[ previous ] [ next ] [ threads ]
 From:  "Daniele Guazzoni" <daniele dot guazzoni at gcomm dot ch>
 To:  <m0n0wall at lists dot m0n0 dot ch>
 Subject:  RE: [m0n0wall] Wireless g 54M
 Date:  Wed, 18 Aug 2004 23:37:54 +0200
Didn't tried yet to get further but it's not gonna be any more than that.
The biggest problem at those distance is not attenuation but delay.
The timing of 802.11 are a big concern.
We used Linksys a pair of WAP54g and got the support of a former Broadcom
engineer to "trick out" the timing.

The original idea was to build a 802.11g backbone around the Zurich big area
to feed 802.11b hotspots.
This was the reason for the legal levels.

Beside the timing, the preamps for 2.4GHz where quite an headache as they
got the misbehaviour to start oscillating.
I spent almost 2 weeks to get this selfmade preamps to work correctly.

And finally with a 50km link you just need to think "now I touch the dish"
and your signal is already fading...
1 degree of antennas alignement error results in no link at all.

Without professional equipment and skills I would say that 20 to 30km is the
pain limit.


-----Original Message-----
From: Pat Ellison [mailto:pat at zbit dot net] 
Sent: Mittwoch, 18. August 2004 22:48
To: daniele dot guazzoni at gcomm dot ch; m0n0wall at lists dot m0n0 dot ch
Subject: RE: [m0n0wall] Wireless g 54M

>- my personal legal record (within the +20dBm limit) is 50km with 
>  using 1.2m dishes and a 25dB 2.4GHz RX-preamp.

That's a good haul! 50km I don't think I ever heard of that distance.
What's your illegal record?  ;) ha ha

I am not connected in any way to any policing authorities!

-----Original Message-----
From: Daniele Guazzoni [mailto:daniele dot guazzoni at gcomm dot ch]
Sent: Tuesday, August 17, 2004 6:17 PM
To: m0n0wall at lists dot m0n0 dot ch
Subject: RE: [m0n0wall] Wireless g 54M

Out of topic but...

my 2 cents about wireless:

About bitrate and throughput:
- with 802.11b you have a RF bitrate of 11 Mbps wich means a pure IP
throughput of about 4 Mbps.
- 802.11g provides 54 Mbps of RF bitrate, also about 22 Mbps of IP
- 802.11a has about the same properties of the "g" standard.
- some chipset maker have introduced proprietary (but backward compatible)
pseudo-standards like 22 and 100 Mbps
  This is obtained by grouping several frame together before RF encoding,
thus reducing the overhead.
  Practically you get more "speed" only if you can fill up frames and this
is not always the case.
- Of course if you are 500m apart from your AP you will have more errors,
retries and therefore less throughput.

Technology and Compatibility:
- 802.11b and "g" works on 2.4GHz and g is backward compatible with b
- 802.44a is on 5GHz and compatible with itself.
- Depending on your local low you may be allowed or not to use 802.11a for
outer links (not inside a building).

World standards:
- I know in US you can use more TX power than in Europe (ETSI limit to
+20dBm EIRP = 100mW)

- the "g" standard use more bandwidth then "b" and is therefore less
  For instance Linksys specs "b" with -80dBm and "g" with -65dBm.
- since "a" and "b/g" are on different frequency ranges, radio propagation
may cosiderably differ.
- my personal legal record (within the +20dBm limit) is 50km with 802.11g
using 1.2m dishes and a 25dB 2.4GHz RX-preamp.
  The link provided 19Mbps of IP thoughput.

Settings and tips:
- on most "new" AP for 802.11g you can set the operating mode to "g-only"
thus you switch off the "b" compatibility.
  This ensures that no one can tear down your AP to "b" speeds.
- Be sure you upgrade the firmware !!!
  There may be huge difference between a "g-draft" and the standard release.
- Choose carefully the emplacement of your AP and antennas.
  Run some tests, for instance the distance between antennas and walls is
  A move few cm (or inches) can act as a wonder.
- Some AP allows to set the TX power (either per default or via "hacked"
  Also the default TX power varies: most AP in europe are set to +15dBm,
other to +12. I also saw a +8.
  6dB less means about half the distance !!!
- I you cannot "push" more out of your AP, get external antennas.


-----Original Message-----
From: Fred Wright [mailto:fw at well dot com]
Sent: Dienstag, 17. August 2004 22:41
To: m0n0wall at lists dot m0n0 dot ch
Subject: RE: [m0n0wall] Wireless g 54M

On Tue, 17 Aug 2004, Pat Ellison wrote:

> Now that is my problem.... I've been so happy with this system that I 
> neglected to learn anything about the A/B/G...etc systems. Not I'm 
> forced to become a quick study..

One thing I don't know is whether the A/B/G cards can run A simultaneously
with B/G.  My guess is that the answer is no, since it's *very* difficult to
design something that can send and receive at the same time through the same
antenna, even in different frequency bands, and sharing the transmit/receive
scheduling between the two would get messy, if it's even possible.

On Tue, 17 Aug 2004, Chet Harvey wrote:

> Just a side note, I am always curious as to why people think they need 
> such fast rates like 75 to 100mbps. Most businesses I see have 512 or 
> 768
> Some go crazy and get 10mb pipes so unless you are moving A LOT of 
> data internally I just dont get it. I'd rather have range and signal 
> strength over high transfer. As signal strength drops so does transfer.

I presume the "512 or 768 pipes" you're talking about is for the WAN.  I
can't imagine anyone wanting a LAN that slow.  It's not even that fast for a
WAN; I have 6M/600k DSL at my *house*. :-)

Anytime you start doing something ike file sharing, it's almost impossible
to have too much bandwidth.  Or using VNC, though that's mostly VNC's fault.
:-) VNC is annoyingly slow over Gb Ethernet.

On Tue, 17 Aug 2004, Robert Staph wrote:
> From: "Fred Wright" <fw at well dot com>
> Sent: Monday, August 16, 2004 3:53 PM
> > 1) How poor the range usually is for G (*much* shorter than B with 
> > the same equipment).
> Actually I've found my Linksys WAP54G to have superior range than my 
> Linksys BEFW11S4.  I can get 36M out to where the B signal was 
> dropping to 5M.  I

Well, I did say "same equipment". :-)

> > 2) How the presence of a single B station causes all the Gs to drop 
> > down to B.  I'm not sure it even needs to be an "authorized" B 
> > station to have this effect.
> I have a B AP about 50-60 feet from my G AP, no issues getting beyond 
> B speeds when connected to the G AP.  Different SSID's and channels (1 
> and 6), one WEP one not, one does its own DHCP other doesn't, two very 
> different configs.  However since they do not share a SSID, I would 
> think it possible a B station and G station using the same SSID and 
> config might cause the G client to dumb down and not talk G (or if the 
> AP's are very, very close; or both are passing as much traffic as they 
> can

I wasn't referring to independent networks, but to the "upward
compatibility" feature of G with B.  This is in the same sense that 66MHz
PCI is "upward compatible" with 33MHz PCI, i.e. plug in a single 33MHz card
and everything switches to 33MHz.  Similarly, *on the same network*, the
presence of a B station drops everything to B speeds.

On Tue, 17 Aug 2004, Pat Ellison wrote:

> I know what you mean.. it's like a family going to Circuit City for a
> computer..
> "So sir, what are you looking to do with your new computer?... Well 
> the wife plays Solitaire.."
> so the salesman needs to make sure that the computer is suited for his 
> needs.. "oh well then, We have this 3.2 Ghz Intel with Hyper 
> Threading(tm)..."

Hey, don't knock HyperThreading.  Recently a case came up where Linux needed
HT just to be able to output text to the console at a reasonable speed. :-)

					Fred Wright

To unsubscribe, e-mail: m0n0wall dash unsubscribe at lists dot m0n0 dot ch
For additional commands, e-mail: m0n0wall dash help at lists dot m0n0 dot ch

To unsubscribe, e-mail: m0n0wall dash unsubscribe at lists dot m0n0 dot ch
For additional commands, e-mail: m0n0wall dash help at lists dot m0n0 dot ch