getting good range .. - ditching USB wireless adaptor. What PCI one?

I have a MIMO 802.11g router, and USB wireless adaptor.

range is insufficient.

I measured signal to noise ratio in netstumbler on the comp with the
usb thing. It went from -85dbi to -95dbi and disconnected (dunno if
was db or dbm or dbi ). But apparently that indicates the outcome of
signal strength /power over the distance (and there's no other easy
way of measuring signal level or power). What is a good value for
that signal to noise ratio?

I read that USB wireless has a rubbish internal "chip" antenna , and
that makes for small range. I have one and it doesn't go past 2
flights of stairs.

I'd like a wireless 802.11g MIMO PCI Card, that comes with a
reasonable antenna, but i'd also like it to have a connector for an
antenna. (so I only have to buy an external antenna if it's not good
without one).

Changing the router isn't an option at the moment.

(I did try pointing the aerials, including so the length of the aerial
as oppose to the tip, faces the wireless usb adaptor, but no use. I
didn't try moving the wireless usb adaptor around on a usb cable. I
don't plan on getting a wireless usb adaptor with an antenna, but I
might)

I'd like to know the power of these things before buying.. I've heard
that Buffalo make powerful wireless stuff. But they don't seem to have
any PCI, and I can't see the power rating.

(the buffalo website is a silly design anyway. each irrelevant
picture links to a directory, the directory's name reveals what it's
about. - so if you ever want to browse it, that's the way!)

I know linksys show power rating, but a few have recommended buffalo.

"(E-Mail Removed)" <(E-Mail Removed)> hath wroth:

>I have a MIMO 802.11g router, and USB wireless adaptor.

You have a problem. Let me know when you are able to read the label
and supply the vendor and model numbers of your inadequate equipment.

>range is insufficient.

So are your expectations. Could I trouble you to describe what manner
of range you were expecting? Over what terrain? Through how many
walls?

>What is a good value for
>that signal to noise ratio?

It depends on the connection speed. For a BER (bit error rate) of
about 1 error in 10^5 bits, or about 10% PER (packet error rate):
Speed SNR(dB)
11 6.99
5.5 5.98
2 1.59
1 -2.92
54 24.6
48 24.1
36 18.8
24 17.0
18 10.8
12 9.0
9 7.8
6 6.0
See:
<http://groups.google.com/group/alt.internet.wireless/msg/b806cb8cf93c9724>
for an explanation. What I managed to leave out is that these figures
are for a fairly lousy error rate. One needs to add in the
anticipated fade margin to get real numbers. Add about 20dB to each
of the SNR numbers for what is considered "good values" for viewing
with Netstumbler.

>I read that USB wireless has a rubbish internal "chip" antenna , and
>that makes for small range. I have one and it doesn't go past 2
>flights of stairs.

My stairs do not fly. You're correct. USB dongles have very small
antennas. Antenna gain is very roughly proportional to the size of
the antenna. The tiny chip or PIFA antennas have a gain of about
-6dBi to optimistically 0dBi gain. A typical non-MIMO PCMCIA card
will have an antenna gain of about 2dBi, but it's rather directional.

Please note that your flight of stairs implies that you have a change
of vertical elevation and that you might be going through a floor or
ceiling. The pattern from the USB dongle is a close approximation of
a hemisphere, so orientation isn't terribly important. However,
that's not the case with the typical access point vertical coaxial
antenna, especially if it's an add-on higher gain antenna. These
trade gain in the vertical and horizontal direction (i.e through the
floor and ceiling) in trade for more gain in the horizontal direction.

>I'd like a wireless 802.11g MIMO PCI Card, that comes with a
>reasonable antenna, but i'd also like it to have a connector for an
>antenna. (so I only have to buy an external antenna if it's not good
>without one).

Forget it for now. MIMO functionality relys heavily on the matching
between 3 or more antennas. As a result, an external antenna would
need to have 3 antennas, 3 coax cables, and 3 connectors. There are
some with external antennas, such as Linksys WMP300N but they're all
PCI cards, where there's no other alternative to locating an antenna.
All the other clients and access points have non-removable antennas or
internal antennas.

>Changing the router isn't an option at the moment.

Change what router into what? If you're going to practice
enchantments and transmogrification, perhaps a practicing wizard or
alchemist would be more helpful.

>(I did try pointing the aerials, including so the length of the aerial
>as oppose to the tip, faces the wireless usb adaptor, but no use. I
>didn't try moving the wireless usb adaptor around on a usb cable. I
>don't plan on getting a wireless usb adaptor with an antenna, but I
>might)

Indecision is the key to flexibility.

>I'd like to know the power of these things before buying.. I've heard
>that Buffalo make powerful wireless stuff. But they don't seem to have
>any PCI, and I can't see the power rating.

The transmit power is usually specified on the manufactories data
sheets. If that's not available, it can be found on the FCC ID search
page under the type certification test data. Buffalo makes various
models. The WHR-HP-G54 transmits up to 250mw. 50mw is the default
value.

>(the buffalo website is a silly design anyway. each irrelevant
>picture links to a directory, the directory's name reveals what it's
>about. - so if you ever want to browse it, that's the way!)

Right. The product quality must be directly related to the quality of
the web site.

>I know linksys show power rating, but a few have recommended buffalo.

Both manufacturers have their benefits and detriments. I don't want
to go into a point by point comparison. See:
<http://www.smallnetbuilder.com/component/option,com_chart/Itemid,189/>
for reviews and comparison charts.

Incidentally, there may be some confusion here when using the latest
MIMO buzzword. In the context of this discussion, there are 3 types
of MIMO.
1. Airgo (Qualcomm) based chipset which in my opinion is "true" MIMO.
<http://www.cdmatech.com/products/wlan.jsp>
2. Ruckus Wireless based chipsets, which are basically beam steering,
beam forming, or antenna aiming systems.
<http://www.ruckuswireless.com/technology/beamflex.php>
3. Gross misuse of the MIMO term by Buffalo on the WHR-HP-G54
claiming "MIMO like performance" with having any features that even
remotely involve MIMO technology.
<http://www.buffalotech.com/technology/our-technology/high-power/>

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann <(E-Mail Removed)> hath wroth:

>>What is a good value for
>>that signal to noise ratio?
>
>It depends on the connection speed. For a BER (bit error rate) of
>about 1 error in 10^5 bits, or about 10% PER (packet error rate):
> Speed SNR(dB)
> 11 6.99
> 5.5 5.98
> 2 1.59
> 1 -2.92
> 54 24.6
> 48 24.1
> 36 18.8
> 24 17.0
> 18 10.8
> 12 9.0
> 9 7.8
> 6 6.0
>See:
><http://groups.google.com/group/alt.internet.wireless/msg/b806cb8cf93c9724>
>for an explanation. What I managed to leave out is that these figures
>are for a fairly lousy error rate. One needs to add in the
>anticipated fade margin to get real numbers. Add about 20dB to each
>of the SNR numbers for what is considered "good values" for viewing
>with Netstumbler.

Duh. I forgot that Netstumbler does its probe requests at the slowest
speed (1Mbit/sec). Therefore, the wireless speeds do not enter into
the puzzle when using Netstumbler. However, if you are looking at the
SNR reported by the wireless client, it should match the table plus
about 20dB. The way the access point operates is to adjust the speed
so that the error rate is reasonable. I'm not sure of the exact
target value, but I think 20dB is about right. If the error rate
creeps up, then the access point slows down the wireless speed to
compensate. Therefore, with the access point speed set to "auto", you
should always see about 20dB or better SNR.

I guess about 20dB would be a good minimum SNR for 1Mbit/sec with
Netstumbler.

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann <(E-Mail Removed)> hath wroth:

Perhaps I'll get it right on the 3rd try. (At home today with a fever
and mild cold. Brain is currently somewhat malfunctional).

>Duh. I forgot that Netstumbler does its probe requests at the slowest
>speed (1Mbit/sec). Therefore, the wireless speeds do not enter into
>the puzzle when using Netstumbler. However, if you are looking at the
>SNR reported by the wireless client, it should match the table plus
>about 20dB. The way the access point operates is to adjust the speed
>so that the error rate is reasonable.

So far, that's correct. The access point adjusts the speed so that
the *error rate* is constant, not the SNR. The SNR will vary with the
connection speed.

>I'm not sure of the exact
>target value, but I think 20dB is about right.

I'm not sure of the exact target value, but I think a BER of about
1*10^6 is about right.

>If the error rate
>creeps up, then the access point slows down the wireless speed to
>compensate.

>Therefore, with the access point speed set to "auto", you
>should always see about 20dB or better SNR.

Therefore, with the access point speed set to "auto", you should
always see 6dB to perhaps 40dB or better SNR depending on wireless
speed.

>I guess about 20dB would be a good minimum SNR for 1Mbit/sec with
>Netstumbler.

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann <(E-Mail Removed)> wrote:
> The pattern from the USB dongle is a close approximation of
> a hemisphere, so orientation isn't terribly important.

I don't agree.

--
Clarence A Dold - Hidden Valley Lake, CA, USA GPS: 38.8,-122.5
http://cdold.home.mchsi.com/Solar-generation.htm

(E-Mail Removed) hath wroth:

>Jeff Liebermann <(E-Mail Removed)> wrote:
>> The pattern from the USB dongle is a close approximation of
>> a hemisphere, so orientation isn't terribly important.

>I don't agree.

Picky, picky. Ok, maybe not so good an approximation of a hemisphere.
The basic idea is that the smaller the antenna, the better
approximation of a hemispherical radiator it appears. The bigger PIFA
antennas will be more erratic. Of course, there's a big hole in the
direction of the dongle and cable.

Here are some realistic chip antenna patterns:
<http://www.rainsun.com/Products/LTCC/LTCC.html>
<http://www.antennafactor.com/documents/ANT-xxx-CHP-x_Data_Guide.pdf>
Yech. Ok, I give up. Something kinda sorta roughly vaguely
resembling a hemisphere.

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann <(E-Mail Removed)> wrote:
> (E-Mail Removed) hath wroth:

> >Jeff Liebermann <(E-Mail Removed)> wrote:
> >> The pattern from the USB dongle is a close approximation of
> >> a hemisphere, so orientation isn't terribly important.

> >I don't agree.

> Picky, picky. Ok, maybe not so good an approximation of a hemisphere.
> The basic idea is that the smaller the antenna, the better
> approximation of a hemispherical radiator it appears. The bigger PIFA
> antennas will be more erratic. Of course, there's a big hole in the
> direction of the dongle and cable.

Which way are you suggesting that the hemisphere is oriented? I always
thought that you were picturing a hemisphere over the top of a dongle, with
the cable exiting the bottom. That would mean the cable would be the
nonexistent hemi of the sphere.

> Here are some realistic chip antenna patterns:
> <http://www.rainsun.com/Products/LTCC/LTCC.html>
> <http://www.antennafactor.com/documents/ANT-xxx-CHP-x_Data_Guide.pdf>
> Yech. Ok, I give up. Something kinda sorta roughly vaguely
> resembling a hemisphere.

Those are the kind of lumps I see from signal strength measurements, except
that I didn't see an equal back lobe. Putting that chip antenna on one
side of a coper plated board might have a different signal pattern that the
theoretical chip in space. The signal seemed to be moderately directional
off the face of the dongle. Out the front of the can in my photo, if the
can weren't there.

http://www.rahul.net/dold/clarence/u...42-800x600.jpg

--
Clarence A Dold - Hidden Valley Lake, CA, USA GPS: 38.8,-122.5

On 3 May, 05:05, Jeff Liebermann <j...@comix.santa-cruz.ca.us> wrote:
> "jameshanle...@yahoo.co.uk" <jameshanle...@yahoo.co.uk> hath wroth:
>
> >I have a MIMO 802.11g router, and USB wireless adaptor.
>
> You have a problem. Let me know when you are able to read the label
> and supply the vendor and model numbers of your inadequate equipment.
>

I will. But the reason why I didn't say, was because suppose at
another point I have to fix a different model for a friend, then the
knowledge here isn't much use, isn't transferrable.

if the answer here only applies to one model. Or applies to some
models and not others, then I wouldn't know why.

>From what you've said, you are trying to ascertain what type of MIMO
router it is, of the types you mentioned.

But I don't know how to ascertain it, so I have to tell you the exact
model numbers. I hope that's the case anyway. I like to know how to
solve it myself. Not to just say here's all the information, you tell
me the answer. Otherwise faced with a similar situation I wouldn't
know. Also, this is all archived, so others can benefit if the
explanation isn't tailored to me.

Belkin wireless G+ MIMO router (it has 2 antennas)
Belkin F5D9230-4
http://catalog.belkin.com/IWCatProdu...duct_Id=203415

Belkin wireless G+ MIMO USB adaptor
Belkin UK F5D7050uk
http://catalog.belkin.com/IWCatProdu...duct_Id=179753






> >range is insufficient.
>
> So are your expectations. Could I trouble you to describe what manner
> of range you were expecting? Over what terrain? Through how many
> walls?
>

2 flights of stairs.. All indoors. I haven't really measured. It's a
friend's house.

> >What is a good value for
> >that signal to noise ratio?
>
> It depends on the connection speed. For a BER (bit error rate) of
> about 1 error in 10^5 bits, or about 10% PER (packet error rate):
> Speed SNR(dB)
> 11 6.99
> 5.5 5.98
> 2 1.59
> 1 -2.92
> 54 24.6
> 48 24.1
> 36 18.8
> 24 17.0
> 18 10.8
> 12 9.0
> 9 7.8
> 6 6.0
> See:
> <http://groups.google.com/group/alt.internet.wireless/msg/b806cb8cf93c...>
> for an explanation. What I managed to leave out is that these figures
> are for a fairly lousy error rate. One needs to add in the
> anticipated fade margin to get real numbers. Add about 20dB to each
> of the SNR numbers for what is considered "good values" for viewing
> with Netstumbler.
>

So -80 to -95 is pretty bad!!!!!!

I know somebody that is happy with his -40 !

My connection was working at -85, then went down to -95 and
disconnected.

> >I read that USB wireless has a rubbish internal "chip" antenna , and
> >that makes for small range. I have one and it doesn't go past 2
> >flights of stairs.
>
> My stairs do not fly. You're correct. USB dongles have very small
> antennas. Antenna gain is very roughly proportional to the size of
> the antenna. The tiny chip or PIFA antennas have a gain of about
> -6dBi to optimistically 0dBi gain. A typical non-MIMO PCMCIA card
> will have an antenna gain of about 2dBi, but it's rather directional.
>
> Please note that your flight of stairs implies that you have a change
> of vertical elevation and that you might be going through a floor or
> ceiling. The pattern from the USB dongle is a close approximation of
> a hemisphere, so orientation isn't terribly important. However,
> that's not the case with the typical access point vertical coaxial
> antenna, especially if it's an add-on higher gain antenna.

> These
> trade gain in the vertical and horizontal direction (i.e through the
> floor and ceiling) in trade for more gain in the horizontal direction.
>

sounds like the antenna is some kind of ecomonist.

I was told that they work like this

> | <
> | <
> | <

So I see how your description might fit that technical fact. But the
person that told me that technical fact said I could orient it such
that the side of it instead of the tip, faces the wireless adaptor.

So why does your tradeoff statement matter?

Anyhow, for some reason, reorienting both antenna so instead of the
tip facing the wireless usb adaptor, the side did, it didn't help.

> >I'd like a wireless 802.11g MIMO PCI Card, that comes with a
> >reasonable antenna, but i'd also like it to have a connector for an
> >antenna. (so I only have to buy an external antenna if it's not good
> >without one).
>
> Forget it for now. MIMO functionality relys heavily on the matching
> between 3 or more antennas. As a result, an external antenna would
> need to have 3 antennas, 3 coax cables, and 3 connectors. There are
> some with external antennas, such as Linksys WMP300N but they're all
> PCI cards, where there's no other alternative to locating an antenna.
> All the other clients and access points have non-removable antennas or
> internal antennas.
>

So is a PCI card with one antenna is as good as without an antenna? Is
a PCI card as bad as a USB adaptor?

I heard that a MIMO router, MIMO is optional. The client may not
connect using MIMO.. e.g. or i.e. If it's not a MIMO adaptor.

What if the adaptor isn't MIMO? It could have one really good antenna
attached. Couldn't it?




> >Changing the router isn't an option at the moment.
>
> Change what router into what?

It's an expression. It means returning it, selling it, throwing it
out, discarding it, not using it. And replacing it with a different
one.

It's an option. And sometimes it's the best option.

For example. Suppose you have a MBRD that breaks, you change it for
another one.


> If you're going to practice
> enchantments and transmogrification, perhaps a practicing wizard or
> alchemist would be more helpful.
>

In computers, when things don't work, it helps to change those things
for something else that works or works better.

Or at least recognise that as an option.

> >(I did try pointing the aerials, including so the length of the aerial
> >as oppose to the tip, faces the wireless usb adaptor, but no use. I
> >didn't try moving the wireless usb adaptor around on a usb cable. I
> >don't plan on getting a wireless usb adaptor with an antenna, but I
> >might)
>
> Indecision is the key to flexibility.
>

I want to know all my options.
And if that's too general a question for you, then the most important
thing, is in particular, how far I can go with the PCI adaptor.

Are you implying that MIMO needs 3 aerials , else it's as good as no
aerials?. And that the same goes for the adaptor that connects to
it?

a wireless PCI adaptor with an external aerial is a waste of time.
'cos it's MIMO.

is a MIMO PCI adaptor without 3 aerials as bad as a Wireless USB
adaptor?

What if that adaptor isn't MIMO? Can't the router choose not to use
MIMO. And so your implication doesn't apply. One aerial , is
succeeded by 2. Or I could have one powerful aerial.



> >I'd like to know the power of these things before buying.. I've heard
> >that Buffalo make powerful wireless stuff. But they don't seem to have
> >any PCI, and I can't see the power rating.
>
> The transmit power is usually specified on the manufactories data
> sheets. If that's not available, it can be found on the FCC ID search
> page under the type certification test data. Buffalo makes various
> models. The WHR-HP-G54 transmits up to 250mw. 50mw is the default
> value.
>

what should I be be looking at to get an idea of strength of signal,
and range.

dbi? or mw?

what numbers?

> >(the buffalo website is a silly design anyway. each irrelevant
> >picture links to a directory, the directory's name reveals what it's
> >about. - so if you ever want to browse it, that's the way!)
>
> Right. The product quality must be directly related to the quality of
> the web site.
>

I never said it was. The product quality may be good, so all the more
reason why it helps to have a tip on browsing the poorly designed
site.

> >I know linksys show power rating, but a few have recommended buffalo.
>
> Both manufacturers have their benefits and detriments. I don't want
> to go into a point by point comparison. See:
> <http://www.smallnetbuilder.com/component/option,com_chart/Itemid,189/>
> for reviews and comparison charts.
>
> Incidentally, there may be some confusion here when using the latest
> MIMO buzzword.

ok

> In the context of this discussion,

ok!!

>there are 3 types
> of MIMO.
> 1. Airgo (Qualcomm) based chipset which in my opinion is "true" MIMO.
> <http://www.cdmatech.com/products/wlan.jsp>
> 2. Ruckus Wireless based chipsets, which are basically beam steering,
> beam forming, or antenna aiming systems.
> <http://www.ruckuswireless.com/technology/beamflex.php>
> 3. Gross misuse of the MIMO term by Buffalo on the WHR-HP-G54
> claiming "MIMO like performance" with having any features that even
> remotely involve MIMO technology.
> <http://www.buffalotech.com/technology/our-technology/high-power/>
>

thanks for the list.

I guess i'd have to google and read more of something to ascertain
which type a MIMO router is. But i've mentioned my model.

"(E-Mail Removed)" <(E-Mail Removed)> hath wroth:

>I will. But the reason why I didn't say, was because suppose at
>another point I have to fix a different model for a friend, then the
>knowledge here isn't much use, isn't transferrable.

Most range and performance problems can be generalized. There's a bit
on the topic in the FAQ at:
<http://wireless.wikia.com/wiki/Wi-Fi#Performance_and_Speed>
The usual problem is unrealistic expectations based on overly
optimistic manufacturers promises.

However, if you're having problems with a specific model router,
specific model client, in a specific location, with specific
expectations, methinks this is not the appropriate place for a general
course in RF propagation. In addition, there are at least radically 3
different types of MIMO and one vendor with MIMO hype, to add to the
confusion. Each has its own propagation and preformance issues.

>if the answer here only applies to one model. Or applies to some
>models and not others, then I wouldn't know why.

I would be elated if you would or could understand what I posted (and
corrected twice). Some of it doesn't apply to different models, but
you have to know what's inside and how it works. I really don't want
to write an industry wide technology survey.

>>From what you've said, you are trying to ascertain what type of MIMO
>router it is, of the types you mentioned.
>
>But I don't know how to ascertain it, so I have to tell you the exact
>model numbers.

How to ascertain the type of MIMO you have:
1. Find the FCC ID number on the serial number tag.
2. Go to the FCC ID web pile and inscribe the FCC ID into the search
page grantee code and product code. If it doesn't work, add a "-" in
front of the product code.
3. Go to the internal photos and extract the part number of the
chipset.
4. Go to the chipset manufacturers web pile and lookup the
specifications and application notes on the chipset.
5. Also check the wireless product reviews at:
<http://smallnetbuilder.com/content/view/85/106/>

>I hope that's the case anyway. I like to know how to
>solve it myself. Not to just say here's all the information, you tell
>me the answer. Otherwise faced with a similar situation I wouldn't
>know. Also, this is all archived, so others can benefit if the
>explanation isn't tailored to me.

Perhaps you didn't notice, but my initial reply was NOT anything
resembling an answer. That was due to your unwillingness to supply
sufficient information to calculate an answer. What I supplied was
the generalized theory of how such things work.

>Belkin wireless G+ MIMO router (it has 2 antennas)
>Belkin F5D9230-4
>http://catalog.belkin.com/IWCatProdu...duct_Id=203415

Thanks.
<http://smallnetbuilder.com/content/view/24214/96/>
Looks like Airgo "TrueMIMO" Pre-N chipset.
<http://smallnetbuilder.com/content/view/24214/96/1/6/>
<http://www.cdmatech.com/products/wlan.jsp>

>Belkin wireless G+ MIMO USB adaptor
>Belkin UK F5D7050uk
>http://catalog.belkin.com/IWCatProdu...duct_Id=179753

Ummmm.... where does it say MIMO on the USB data sheet? It's not a
MIMO USB adapter because it only has one antenna. It will work with a
MIMO router, but there's nothing MIMO about the basic product.

Are you sure it's not a F5D7050? The US version is MIMO:
<http://catalog.belkin.com/IWCatProductPage.process?Product_Id=272959>

No clue who's chipset is inside. Check on FCC ID web site or post the
FCC ID and I'll do the digging.

>> So are your expectations. Could I trouble you to describe what manner
>> of range you were expecting? Over what terrain? Through how many
>> walls?

>2 flights of stairs.. All indoors. I haven't really measured. It's a
>friend's house.

If you can't measure, try guessing. Also, you didn't answer the
question. What range were you expecting and going through what manner
of interior construction material? See list of losses in the FAQ at:
<http://wireless.wikia.com/wiki/Wi-Fi#Attenuation>
If you are trying to penetrate some of the high attenuation materials,
you're not going to go very far, with or without MIMO.

>> >What is a good value for
>> >that signal to noise ratio?

>So -80 to -95 is pretty bad!!!!!!

Not exactly. That's the noise level, not the SNR (signal to noise
ratio). Netstumbler gives two numbers. Signal level and Noise level.
If you subtract the two, you'll get SNR. Unfortunatly, it's somewhat
subjective.

For example, see:
<http://home.carolina.rr.com/harothberg/Test2.htm>
and note the Netstumbler screen capture. The left part is about
-85dbm signal, and -95dBm noise for a SNR= 10dB. The right part is
about -67dBm signal, and -98dBm noise for a SNR= 31dB. However, the
Orinoco client manager says SNR= 16dB, which is nowhere near either
result. Oh well.

>I know somebody that is happy with his -40 !
>
>My connection was working at -85, then went down to -95 and
>disconnected.

Which went down? The signal or the noise? Using Netstumbler or
something else?

>> These
>> trade gain in the vertical and horizontal direction (i.e through the
>> floor and ceiling) in trade for more gain in the horizontal direction.

>sounds like the antenna is some kind of ecomonist.

True. There's no such thing as a free lunch or free antenna gain. A
higher gain antenna trades beam width for gain. It also trades VSWR
bandwidth, but that's a different problem.

>I was told that they work like this
>
> > | <
> > | <
> > | <

Ummm... Whazzat?

>So I see how your description might fit that technical fact. But the
>person that told me that technical fact said I could orient it such
>that the side of it instead of the tip, faces the wireless adaptor.

The side of what antenna?

>So why does your tradeoff statement matter?

Easy. Your Belkin access point has the bulk of its gain and therefore
it's range in the horizontal plane. If you try to connect with
something that is above (or below) the horizontal plane, you're going
to see less gain.

As the antennas cannot be replaces on the Belkin access point, there's
no point in going into the effects of increasing their gain in order
to possibly improve performance. However, if they could be replaced,
then you would get more gain and range in the horizontal plane, and
less gain and range vertically (i.e. up your stairs).

>Anyhow, for some reason, reorienting both antenna so instead of the
>tip facing the wireless usb adaptor, the side did, it didn't help.

>So is a PCI card with one antenna is as good as without an antenna? Is
>a PCI card as bad as a USB adaptor?

As good as compared to what? MIMO generally has better range because
it tolerates reflections better than ordinary 802.11g. Depending on
what speed you're expecting, range will be better with ordinary
802.11g speeds. The PCI cards generally have much better antennas
than USB and will therefore be correspondingly better. In general,
the bigger the antenna, the better it works.

>I heard that a MIMO router, MIMO is optional.

Generally true. You can turn it off in the access point and you'll
end up with the equivalent of an ordinary 802.11g access point. You
might also have to turn off some of the other proprietary enhancements
such as turbo-G, afterburner, burst mode, 125Mbits/sec, etc.

>The client may not
>connect using MIMO.. e.g. or i.e. If it's not a MIMO adaptor.

Correct and your Belkin F5D7050uk is not a MIMO client. However, the
F5D7050 is a MIMO client adapter:
<http://catalog.belkin.com/IWCatProductPage.process?Product_Id=272959>

>What if the adaptor isn't MIMO? It could have one really good antenna
>attached. Couldn't it?

Yes, it could but it's unlikely that anything crammed into a USB
dongle enclosure has a decent gain antenna. Even with MIMO, a crappy
antenna works badly.

>It's an option. And sometimes it's the best option.

Changing your expectations is another possibility. I don't have
enough info to recommend anything specific. However, I can tell you
that when the signal gets weak, *ALL* (and I do mean *ALL*) the
wireless access points reduce speed in an effort to improve
sensitivity. Once that happens, all the MIMO, turbo-G, super-G,
whatever enhancements are effectively disabled. Range and speed can
be traded for each other.

>In computers, when things don't work, it helps to change those things
>for something else that works or works better.

Sure, but that's an expensive way of troubleshooting. I'm partial to
defining requirements, understanding the technology, and selecting
something reasonable based on these. There's also a problem with
"works better". That's fine, but the not necessarily universal for
all environments and applications. For example, different
technologies do better for reflective environments, long range, video,
large numbers of users, coffee shop environment, outdoors, indoors,
and such. My guess is they're all about 85% identical, but the last
15% is what provides the product differentiation.

>Or at least recognise that as an option.

Replacement is always an option. I'm just suggesting that you not do
it randomly or in an uninformed manner.

>I want to know all my options.
>And if that's too general a question for you, then the most important
>thing, is in particular, how far I can go with the PCI adaptor.

That's actually easy to calculate. The problem is that your
unspecified indoor arrangement is a key part of the puzzle. I also
need to know what speed you're expecting (speed and range can be
traded). I've done the calculations for outdoor links many times in
this newsgroups. For example:
<http://groups.google.com/group/alt.internet.wireless/msg/5a27b03fc5ef03e1>
Aim for a 20dB fade margin. Specs can be scraped from the
manufactureres data sheets. If you have problems, post your numbers
and I'll try to help.

>Are you implying that MIMO needs 3 aerials , else it's as good as no
>aerials?. And that the same goes for the adaptor that connects to
>it?

Ruckus MIMO is actually a steerable single antenna. Usually, it's a
big circuit board antenna inside the wireless router. No external
antenna needed or possible.

True MIMO must meet the 802.11n preliminary specifications, which
requires 3 antennas.

Airgo has an oddity which isn't really MIMO because it doesn't meet
the 802.11n specification, but which combines two streams. It acts
like MIMO, but uses only two antennas. That's your Belkin:
<http://80211n.wifinetnews.com/archives/2005/06/true_g_drops_pr.html>

>a wireless PCI adaptor with an external aerial is a waste of time.
>'cos it's MIMO.

I don't think you'll want a PCI card with an internal antenna. The RF
won't make it outside the metal box. With PCI, it has to be external.

>is a MIMO PCI adaptor without 3 aerials as bad as a Wireless USB
>adaptor?

A MIMO PCI adapter, with no antenna, is not going to work. I don't
see what you're asking.

>What if that adaptor isn't MIMO? Can't the router choose not to use
>MIMO. And so your implication doesn't apply. One aerial , is
>succeeded by 2. Or I could have one powerful aerial.

A non-MIMO client can connect to a MIMO router without difficulties.
The MIMO routers are all downwards compatible.

Please understand that two antennas are not twice as good as one, or
three antennas are not three times as good as one. The only thing the
multiple antennas do for the system is deal with reflections and their
position, allow for combining data streams to get more speed.

>what should I be be looking at to get an idea of strength of signal,
>and range.

Good question. I think a benchmark test is in order. Download IPerf
from:
<http://dast.nlanr.net/Projects/Iperf/>
Plug a fast PC into the Belkin wireless MIMO router using CAT5
ethernet cable. Run iperf as a server in a DOS window as:
iperf -s
On a laptop, connect to the Belkin router via wired or wireless. Run:
iperf -c ip_address_of_server
You should get some TCP thruput performance figures. Play with the
options and arguements.

Now, move away with the laptop and check your thruput at different
places in the house. Never mind signal strength, SNR, and such for
now as these don't mean much when not moving data. Just range and
performance. Walk around and you'll get a fair idea of how well your
system covers the house.

If you have time and patience, try a line of sight test. Measure out
a range and measure the thruput. It should be fairly constant and
slowly dribble down to zero as you get to the extreme range.

Now, do the same thing with one big change. Change the wireless speed
in the access point from "auto" to 54Mbits/sec. Do the walk test
again. This time, it will probably be quite constant until you get to
some threshold, where it will drop very quickly to zero. That's your
maximum range at maximum speed point.

You'll find plenty of such graphs under various reviews at
SmallNetBuilder.com.

>dbi? or mw?

dBi is antenna gain (over a theoretical isotropic radiator).
mw is power output in milliwatts.
Apples and oranges.
Range is measured in meters.
Performance in Mbits/sec.
Signal strength and Noise level are in dBm or dB over 1 milliwatt.
SNR (signal to noise ratio) is in dB which is a ratio.



--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

(E-Mail Removed) hath wroth:

>Which way are you suggesting that the hemisphere is oriented?

Ummm... hemispheres don't have points. However, I get your point. My
guess(tm) is that the maximum gain is in the direction opposite
anything that might act as a reflector. Therefore, for a PCB mounted
chip antenna, located near the narrow edge of the PCB, the maximum
gain would be upwards from the chip antenna, and slightly forward. One
thing for sure is that it's not symmetrical.

>I always
>thought that you were picturing a hemisphere over the top of a dongle, with
>the cable exiting the bottom. That would mean the cable would be the
>nonexistent hemi of the sphere.

When the chip antenna is near the edge of the PCB, there's quite a bit
of RF radiated in the downward direction from the bottom of the PCB.

>> Here are some realistic chip antenna patterns:
>> <http://www.rainsun.com/Products/LTCC/LTCC.html>
>> <http://www.antennafactor.com/documents/ANT-xxx-CHP-x_Data_Guide.pdf>
>> Yech. Ok, I give up. Something kinda sorta roughly vaguely
>> resembling a hemisphere.
>
>Those are the kind of lumps I see from signal strength measurements, except
>that I didn't see an equal back lobe.

It think the plots were done with a test board, which might not have
construction similar to a USB dongle. My guess is that the USB dongle
is far worse than the test board in terms of pattern symmetry.

>Putting that chip antenna on one
>side of a coper plated board might have a different signal pattern that the
>theoretical chip in space.

It's certainly not done in free space. Usually, the data sheet
mumbles something about a ground plane under the chip antenna for
strip line antennas. For meandering 1/4 wave lines, no ground plane
is desired. PIFA requires a ground connection, but no ground plane.

>The signal seemed to be moderately directional
>off the face of the dongle. Out the front of the can in my photo, if the
>can weren't there.
>
>http://www.rahul.net/dold/clarence/u...42-800x600.jpg
--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558