[CUWiN-Dev] Re: Alchemy CPU

[Quantum Scientific]

On Sunday 23 January 2005 22:54, David Young wrote:
> Here's a capable little device,
> <http://www.netgear.com/products/details/WGT634U.php>, that costs
> about $80.  The approach I suggest is to license the reference design,
> change out the parts that are not outdoors-proof, and get an outdoor
> enclosure w/ PoE and radio ports made.  Does it make sense?

Yes.  My only reservation is the lack of flexibility with the radio.  That particular router has the Atheros chipset on-board, ie not in a daughtercard, so it could never be changed.  This does mean lower cost though.  Also, the WGT is not directly built from an Atheros reference design as far as I can tell;  only loosely based on the miniPCI reference design.  

I agree, the AR5004 chipset is wonderful, with purported Rx sensitivity of -105dBm and low current draw, but I have been unable to increase Tx power in this unit due to the MicroSemi 5510 power amp's limitation of +20dBm (100mW).  With Atheros' dynamic power adjustment, that's probably more like +18dBm (64mW).  Replacing this with the 5512 PA is questionable, due to different pinout and limited physical space, unless we have complete control over the circuit.  This would mean FCC certification by a lab though.  Macom makes the best amps, but they need external matching to 50 ohms, which requires much better than the pitiful software tools I have.

Netgear has been unable to tell me simply what the Tx power and Rx sensitivity is for the WG511U, WG311T, or WGT634U after a month of trying, even at the highest levels of Tech Support.  So I doubt the real brains are in Netgear.  They seem to only throw others' designs.


I am interested in reliable signal.  This means:
- horizontal polarization, or
- high power, or 
- 700/900MHz  

Horizontal polarization is great with its 20dB seperation from vertical, and is semi-immune to lightning (which is vertical) and most other noise, but omni antennas are expensive. (>$150)  

High power can be planned for.  

700 or 900MHz is the best option for low noise, penetrating trees & structures, *and* working over distance, but the radio will be more expensive.  Motorola's newest Canopy uses 900MHz, and 700MHz is an untapped unlicensed band.  You know, a pine needle is about the length of a 2.4GHz 1/4 wave antenna... so a fir tree is like a big, resonating 802.11-absorbing pillow.