[Fwd: RF performance of a AR2317 based SoC device - was: Re: Questions about the Mesh Potato]

Interesting bit of info...

-------- Original Message --------
Subject: RF performance of a AR2317 based SoC device - was: Re:
Questions about the Mesh Potato
Date: Sun, 16 Aug 2009 11:56:38 +0200
From: elektra <onelektra@gmx.net>
Reply-To: village-telco-dev@googlegroups.com
To: village-telco-dev@googlegroups.com
References:
<6b08b4a5-6847-4784-9828-bb9603d5bb94@w6g2000yqw.googlegroups.com>
<1250382825.31167.14.camel@localhost>

Hello Lew!

As the Mesh-Potato hasn't reached the state of its first mass production
design we don't have exact figures of transmitter and receiver performance
yet.

But let me show you some example calculations based on the D-Link DIR-300.

The Mesh-Potato is based on the Atheros AR2317 Wireless SoC (SoC =
System on
Chip) integrated circuit. The D-Link DIR-300 is a commodity product for
SOHO
(Small Office / Home Office) applications, which is using the same chip.
The
D-Link DIR-300 manual contains these specs regarding transmit (TX) and
receive (RX) performance:

TX power: 15 dBm +/- 2 dBm

RX sensitivity:
• 54 Mbit/s OFDM, 10 % PER, -68 dBm)
• 48 Mbit/s OFDM, 10 % PER, -68 dBm)
• 36 Mbit/s OFDM, 10 % PER, -75 dBm)
• 24 Mbit/s OFDM, 10 % PER, -79 dBm)
• 18 Mbit/s OFDM, 10 % PER, -82 dBm)
• 12 Mbit/s OFDM, 10 % PER, -84 dBm)
• 11 Mbit/s CCK, 8% PER, -82 dBm)
• 9 Mbit/s OFDM, 10 % PER, -87 dBm)
• 6 Mbit/s OFDM, 10 % PER, -88 dBm)
• 5,5 Mbit/s CCK, 8% PER, -85 dBm)
• 2 Mbit/s QPSK, 8% PER, -86 dBm)
• 1 Mbit/s BPSK, 8% PER, -89 dBm)

PER = Packet Error Rate

Let's do some simple and optimistic radio performance calculations based on
these figures.

The DIR-300 is shipped with a sleeve dipole antenna, which - if implemented
well - should have a gain of 2.15 dBi. It would be nitpicking to
mention the
0.2dB loss of the R-SMA antenna plug ;-) So we can just assume a antenna
gain
of 2dBi.

I guess the values given by D-Link refer to the performance measured at the
antenna socket, but I could be wrong.

Based on these figures one can make a quick range / data rate estimation
between two DIR-300 under ideal conditions (free fresnel zone, considering
attenuation caused by free space loss only, no interference on the radio
channel). If you want to make calculations with obstructions in the
propagation path I'd strongly recommend playing with the free "radio
mobile"
software, if you don't know that software already. It is a truly amazing
software.

A 1 km radio link under ideal conditions (clear fresnel zone) has a free
space
attenuation of ~100 dB @ 2.45 GHz, 2km attenuation is ~106 dB, 4 km
attenuation is ~112 dB, 8 km is ~118 dB and so on.

With 15 dBm effective radiated transmit power the signal after 1 km is down
at -85 dBm (subtracting 100 dB free space loss). Taking the gain of the
antennas into account (+2 dB when we transmit on one side, + 2 dB when we
receive on the other side) we should have a signal level at the antenna
port
of the RX side of -81 dBm, which should be good enough to decode 18
Mbit. If
we double the distance to 2km we can still operate at 6 or 9 Mbit (-87
dBm).
Increasing the distance by 50 % more would add 3 dB more free space
loss, so
we end up at -90 dBm signal level after 3km. This is now lower than the -89
dBm for the lowest bitrate of 1 Mbit. However at 1 Mbit the TX power is
slightly higher, so with a bit of luck the connectivity will be at its
absolute edge after 3 km, but still there.

Keep in mind these estimations are based under the assumption of ideal
conditions: No rain or hail storm. No interference from other radios. Low
noise floor. Absolutely no obstructions or reflections in the fresnel zone.
No birds sitting on the antennas. The specs provided are assumed to be
correct. The timing of the MAC layer must be adjusted with the athctrl
utility for distances greater than 1 km.

However the DIR-300 is just a commodity product intended to be used indoors
and not a reference design intended to show off with brilliant radio
performance. Its design goal is least cost with moderate radio
performance. I
can buy those devices for ~27 € per unit retail in Germany via mail-order
plus shipping, which is really cheap. Actually the AR2317 chipset can
perform
better than the specs of the DIR-300, but it is good enough for the market
and the application, given the price.

The MP is not a SOHO product. It will be primarily used outdoors in densely
and sparsely populated areas (and anything in between, of course). The
great
many will be used in densely populated areas, but it should be also capable
to make a difference in rural areas. For both applications good receiver
performance is critical. In densely populated areas the network should
operate at the highest possible bitrate - simply to keep the amount of
airtime consumed by each transmission low and the channel capacity high.
Good receiver performance helps to keep the data rate fast and allows to
keep
the TX power levels low. The radios need to detect and coordinate
channel use
to avoid collisions on the radio channels as well. Low receiver performance
will have a negative effect on channel capacity and increase the noise
floor.

The TX power of the MP can be adjusted to 8 different levels:

0 dBm (1 mW)
5 dBm (3 mW)
7 dBm (5 mW)
9 dBm (7 mW)
11 dBm (12 mW)
13 dBm (19 mW)
15 dBm (31 mW)
17 dBm (50 mW)

In densely populated areas where people live door by door a few mW TX power
can be enough.

In sparsely populated areas good RX performance helps to achieve good
range,
as we have seen before. Interference from other wireless networks is not so
critical, so if we put the devices high enough we can often work with
good RF
conditions. A internal antenna with a slight gain of 3 dBi would be
compliant
with the regulatory limit of 20 dBm effective radiated power present in
many
countries.

Cheers,
Elektra

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google
Groups "village-telco-dev" group.
To post to this group, send email to village-telco-dev@googlegroups.com
To unsubscribe from this group, send email to
village-telco-dev+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/village-telco-dev?hl=en
-~----------~----~----~----~------~----~------~--~---