Coverage Patterns

Friend of mine is getting cable Internet service, and is thinking
about going with a wireless router. House is a two-story colonial,
frame construction. Is there somewhere I could see the typical
coverage pattern of such routers? That is, something that would help
us decide where in the house the best location for it would be.

John Jones, Detroit

On Wed, 06 Oct 2004 13:03:51 GMT, John Jones
<(E-Mail Removed)> wrote:

>Friend of mine is getting cable Internet service, and is thinking
>about going with a wireless router. House is a two-story colonial,
>frame construction. Is there somewhere I could see the typical
>coverage pattern of such routers? That is, something that would help
>us decide where in the house the best location for it would be.

I looked but couldn't find anything for the just a router and a cheezy
coaxial antenna. It wouldn't do much good because most of such
patterns are done in "free space" (which means suspended in mid-air)
or over a mythical ground plane (which may not be realistic). You may
wanna look at:
http://www.freeantennas.com/2400/
for some nice JPG's of 4NEC2 program output for a WAP11 and DWL-1000
boxes in various configurations. Basically, the pattern is ideally
something like a donut, but is easily distorted by nearby objects.

Two story houses are always a problem. If you increase the gain of
the omni antenna, you narrow the vertical beamwidth, which reduces
coverage above and below the plane of the antenna. If you switch to
an panel (patch) antenna, you end up with a similar vertical radiation
problem.

As a rule-of-thumb for 2.4GHz:
Going through one wall is usually not a problem.
Two walls is a crap shoot and depends on equipment.
Three walls are impossible.
If there is aluminium foil or chicken wire in the walls, forget it.
A floor is thicker than a wall and more difficult to penetrate.
Radio waves do not make right angle turns down hallways.

I've tried using single radios on two story homes with limited
success. If there's an open stair well, with few partitions, it will
work with the radio near the stair well. However, mounted away from
any vertical passage ways, I usually end up with weird coverage
patterns. Most commonly, I get a varying signal, but can't maintain a
connection due to multipath and reflective nulls. I've had some
success using two diversity antennas. One remains on the access point
and the other ends up extended via coax cable to an antenna on another
floor. The problem is that the necessary coax cable is big, lossy,
ugly, and expensive. It's cheaper and better just to buy a 2nd
wireless access point instead of trying to stretch the antenna reach.

I suggest a "site survey" which in this case means installing a
temporary access point, and walking around the house with a PDA or
laptop running Netstumbler.


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558

Sorry Jeff, but I have to disagree with you. Three walls are not
"impossible". I just got a Netgear WGT624 Wireless 108mps Router and I'm
going through floor framing and SIX walls total. Typical 2-story house with
wood framing and this router totally blankets my house with a Wi-Fi signal.
At the farthest reach, the signal reduces to 11mps, which is still very
good. Directly below my router is my dining room where my wife surfs the
net with a 54mps connection.

I switched from a 54mps Linksys Wireless Access Point which barely worked at
the farthest reach. This new Netgear router is great and solved many
problems.

- Tom C.


"Jeff Liebermann" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> On Wed, 06 Oct 2004 13:03:51 GMT, John Jones
> <(E-Mail Removed)> wrote:
>
> >Friend of mine is getting cable Internet service, and is thinking
> >about going with a wireless router. House is a two-story colonial,
> >frame construction. Is there somewhere I could see the typical
> >coverage pattern of such routers? That is, something that would help
> >us decide where in the house the best location for it would be.
>
> I looked but couldn't find anything for the just a router and a cheezy
> coaxial antenna. It wouldn't do much good because most of such
> patterns are done in "free space" (which means suspended in mid-air)
> or over a mythical ground plane (which may not be realistic). You may
> wanna look at:
> http://www.freeantennas.com/2400/
> for some nice JPG's of 4NEC2 program output for a WAP11 and DWL-1000
> boxes in various configurations. Basically, the pattern is ideally
> something like a donut, but is easily distorted by nearby objects.
>
> Two story houses are always a problem. If you increase the gain of
> the omni antenna, you narrow the vertical beamwidth, which reduces
> coverage above and below the plane of the antenna. If you switch to
> an panel (patch) antenna, you end up with a similar vertical radiation
> problem.
>
> As a rule-of-thumb for 2.4GHz:
> Going through one wall is usually not a problem.
> Two walls is a crap shoot and depends on equipment.
> Three walls are impossible.
> If there is aluminium foil or chicken wire in the walls, forget it.
> A floor is thicker than a wall and more difficult to penetrate.
> Radio waves do not make right angle turns down hallways.
>
> I've tried using single radios on two story homes with limited
> success. If there's an open stair well, with few partitions, it will
> work with the radio near the stair well. However, mounted away from
> any vertical passage ways, I usually end up with weird coverage
> patterns. Most commonly, I get a varying signal, but can't maintain a
> connection due to multipath and reflective nulls. I've had some
> success using two diversity antennas. One remains on the access point
> and the other ends up extended via coax cable to an antenna on another
> floor. The problem is that the necessary coax cable is big, lossy,
> ugly, and expensive. It's cheaper and better just to buy a 2nd
> wireless access point instead of trying to stretch the antenna reach.
>
> I suggest a "site survey" which in this case means installing a
> temporary access point, and walking around the house with a PDA or
> laptop running Netstumbler.
>
>
> --
> Jeff Liebermann (E-Mail Removed)
> 150 Felker St #D http://www.LearnByDestroying.com
> Santa Cruz CA 95060 AE6KS 831-336-2558

On Wed, 06 Oct 2004 09:15:14 -0700, Jeff Liebermann
<(E-Mail Removed)> wrote:

>Basically, the pattern is ideally
>something like a donut, but is easily distorted by nearby objects.

That's what concerned me: if we put the router on the second floor,
would his laptop in the first floor family room, or the outside deck
next to that room, be able to connect.

>Two story houses are always a problem. If you increase the gain of
>the omni antenna, you narrow the vertical beamwidth, which reduces
>coverage above and below the plane of the antenna. If you switch to
>an panel (patch) antenna, you end up with a similar vertical radiation
>problem.
>
>As a rule-of-thumb for 2.4GHz:
>Going through one wall is usually not a problem.
>Two walls is a crap shoot and depends on equipment.
>Three walls are impossible.
>If there is aluminium foil or chicken wire in the walls, forget it.
>A floor is thicker than a wall and more difficult to penetrate.
>Radio waves do not make right angle turns down hallways.

Would a "g" router follow the same general pattern ("g"s are 5.8GHz,
right)?

>I've tried using single radios on two story homes with limited
>success. If there's an open stair well, with few partitions, it will
>work with the radio near the stair well. However, mounted away from
>any vertical passage ways, I usually end up with weird coverage
>patterns. Most commonly, I get a varying signal, but can't maintain a
>connection due to multipath and reflective nulls. I've had some
>success using two diversity antennas. One remains on the access point
>and the other ends up extended via coax cable to an antenna on another
>floor. The problem is that the necessary coax cable is big, lossy,
>ugly, and expensive. It's cheaper and better just to buy a 2nd
>wireless access point instead of trying to stretch the antenna reach.

OK

>I suggest a "site survey" which in this case means installing a
>temporary access point, and walking around the house with a PDA or
>laptop running Netstumbler.

He's got a "work" laptop, but I don't know what OS it's running, and
whether or not he'd have sufficient admin rights to install
NetStumbler. Good thought, though. And the other problem is: how can
we test this BEFORE Comcast does the install of the cable modem
splitter?

John Jones, Detroit

On Thu, 7 Oct 2004 14:30:04 -0700, "Tom C." <(E-Mail Removed)>
wrote:

>Typical 2-story house with
>wood framing and this router totally blankets my house with a Wi-Fi signal.

Where do you have it installed?

John Jones, Detroit

I installed the router on a shelf above my desk at the very back wall of our
house. Our house is longer than it is wide, so the backyard is out my
office window and the signal reaches up to the garage and holds more than
50% of the signal with around 75% signal quality. Surfing the web is very
fast - no slowdowns at all.


"John Jones" <(E-Mail Removed)> wrote in message
news(E-Mail Removed)...
> On Thu, 7 Oct 2004 14:30:04 -0700, "Tom C." <(E-Mail Removed)>
> wrote:
>
> >Typical 2-story house with
> >wood framing and this router totally blankets my house with a Wi-Fi
signal.
>
> Where do you have it installed?
>
> John Jones, Detroit

Oh yah - one more thing . . . my office is on the 2nd floor.


"John Jones" <(E-Mail Removed)> wrote in message
news(E-Mail Removed)...
> On Thu, 7 Oct 2004 14:30:04 -0700, "Tom C." <(E-Mail Removed)>
> wrote:
>
> >Typical 2-story house with
> >wood framing and this router totally blankets my house with a Wi-Fi
signal.
>
> Where do you have it installed?
>
> John Jones, Detroit

On Thu, 7 Oct 2004 14:30:04 -0700, "Tom C." <(E-Mail Removed)>
wrote:

>Sorry Jeff, but I have to disagree with you. Three walls are not
>"impossible". I just got a Netgear WGT624 Wireless 108mps Router and I'm
>going through floor framing and SIX walls total. Typical 2-story house with
>wood framing and this router totally blankets my house with a Wi-Fi signal.
>At the farthest reach, the signal reduces to 11mps, which is still very
>good. Directly below my router is my dining room where my wife surfs the
>net with a 54mps connection.
>
>I switched from a 54mps Linksys Wireless Access Point which barely worked at
>the farthest reach. This new Netgear router is great and solved many
>problems.
>
>- Tom C.

Whee... Why don't I have such good luck? I have never made it
through that many walls with a reliable connection. What's your
secret? What client radios are you using? Any open doorways and
hallways? Who's your sorcerer?

Let's play with the numbers and see where they lead.

The WRT624 belches about +13dBm into a +2dBi antenna for a +15dBm
EIRP. At the other end is probably (my guess), an laptop with a built
in antenna. At 11Mbits/sec, it has about -83dBm sensitivity. Antenna
gain is a disgusting -2dBi.

When I did my own testing in 1999, the average wood framed residential
wall was good for -6dB of loss, while the floors were -10dB. Note
that this is only the *ADDED* loss contributed by the walls and
floors, and do not include any path loss over a distance.

My guess is your overall range is about 75ft from one end of the house
to the other.

Free space loss (dB) = 96.6 + 20 log F + 20 log D
Where:
F = Frequency (GHz)
D= Distance (miles)
logs are base 10, 5280 ft/mile
FSL = 96.6 + (20 log 2.4) + (20 log 75/5280)
FSL = 96.6 + 7.6 -40 = 64.2 dB loss.
(Somebody please check my math. I haven't had my morning coffee yet).

So, we start with the TX EIRP of +15dBm, loose 64.2dB over the
distance, and end up with a receive sensitivity of -83dBm. The fade
margin is:
+15dBm - 64.2dB -(-83dBm) = 34dB
That means that we can lose 34dB more signal and still just barely
communicate with the laptop over a distance of 75ft.

With my measurements of about -6dB for walls, and -10dB for floors,
that will give you about 5 walls, or 1 floor and 3 walls which can be
penetrated.

However, please note that the receiver sensitivity reference level is
at a BER (bit error rate) of 10^5, which is fairly lossy and would not
provide reliable communications. You would need at least +10dB more
signal to make it reliable. Therefore, my guess would be 4 walls, or
1 floor and 2 walls maximum.

Drivel: When I was making the loss measurements, we had a very
difficult time making sense of the numbers. A wall should be 6dB per
wall and should be linear for additional walls. Instead, additional
walls contributed less and less loss for each additional wall. After
about 4 walls, the signal level was strictly free space loss and the
additional walls had little effect. Huh? Eventually, we figured out
that as the number of walls increased, the tendency for the signal to
leak AROUND the walls became more pronounced. It was going out
windows, bouncing around stairwells, and generally avoided going
through the walls. This might be what's happening here. Any
possibility that instead of going through 6 walls, your signal is
going around some of them?


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558

On Thu, 07 Oct 2004 21:33:31 GMT, John Jones
<(E-Mail Removed)> wrote:

>Would a "g" router follow the same general pattern ("g"s are 5.8GHz,
>right)?

You've got the letters mixed up.
802.11 2.4GHz 1 or 2 Mbits/sec
802.11a 5.7GHz OFDM up to 54Mbits/sec
802.11b 2.4GHz 1, 2, 5.5, and 11 Mbits/sec
802.11g 2.4Ghz OFDM up to 54Mbits/sec
(or more with proprietary enhancements).

At 2.4GHz, the antenna patterns would be the same no matter what type
of modulation or data rate. However (insert drum roll), the faster
you go, the "stronger" the signal you need to maintain the data rate.
More accurately, a better S/N (signal to noise) ratio is needed at
higher data rates. The pattern is the same, but you'll need to be
closer to the access point in order to obtain the faster speeds.

5.7Ghz antennas have about twice the gain (+3dB) as at 2.4GHz, for the
same size antenna. However, nothing is free in this world, so the
vertical beamwidth is even narrower at 5.7GHz. Also, free space loss
is about +3dB more at 5.7Ghz. Unfortunately, many dual band antennas,
sound on 802.11a/b/g multimode access points are nothing but a best
compromise between 2.4Ghz and 5.7Ghz operation. The result is usually
a rather ugly looking "cloverleaf" type of pattern, that puts lobes
where you least need them, and almost eliminates the horizontal main
lobe at 5.7GHz.

>He's got a "work" laptop, but I don't know what OS it's running, and
>whether or not he'd have sufficient admin rights to install
>NetStumbler. Good thought, though. And the other problem is: how can
>we test this BEFORE Comcast does the install of the cable modem
>splitter?

Laptops are good. I run into the problem of the "corporate" laptop
all the time. Many users do not have administrator rights to their
own machines. So, I setup a bootable Linux cdrom with all the
necessary tools. See:
http://www.remote-exploit.org/?page=auditor
for a very nice cdrom full of wireless and wired LAN tools. Boot it
and run. If the customer is paranoid, pull the hard disk and leave it
out while testing. Try Kismet instead of Netstumbler:
http://www.kismetwireless.net/documentation.shtml


--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831.336.2558 voice http://www.LearnByDestroying.com
# (E-Mail Removed)
# 831.421.6491 digital_pager (E-Mail Removed) AE6KS

Very nice cdrom of software. Sure seems to be downloading slow via HTTP.
Couldn't get the FTP to connect using WSFTP

--
Bob Alston

bobalston9 AT aol DOT com
"Jeff Liebermann" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> On Thu, 07 Oct 2004 21:33:31 GMT, John Jones
> <(E-Mail Removed)> wrote:
>
>>Would a "g" router follow the same general pattern ("g"s are 5.8GHz,
>>right)?
>
> You've got the letters mixed up.
> 802.11 2.4GHz 1 or 2 Mbits/sec
> 802.11a 5.7GHz OFDM up to 54Mbits/sec
> 802.11b 2.4GHz 1, 2, 5.5, and 11 Mbits/sec
> 802.11g 2.4Ghz OFDM up to 54Mbits/sec
> (or more with proprietary enhancements).
>
> At 2.4GHz, the antenna patterns would be the same no matter what type
> of modulation or data rate. However (insert drum roll), the faster
> you go, the "stronger" the signal you need to maintain the data rate.
> More accurately, a better S/N (signal to noise) ratio is needed at
> higher data rates. The pattern is the same, but you'll need to be
> closer to the access point in order to obtain the faster speeds.
>
> 5.7Ghz antennas have about twice the gain (+3dB) as at 2.4GHz, for the
> same size antenna. However, nothing is free in this world, so the
> vertical beamwidth is even narrower at 5.7GHz. Also, free space loss
> is about +3dB more at 5.7Ghz. Unfortunately, many dual band antennas,
> sound on 802.11a/b/g multimode access points are nothing but a best
> compromise between 2.4Ghz and 5.7Ghz operation. The result is usually
> a rather ugly looking "cloverleaf" type of pattern, that puts lobes
> where you least need them, and almost eliminates the horizontal main
> lobe at 5.7GHz.
>
>>He's got a "work" laptop, but I don't know what OS it's running, and
>>whether or not he'd have sufficient admin rights to install
>>NetStumbler. Good thought, though. And the other problem is: how can
>>we test this BEFORE Comcast does the install of the cable modem
>>splitter?
>
> Laptops are good. I run into the problem of the "corporate" laptop
> all the time. Many users do not have administrator rights to their
> own machines. So, I setup a bootable Linux cdrom with all the
> necessary tools. See:
> http://www.remote-exploit.org/?page=auditor
> for a very nice cdrom full of wireless and wired LAN tools. Boot it
> and run. If the customer is paranoid, pull the hard disk and leave it
> out while testing. Try Kismet instead of Netstumbler:
> http://www.kismetwireless.net/documentation.shtml
>
>
> --
> # Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
> # 831.336.2558 voice http://www.LearnByDestroying.com
> # (E-Mail Removed)
> # 831.421.6491 digital_pager (E-Mail Removed) AE6KS


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.774 / Virus Database: 521 - Release Date: 10/7/2004