Cheaply measuring wireless signal strength

[Mark Carroll]

802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz. I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing.

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something? (I've made plenty of digital circuits, but
with analog I know little more than how do simple things with RC
oscillators, op amps, etc.) Or, really, anything that responds to the
high-frequency signal in a manner that's easier to detect with cheaper
equipment?

Just looking for ideas. (-: Tiny GHz signals are not what I'm used to
dealing with! I'm not even sure how tuned to the application the
antenna needs to be so long as it's roughly the right length; it's
probably relevant that I'm interested much more in relative signal
strength than being able to accurately measure absolute power, so some
losses in the signal path may be quite acceptable if they're
consistent.

-- Mark

 

[Tony Morgan]

In message <0ii*5D31p@news.chiark.greenend.org.uk>, Mark Carroll
<markc@chiark.greenend.org.uk> writes

802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz.

13 I think.. Varies country to country.. 11 in the US I think.

I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing.

Indeed, and everyone wants one. If it were as easy and cheap as you
think there would be dozens on the market but there are (AFAIK) only two
and those aren't very well reviewed.

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

High frequency oscilloscopes are *very* expensive and AFAIK there aren't
any available at anywhere near the GHz range.

Spectrum analysers for the GHz range (at about Ł3,000) are the nearest
thing, and they mix a frequency to the signal and display the
difference.

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something? (I've made plenty of digital circuits, but
with analog I know little more than how do simple things with RC
oscillators, op amps, etc.) Or, really, anything that responds to the
high-frequency signal in a manner that's easier to detect with cheaper
equipment?

RF design at GHz frequencies is bordering on a black art. And, since
we're dealing here with low-level signals noise is a real issue.
Cross-coupling another (thing suddenly bursts into oscillation jamming
the very thing you're trying to detect/measure )

Just looking for ideas. (-: Tiny GHz signals are not what I'm used to
dealing with! I'm not even sure how tuned to the application the
antenna needs to be so long as it's roughly the right length;

That is tuning where RF is concerned. Aerials are (or should be)
resonant at the centre frequency. Look up "Q" in the text books. Q is
synonymous with gain. And the thicker the aerial the greater the
bandwidth but the lower the Q. Aerial tuning is based on a half-wave
dipole - at 2.45GHs that's about 6.25cm.

it's
probably relevant that I'm interested much more in relative signal
strength than being able to accurately measure absolute power, so some
losses in the signal path may be quite acceptable if they're
consistent.

The only thing that occurs to me is that all 802.11 devices have a
receiver in them. And receivers generally use AGC (automatic gain
control) using a derived voltage level corresponding to signal strength.
Since AGC voltage is effectively a varying DC level proportional to
signal strength, it might be possible to drive an inexpensive meter via
an op-amp (for isolation). PCMIA/Cardbus and USB (better [1]) wireless
devices for laptops are about Ł30 to Ł40 and (I imagine) it wouldn't be
too difficult to "break-out" the AGC line for your op-amp/meter giving
you a relatively accurate indication of signal strength..

Just a thought

[1] A USB device like the Linksys WUSB11 seems better
because it is easy to provide (battery) power to the device
via the power pins on the USB plug. The WUSB11also has
a real vertically polarised aerial which means you're
extracting more signal from the ether to play with.

Anyway, best of luck - let me know if you have any success.
--
Tony Morgan
Smile in the face of adversity - and adversity will probably
think you're taking the piss and kick the shit out of you.

 

[CWatters]

If you have an IPAQ, perhaps look at the YellowJacket series from...

http://www.bvsystems.com

eg

http://www.bvsystems.com/Products/WLAN/Yellowjacket%2B/yellowjacket%2B.htm

(slow site)

 

[Robert Lacoste]

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

Oscilloscopes don't (thy will display the sum of all signals around, and
will have a very low sensitivity compared to RF signals). Moreover I'm
afraid that a cheap 2.4GHz oscilloscope doesn't exist... The general-purpose
test equipment you really need exists and is a called a spectrum analyzer.
It displays the signal amplitude over frequency and not over time, and has a
very high sensitivity due to log amplifiers. Unfortunatly that isn't what
you I think you will call a cheap equipment... Old units going to 3GHz could
be found may be for $1000 (look for HP141T/HP8555, etc), new units starts at
$4-5K

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something?

It will be difficult. It will be nearly impossible to find a narrow band
pass filter in the 2.4GHz range that will be selective enough to get only
the Wifi bands, except if you knows an expert in cavity designs. The usual
scheme is to use a larger band pass filter, then a downconverter, then a
narrow band pass filter, then a level detector, etc, but that's not an easy
design.

Just looking for ideas.

I've seen some adds from commpanies using actual 802.11 access points and
just providing a specific software that let see the average power level on
each channel. That is probably the most cost effective solution for what you
are looking for, even if the precision will probably not be high...

Good luck,
--
Robert Lacoste - ALCIOM : The mixed signals experts
http://www.alciom.com

 

[mike]

Mark Carroll wrote:

802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz. I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing.

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something? (I've made plenty of digital circuits, but
with analog I know little more than how do simple things with RC
oscillators, op amps, etc.) Or, really, anything that responds to the
high-frequency signal in a manner that's easier to detect with cheaper
equipment?

Just looking for ideas. (-: Tiny GHz signals are not what I'm used to
dealing with! I'm not even sure how tuned to the application the
antenna needs to be so long as it's roughly the right length; it's
probably relevant that I'm interested much more in relative signal
strength than being able to accurately measure absolute power, so some
losses in the signal path may be quite acceptable if they're
consistent.

-- Mark

I must be missing something here.
If you put netstumbler on a laptop with a wireless card it will
give you a list of all the signals it finds and their strength,
s/n ratio etc. Won't that do what you want?
mike

--
Bunch of stuff For Sale and Wanted at the link below.
laptops and parts Test Equipment
4in/400Wout ham linear amp.
Honda CB-125S
400cc Dirt Bike 2003 miles $550
Police Scanner, Color LCD overhead projector
Tek 2465 $800, ham radio, 30pS pulser
Tektronix Concept Books, spot welding head...
http://www.geocities.com/SiliconValley/Monitor/4710/

 

[K Bloch]

O'scope will not work for what you want. A reasonable fast o'scope has
a bandwidth of about 100 mhz. Something for 2.4 ghz would be very
pricey.

As far as cheap a used spectrum analyzer that will do at least 2.5 ghz
will also be very expensive. Best bet is to buy a card that can
monitor in promiscous mode such as an Orinoco/Proxim Gold Card or
thier dual 802.11 A/G card. This way you can run an application like
netstumbler and determine what channels are in use, the relative
signal strength etc. Total cost is about $100 for the card plus a
laptop computer of your choice.

Also by using a laptop you will have real world results that a user is
likely to experience.


Mark Carroll <markc@chiark.greenend.org.uk> wrote in message news:<0ii*5D31p@news.chiark.greenend.org.uk>...

802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz. I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing.

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something? (I've made plenty of digital circuits, but
with analog I know little more than how do simple things with RC
oscillators, op amps, etc.) Or, really, anything that responds to the
high-frequency signal in a manner that's easier to detect with cheaper
equipment?

Just looking for ideas. (-: Tiny GHz signals are not what I'm used to
dealing with! I'm not even sure how tuned to the application the
antenna needs to be so long as it's roughly the right length; it's
probably relevant that I'm interested much more in relative signal
strength than being able to accurately measure absolute power, so some
losses in the signal path may be quite acceptable if they're
consistent.

-- Mark

 

[K Bloch]

O'scope will not work for what you want. A reasonable fast o'scope has
a bandwidth of about 100 mhz. Something for 2.4 ghz would be very
pricey.

As far as cheap a used spectrum analyzer that will do at least 2.5 ghz
will also be very expensive. Best bet is to buy a card that can
monitor in promiscous mode such as an Orinoco/Proxim Gold Card or
thier dual 802.11 A/G card. This way you can run an application like
netstumbler and determine what channels are in use, the relative
signal strength etc. Total cost is about $100 for the card plus a
laptop computer of your choice.

Also by using a laptop you will have real world results that a user is
likely to experience.


Mark Carroll <markc@chiark.greenend.org.uk> wrote in message news:<0ii*5D31p@news.chiark.greenend.org.uk>...

802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz. I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing.

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something? (I've made plenty of digital circuits, but
with analog I know little more than how do simple things with RC
oscillators, op amps, etc.) Or, really, anything that responds to the
high-frequency signal in a manner that's easier to detect with cheaper
equipment?

Just looking for ideas. (-: Tiny GHz signals are not what I'm used to
dealing with! I'm not even sure how tuned to the application the
antenna needs to be so long as it's roughly the right length; it's
probably relevant that I'm interested much more in relative signal
strength than being able to accurately measure absolute power, so some
losses in the signal path may be quite acceptable if they're
consistent.

-- Mark

 

[Tony Morgan]

In message <3f5df883$0$20948$7a628cd7@news.club-internet.fr>, Robert
Lacoste <rlacoste@alciom.com> writes
Snipped...

The usual scheme is to use a larger band pass filter, then a
downconverter, then a narrow band pass filter, then a level detector,
etc,

AFAIK the usual scheme is to use a balanced mixer with a LO of (say)
2GHz. Then a 400KHz bandpass with an endpoint on upper or lower sideband
to give you a signal for analysis/inspection/processing.

Obviously since you are a mixed signal expert - I must be wrong
--
Tony Morgan
Smile in the face of adversity - and adversity will probably
think you're taking the piss and kick the shit out of you.

 

[Dbowey]

Mark posted, in part:
<< 802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz. I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing. >>


How about $24.95 plus shipping? Kensington WiFi Finder detects 802.11 a and g
from up to 200'. Order from MACMALL at 800 622-6255 (their part # is 241768).
It's a stand-alone unit that has 3 LEDs to indicate signal strength. You could
probably modify it with a meter driver very easily.

Don

 

[Tony Morgan]

In message <Xns93F2835D8BFRusHcomputersystems@193.110.122.80>, RusH
<rush@kiti.pulse.pdi.net> writes

not all, it wont show you microwave oven for example

You using a microwave oven as an AP?
--
Tony Morgan
Smile in the face of adversity - and adversity will probably
think you're taking the piss and kick the shit out of you.

 

[RusH]

mike <spamme0@juno.com> wrote in news:3F5E208A.90605@juno.com:

I must be missing something here.

yes, you are

If you put netstumbler on a laptop with a wireless card it will
give you a list of all the signals it finds and their strength,
s/n ratio etc. Won't that do what you want?

not all, it wont show you microwave oven for example

--
Oh hell yes! Think of a computer as a tiny, super complex street
hooker. The more you put in.. Wait. No. That's not a good
analogy... Um... A computer is like a piece of paper. Er. No.
Um. I really have no idea how they work.

 

[Michael A. Terrell]

Tony Morgan wrote:

AFAIK the usual scheme is to use a balanced mixer with a LO of (say)
2GHz. Then a 400KHz bandpass with an endpoint on upper or lower sideband
to give you a signal for analysis/inspection/processing.

Obviously since you are a mixed signal expert - I must be wrong
--

2.4 GHz - 2.0 GHz leaves 400 MHZ, not 400 KHz.
--


Michael A. Terrell
Central Florida

 

[Tony Morgan]

In message <3F5E7BDE.D39971A0@earthlink.net>, Michael A. Terrell
<mike.terrell@earthlink.net> writes

2.4 GHz - 2.0 GHz leaves 400 MHZ, not 400 KHz.

Thank you soooo.... much.
--
Tony Morgan
Smile in the face of adversity - and adversity will probably
think you're taking the piss and kick the shit out of you.

 

[Robert Lacoste]

AFAIK the usual scheme is to use a balanced mixer with a LO of (say)
2GHz. Then a 400KHz bandpass with an endpoint on upper or lower sideband
to give you a signal for analysis/inspection/processing.

It's what I call a downconverter...

Friendly,
Robert

 

[IVAN KORSHUN]

If you put netstumbler on a laptop with a wireless card it will
give you a list of all the signals it finds and their strength,
s/n ratio etc. Won't that do what you want?
not all, it wont show you microwave oven for example

WLANexpert mode Air Scan:
http://www.allaboutjake.com/network/linksys/wlanexpert.html

Will show you microwave oven

 

[RusH]

2_ivan@mail.ru (IVAN KORSHUN) wrote in news:d84590a1.0309100251.48b327d0
@posting.google.com:

If you put netstumbler on a laptop with a wireless card it will
give you a list of all the signals it finds and their strength,
s/n ratio etc. Won't that do what you want?
not all, it wont show you microwave oven for example

WLANexpert mode Air Scan:
http://www.allaboutjake.com/network/linksys/wlanexpert.html

Will show you microwave oven

yez, this is the only free tool that will do that , but only on a Prism 2 card,
not on all cards like Mike said - that was my point.


--
Oh hell yes! Think of a computer as a tiny, super complex street
hooker. The more you put in.. Wait. No. That's not a good
analogy... Um... A computer is like a piece of paper. Er. No.
Um. I really have no idea how they work.

 

[Gary Tait]

On 09 Sep 2003 15:04:20 +0100 (BST), Mark Carroll
<markc@chiark.greenend.org.uk> wrote:

802.11b wireless networks run on 14 channels from 2.412GHz to
2.484GHz. I want to be able to test the signal strength in different
areas of a building cheaply, but the professional meters are rather
expensive. It would be nice to be able to test the channel strengths
individually, but a measure of signal strength over that whole range
of channels would be better than nothing.

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

(a) Can I just attach a little antenna to oscilloscope probes? Can I
get a fast enough second-hand oscilloscope fairly cheaply? 2.4GHz is
pretty fast! Or, do things just not work like that?

(b) Can I make simple little circuits, that have a little antenna,
that resonate fairly specifically with the 802.11b signal (roughly
over the right range, or even for individual channels) enough to build
up a steadier signal that I can measure more easily with a digital
voltmeter or something? (I've made plenty of digital circuits, but
with analog I know little more than how do simple things with RC
oscillators, op amps, etc.) Or, really, anything that responds to the
high-frequency signal in a manner that's easier to detect with cheaper
equipment?

Just looking for ideas. (-: Tiny GHz signals are not what I'm used to
dealing with! I'm not even sure how tuned to the application the
antenna needs to be so long as it's roughly the right length; it's
probably relevant that I'm interested much more in relative signal
strength than being able to accurately measure absolute power, so some
losses in the signal path may be quite acceptable if they're
consistent.

-- Mark

I don't know who makes them, but there is a simple 2.4 Ghz detctor
sold. I think Fellowes, or other computer accessory vendore has it.

 

[Slick Willy]

http://www.defactowireless.com/teletronics.htm
These folks make an 802.11 card and have a utility that makes it behave
like an ism band spectrum analyzer.


http://users.rcn.com/exergea/accessories/
At the bottom of this page you will find a PCMCIA card designed
specificly for the purpose. $250 new and plenty on eBay for $150 and
under. I bought one of these but havn't received it yet. Supposedly runs
in DOS only and I've seen a few "package" deals including a 486 based
laptop. Could be pretty sweet with an old tablet PC like a Fujitsu
Stylistic.


http://www.aerocomm.com/OEM/SA.htm
This is another product designed specificly for the ISM band. It's a bit
on the large size and fairly pricey if I remember correctly.


Google for "ISM Band Spectrum Analyzer" and you will find quite a few
more traditional spectrum analyzers for the job in the $2k-$5k price
range.


Kensington and a few others make small handheld devices for detecting
wifi networks. They filter out things like cordless phones, cameras, and
microwave ovens. This filtering can be a problem if your looking for a
source of intereferance. Also they have just a few LEDs to show signal
strength. Cost is something like $30 depending on where you buy.

I'm sure something clever can be done on the cheap but it doesn't appear
to be public yet.


In article <0ii*5D31p@news.chiark.greenend.org.uk>,
Mark Carroll <markc@chiark.greenend.org.uk> wrote:

I was wondering if something clever could be done rather cheaply. I'm
out of my depth here, but I was thinking something like:

-- Mark

 

[Al Dykes]

In article <Xns93F2835D8BFRusHcomputersystems@193.110.122.80>,
RusH <rush@kiti.pulse.pdi.net> wrote:

mike <spamme0@juno.com> wrote in news:3F5E208A.90605@juno.com:

I must be missing something here.

yes, you are

If you put netstumbler on a laptop with a wireless card it will
give you a list of all the signals it finds and their strength,
s/n ratio etc. Won't that do what you want?

not all, it wont show you microwave oven for example

--

I was under the impression that "noise" figure reported by netstumbler was
all non-802.11 energy on the channel.



--
Al Dykes
-----------
adykes@panix.com

 

[Steve Roberts]

see http://www.arrl.org/qex/Henkel.pdf

extends poor mans spectrum analyser to almost 5 gigahertz.
cheaply using cavities made from plumbing pipe and hobby shop brass,
plus some low cost microwave vcos.

Steve Roberts