On Sat, 11 Nov 2006 10:15:01 -0800, Jeff Liebermann wrote:
> Nicely done. However, methinks you missed on some assumptions and
> didn't supply the parameters you used to generate your numbers.
>
> See the FAQ at:
> | http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations
> for a sample link calculations. The receiver sensitivity should be at
> whatever connection rate you're expecting. My guess(tm) is that
> you're sharing a cable or DSL line and would like to see at least
> 6Mbit/sec thruput. That means you'll need at least a 12Mbit/sec
> wireless connection. Methinks some safety factor would be helpful, so
> I'll use 18Mbits/sec. The corresponding sensitivity for a Dlink
> DI-624 is -82dBm. Other hardware may be different, but this is in the
> ballpark. Also, everyone lies about rx sensitivity because it's so
> expensive and difficult to measure.
I would probably be sharing a 384 kbps connection from a wireless ISP
which is the only choice other than satellite out there. So, if the link
does no poorer than 1 Mbs, I think we'll be fine. This is just an idea
I'm working on if my brother would like something a bit faster than his
26.4 kbs dialup and I could tie the folks place in too. There will only
be one PC at either end of the link and rarely will both be in operation
at the same time.
> You didn't bother to mention the point to point distance. Reverse
> engineering your -109dB path loss, I'll guess it's 9500 ft (about 1.8
> miles). Is this correct?
I did neglect that. Per the coordinates, Splat! says it's 1.54 miles.
> You also neglected coax cable losses. Unless you buy an integrated
> antenna and radio device, there will be some coax connector and cable
> loss. I'll assume you have an external radio, a short piece of coax,
> and a pair of dish antennas.
I did toss in 2 dB unless I had an integrated unit. I did forget to
subtract 2dB at the far end and add in the 17 dBi antenna gain. OOPS!
> 20dB of fade margin is a good target value. Anything less than about
> 15dB will be flakey.
>
> +20dBm tx power is optimistic. There are some 100mw radios out there
> that claim this power level. Most commodity devices are rated at +17dBm
> and actually deliver about +15dBm xmit power.
>
> Now, we have enough to run the numbers. I'll do a template and you
> change the value to resemble your reality.
>
> TX power = +17dBm
> TX coax loss = 2 dB (3 ft LMR-240 plus connectors) TX ant gain =
> 17dBi
> Distance = 1.8 miles
> RX ant gain = 17dBi
> RX coax loss = 2 dB (3 ft LMR-240 plus connectors) RX sens =
> -82dBm
> Fade margin = unknown
>
> Plugging into:
> | http://www.terabeam.com/support/calculations/som.php I get a fade
> margin of 19.7dB. Close enough. It will work.
>
> Please note that this is the BEST case calculation. It always gets
> worse. You didn't mention if you were impacting the Fresnel Zone on
> your path. There may also be interference and reflection issues. Of
> course much of the hardware isn't up to specifications.
There is one other farmstead in the path, but I think we have pretty good
clearance. Of course, one won't know until it's tried. Like I said, a
tree or two may need to be trimmed.
> Sure. I'll assume a symmetrical transparent bridge. Just look for any
> of the access points that have a "bridge" mode. For example, Linksys
> WAP11, DLink DWL-2100AP, etc. Most access points will also play bridge.
> Check the feature list of web setup:
> http://www.linksysdata.com/ui/
> http://support.dlink.com/emulators/
> Figure on about $100 per radio. You'll probably need a PoE adapter and
> some kind of weatherproof package. Add about $50 per radio. Antennas,
> coax cables, connectors, and mounts will probably cost about $50. Yeah,
> it can be done.
I do appreciate that, Jeff. It helps to get a bit of direction and a bit
of refinement of my rough numbers. I spent about 10 minutes this morning
guessing all of that up. :-)
- Nate >>
--
"The optimist proclaims that we live in the best of all possible worlds,
the pessimist fears this is true."