only mode (also known as greenfield). The closest is 802.11n+a. Apparently you can prevent 802.11a nodes from associating somehow. [http://comments.gmane.org/gmane.linux.drivers.ath9k.devel/10735 source].

Latest revision as of 04:29, 18 April 2014

Note: Legal info on this page is U.S. centric. Also, do not take any of this as legal advice. We are not lawyers.

Channel selection is not as straight-forward as one might expect. In selecting channels we take into account:

  • Minimizing interference / staying off highly used channels
Is a channel prone to interference or high use in a given area?
  • Using a single channel for many-to-many communication in a given area
Most of the mesh nodes have only one radio, so can only be on one channel at a time.
  • Legality in geographical region
Some channels are completely illegal in the U.S.
  • Additional technical legal requirements
Some countries require dynamic frequency selection for some channels

Legal

2.4 GHz

5 GHz

For 5 GHz, there are quite a few available channels, but many require the use of Dynamic Frequency Selection (DFS) and Transmit Power Control (TPC).

Non-DFS channels

The following 40 MHz wide overlapping channels are allowed without DFS for _outdoor_ use:

  • 149 - 5745 MHz (5725-5765)
  • 153 - 5765 MHz (5745-5785)
  • 157 - 5785 MHz (5765-5805)
  • 161 - 5805 MHz (5785-5825)
Beware that you cannot use channel 165 as part of a HT40+/- channel

Maximum power output for these channels at 20 or 40 mhz is 1 watt, with the following caveats:

...the peak power spectral density shall not exceed 17 dBm in any 1-MHz band. If transmitting antennas of directional gain greater than 6 dBi are used, both the maximum conducted output power and the peak power spectral density shall be reduced by the amount in dB that the directional gain of the antenna exceeds 6 dBi. However, fixed point-to-point U-NII devices operating in this band may employ transmitting antennas with directional gain up to 23 dBi without any corresponding reduction in the transmitter peak output power or peak power spectral density. For fixed, point-to-point U-NII transmitters that employ a directional antenna gain greater than 23 dBi, a 1 dB reduction in peak transmitter power and peak power spectral density for each 1 dB of antenna gain in excess of 23 dBi would be required. Fixed, point-to-point operations exclude the use of point-to-multipoint systems, omnidirectional applications, and multiple collocated transmitters transmitting the same information. The operator of the U-NII device, or if the equipment is professionally installed, the installer, is responsible for ensuring that systems employing high gain directional antennas are used exclusively for fixed, point-to-point operations. -- FCC rules

The following 20 MHz wide non-overlapping channels are allowed without DFS for _indoor_ use:

  • 36 - 5180 MHz (5170-5190)
  • 40 - 5200 MHz (5190-5210)
  • 44 - 5220 MHz (5210-5230)
  • 48 - 5240 MHz (5130-5250)

Maximum power output for these channels is 50 mW (!) at 20 or 40 mhz, with the following caveats:

...the peak power spectral density shall not exceed 4 dBm in any 1-MHz band. If transmitting antennas of directional gain greater than 6 dBi are used, both the maximum conducted output power and the peak power spectral density shall be reduced by the amount in dB that the directional gain of the antenna exceeds 6 dBi. -- FCC rules

DFS channels

None of the DFS channels allow more than 250 mW output power. The TPC requirement only applies if the EIRP is more than 500 mW.

The following channels with DFS are allowed for indoor or outdoor use:

  • 52 - 5260 MHz (5250-5270)
  • 56 - 5280 MHz (5270-5290)
  • 60 - 5300 MHz (5290-5310)
  • 64 - 5320 MHz (5310-5330) (possibly 5310-5350?)
  • 100 - 5500 MHz (5490-5510) (possibly 5470-5510?)
  • 104 - 5520 MHz (5510-5530)
  • 108 - 5540 MHz (5530-5550)
  • 112 - 5560 MHz (5550-5570)
  • 116 - 5580 MHz (5570-5590)
  • 132 - 5660 MHz (5650-5670)
  • 136 - 5680 MHz (5670-5690)
  • 140 - 5700 MHz (5690-5710)

Some, if not all of the Atheros AR9xxx chipsets support DFS. It looks like Linux support for DFS in ad-hoc mode has been implemented:

It may be that we _can_ use these channels for point to multi-point rooftop links. More research is required.

Maximum power output for these channels is 250 mW at 20 or 40 mhz, with the following caveats:

...the peak power spectral density shall not exceed 11 dBm in any 1 megahertz band. If transmitting antennas of directional gain greater than 6 dBi are used, both the maximum conducted output power and the peak power spectral density shall be reduced by the amount in dB that the directional gain of the antenna exceeds 6 dBi. -- FCC rules


Spectrum analysis / spectrum usage mapping

We use three different methods to map out spectrum usage.

Wifi channel usage scan

Most wifi chipsets are capable of reporting:

  • How many access points are on each channel
  • With how much power each access point is being received

These measurements don't account for how many clients are using the APs how much of the time and with how much power. There can be several high-powered access points all operating on the same channel in one area, but if they are unused then they will only minimally interfere with other nodes. This type of scan cannot detect if the access points are in use.

See the software tools page for suggestions on using this type of scan.

Atheros spectral scan

The Atheros AR92xx and AR93xx chipsets have a spectral scan mode that outputs FFT data for the entire range of the chip.

Ubiquiti AirOS has a built-in java applet that visualizes the spectral scan output. Unfortunately it does not provide any kind of documented API.

Support for atheros spectral scan has as far as we know not yet been implemented into the open source Linux kernel driver, but there is an effort to support this functionality. Here are some resources:

Frequency analyzer scan

Using a frequency analyzer or software-defined radio with a wide frequency range, it's possible to scan the actual frequency usage across the entire range of interest. Elonics 4000 can scan from 64 mhz to 2.2 ghz and only costs about $25.

OpenWRT setup

5 GHz

Here are some (possibly?) reasonable settings for a point to point or point to multi-point 5 GHz rooftop node:

channel 149
hwmode 11na 
htmode HT40
chanbw 20
txpower 20
noscan 1 (?)

These OpenWRT settings are documented here.

I'm not actually sure what the noscan feature does or whether it's a good idea. There's a warning that it will violate regulatory requirements, but I'm not sure if it's related to DFS or if it's something else. - (Juul (talk))

The ath9k driver does not yet have support for a 802.11n-only mode (also known as greenfield). The closest is 802.11n+a. Apparently you can prevent 802.11a nodes from associating somehow. source.

I'm not sure how to set the guard interval. The default is likely 800 ns, which we may want to lower for shorter links or heighten for longer links.

Hardware capabilities

5 GHz

Data rates for various settings can be looked up in this table.

Nanostation M5

Assuming 40 MHz channel and 800 ns guard interval:

 Setting | Speed        | Modulation | Tx power        | Rx sensitivity |
 ------------------------------------------------------------------------
   MCS12 | 162 Mbit/sec |     16-QAM | 22 dBm / 157 mW |        -84 dBm |
   MCS13 | 216 Mbit/sec |     64-QAM | 20 dBm / 100 mW |        -79 dBm |
   MCS14 | 243 Mbit/sec |     64-QAM | 18 dBm /  63 mW |        -78 dBm |
   MCS15 | 270 Mbit/sec |     64-QAM | 17 dBm /  50 mW |        -75 dBm |