Prev: [How to] wireless network: modem/router + wifi router
Next: WAP54G remote log
From: miso on 21 Jan 2007 01:35

bi241(a)scn.org wrote:
> i played with the linksys WRT54G for a while and somehow got stuck in
> DD-WRT concept of diversity antenna. and also i am skeptical about the
> MIMO title that the Ralink chipset is advertised, after all, the
> AWLL5025 is still within 802.11G specs, while MIMO seems to belong to
> draft-N or whatever...
>
> but yeah in true MIMO design, diversity mean multiple RX antennas, (and
> mulltiple TX antennas too)
> i often see draft-N wireless adpaters with three antennas, just like
> the 802.11G AWLH5025 PCI adapter, but i was too lazy to learn about the
> antenna configurations. let me give it a try. these are possible
> configuration in a three-antenna setup
>
> a) three TX/RX combos
> b) one dedicated RX and two TX/RX combos
> c) one dedicated TX and two TX/RX combos
> d) two dedicated RX and one TX/RX combos
> e) two dedicated TX and one TX/RX combos
>
> i am still clueless about how many radios needed for one "draft-N"
> adaptor. but i do hope that there's no final N. 802.11 B/G is good
> enough for me, and for all, that's all we need for wireless
> internet.... as for networking, i'd prefer and recommend the old
> fashion way...
>
> anyways, in the case of the AWLL5025 USB, the external antenna must be
> the combo TX/RX. Thank you for pointing that out. it makes alot of
> sense... and i think i am gonna hack the antenna after this... yes i am
> serious!
>
>
> cheers

Diversity as receive only has been around for some time. Military of
course, but some high end automobiles had diversity receivers for their
FM broadcast reception.

I can understand diversity receive. The receiver comes up with some
sort of figure of merit on the reception. It is multiple transmit
(output) that confuses me. I don't see how a transmitter decides which
antenna to use. In any event, if everyone had diversity in the receive
path, I don't see why it would be needed in the transmitter.

http://www.smc.com/index.cfm?event=viewProduct&localeCode=EN_USA&cid=9&scid=&pid=346
SMC2532W-B
is the high power card I use in my notebook. It has diversity receive,
but one transmitter. They claim half a mile on the detachable antenna.
Hardly. However, I've done drive-bys of at least a mile with the
biquad. Make that the bad biquad, i.e. the two wire instead of the
coaxial feed. I haven't tried the good biquad over any distance yet.

This is what Jeff considers the good biquad:
http://martybugs.net/wireless/biquad/

I used something like this for the feed:
http://www.hyperlinktech.com/web/connectors/connector_information.php?SKU=AXA-NFNF

Typically they are smooth, not knurled. [Should be about a dollar at
the ham flea market.] Hacksaw one end off, leaving the center pin
exposed. Then position the adapter so you get the right distance
between the antenna and reflector. I cut the copper plate, drilled the
hole, then tinned the area around the hole with solder. This made it
easier to solder the N adapter to the copper plate. You need a propane
torch. [Benzomatic]




<snip>

From: Jeff Liebermann on 21 Jan 2007 13:04
bi241(a)scn.org hath wroth:

>i played with the linksys WRT54G for a while and somehow got stuck in
>DD-WRT concept of diversity antenna. and also i am skeptical about the
>MIMO title that the Ralink chipset is advertised, after all, the
>AWLL5025 is still within 802.11G specs, while MIMO seems to belong to
>draft-N or whatever...

Correct. The marketting baloney that accompanies the Buffalo
WHR-HP-54G is disgusting. They use the term MIMO because it's the
latest and greatest, while there's nothing even remotely related to
MIMO inside the box. What Buffalo is pushing is that you'll get the
same performance as MIMO with their high power transmitter, and offer
some un-reproduceable tests to prove it. It makes me sick to see a
good product polluted by such technically inept marketing.
<http://www.buffalotech.com/technology/our-technology/high-power/>


>but yeah in true MIMO design, diversity mean multiple RX antennas, (and
>mulltiple TX antennas too)
>i often see draft-N wireless adpaters with three antennas, just like
>the 802.11G AWLH5025 PCI adapter, but i was too lazy to learn about the
>antenna configurations.

MIMO is not diversity. What "real" MIMO does is send data from 2 or
more antennas. The data arrives at corresponding antennas and
receivers, but delayed very slightly. The DSP is able to seperate out
the various delays and extract seperate data streams for each delay.
The result is a doubling in available bandwidth. See:
<http://www.airgonetworks.com/mimo/how/>

There's also a skool of thought that claims that MIMO is also adaptive
antenna patterns. This is where the access point contains a smart
antenna system that adds more gain where it's needed, and installs
nulls where it detects interference. See Ruckus Beamflex at:
<http://www.ruckuswireless.com/technology/beamflex.php>
I have a good laugh on this one. The system works, but Netgear
advertises their various routers mounted in the vertical (tower)
position, where the beam forming can't work.

If you're interested in reading about non-MIMO diversity reception
see:
<http://www.commsdesign.com/design_corner/showArticle.jhtml?articleID=16500279>
<http://www.commsdesign.com/design_corner/showArticle.jhtml?articleID=16501888>
<http://www.cisco.com/en/US/tech/tk722/tk809/technologies_tech_note09186a008019f646.shtml>

--
Jeff Liebermann jeffl(a)comix.santa-cruz.ca.us
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
From: miso on 22 Jan 2007 02:53

dold(a)15.usenet.us.com wrote:
> miso(a)sushi.com wrote:
>
> > Actually, what is returned at Fry's is handy to know. If there are lots
> > of returns, try something else. I know those on this usenet group are
> > Linksys fans, but most of what is returned to Fry's is Linksys.
>
> I would say that's hard to judge. In the WiFi sections (more than one,
> since there are some dedicated sections, including Linksys), there are
> many empty shelf spaces, usually for the item you want, plenty of items in
> the wrong shelf space, and some opened/returned/markdown units, with
> markdowns of 12-15 cents, which fascinates me.
>
> I would guess that they don't give Linksys their own dedicated space unless
> they move lots of units, so the number of returns on the shelf would have
> to be noted against the number sold.
>
> Wifi is still an emerging market. Many of the units are purchased with
> unreasonable hopes, like using Wifi while commuting on Caltrain... okay, so
> that works now; or getting free access anywhere, or full coverage in a
> home. Those units probably come back, maybe for refund, maybe to exchange
> for a different brand.
>
> The motherboard section is full of returns, often with no new units on the
> shelf. For stores other than Sunnyvale, I don't know what that means, but
> in Sunnyvale, I have overheard engineers selecting motherboards to take
> back and test/certify against some product or another, after which it will
> be returned to Fry's. But these are the same guys trying to make sure they
> get one that is unopened.
>
> > If I think there are going to be problems with something, I always go
> > to Fry's because they will take back anything.
>
> The home of the thirty day free trial for almost anything.
> I have seen people return stereos with no packaging.
>
> --
> ---
> Clarence A Dold - Hidden Valley Lake, CA, USA GPS: 38.8,-122.5

I just purchased a refurb wifi router from Fry's . Netgear WNR854T for
$60. I wanted a 1gbps router with wan port, but figured might as well
get N wifi at the same time. These are factory refurbs in plain brown
boxes. There was an assortment of netgear products, maybe three of each
model.

Fry's is a poor place to get a mobo. They tend to sell older models of
the better brands, or mystery mobos. For a business, I can't see going
through the work to set up a mobo, only to test then return. More
likely would be to use a junky case and just remove the ram, CPU, and
drives, leaving the mobo in the case in the event you need it again.
Fry's records who returns what. I can't see someone making a habit of
doing returns not getting the evil eye, or at least being force to take
store credit rather than an exchange. I return something maybe every
two or three years.

From: Jeff Liebermann on 22 Jan 2007 03:36
Jeff Liebermann <jeffl(a)comix.santa-cruz.ca.us> hath wroth:

>Nope. Quiz time:
>
>You have a spectrum analyzer that shows an 802.11g spread spectrum
>signal that is approximately 25Mhz wide. Peak power on the scale is
>0dBm to make the calculations easy. The spectrum analyzer IF
>bandwidth is 500KHz. What is the "real" power (ignoring duty cycle)?
>
>I'll supply the answer tomorrow. Note that this is not a trick
>question and the numbers are quite real.

Not even a guess? Oh well. This is a good example of why you can't
directly use a spectrum analyzer in place of a power meter or
bolometer with wideband (SS, video, UWB, etc) signals.

The indicated power level on the display is 0dBm. However, the power
is splattered over a bandwidth of 25Mhz. We want the equivalent
carrier power, or the power of the de-spread signal. So we add:
10 * log(25MHz/0.5MHz) = 10 * log(50) = 17dB
So, the real power is:
0dBm + 17dB = +17dBm

Even this is not exact because the power spectra is not equal power
across the 25Mhz bandwidth and the IF filter is not a brick wall at
500KHz. There's also quite a bit of power outside of the 25MHz
window. The rms (heating) power is also different because of the
transmit duty cycle.

--
Jeff Liebermann jeffl(a)comix.santa-cruz.ca.us
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
From: miso on 22 Jan 2007 15:14

Jeff Liebermann wrote:
> Jeff Liebermann <jeffl(a)comix.santa-cruz.ca.us> hath wroth:
>
> >Nope. Quiz time:
> >
> >You have a spectrum analyzer that shows an 802.11g spread spectrum
> >signal that is approximately 25Mhz wide. Peak power on the scale is
> >0dBm to make the calculations easy. The spectrum analyzer IF
> >bandwidth is 500KHz. What is the "real" power (ignoring duty cycle)?
> >
> >I'll supply the answer tomorrow. Note that this is not a trick
> >question and the numbers are quite real.
>
> Not even a guess? Oh well. This is a good example of why you can't
> directly use a spectrum analyzer in place of a power meter or
> bolometer with wideband (SS, video, UWB, etc) signals.
>
> The indicated power level on the display is 0dBm. However, the power
> is splattered over a bandwidth of 25Mhz. We want the equivalent
> carrier power, or the power of the de-spread signal. So we add:
> 10 * log(25MHz/0.5MHz) = 10 * log(50) = 17dB
> So, the real power is:
> 0dBm + 17dB = +17dBm
>
> Even this is not exact because the power spectra is not equal power
> across the 25Mhz bandwidth and the IF filter is not a brick wall at
> 500KHz. There's also quite a bit of power outside of the 25MHz
> window. The rms (heating) power is also different because of the
> transmit duty cycle.
>
> --
> Jeff Liebermann jeffl(a)comix.santa-cruz.ca.us
> 150 Felker St #D http://www.LearnByDestroying.com
> Santa Cruz CA 95060 http://802.11junk.com
> Skype: JeffLiebermann AE6KS 831-336-2558

This depends on the "intelligence" of your spectrum analyzer. For the
audio spectrum analyzer I have, you just push a button and it computes
the energy in the bandwidth since it has all the information available
to do such a computation. I don't know if there are RF spectrum
analyzers with such intelligence, but I don't see why not.

In datacom (again, back to voice band communications), some modems used
guard tones to indicate the line contained data, not voice. The CCITT
specification is based on energy, so making such measurements (energy
in a bandwidth) is a desirable feature. That is, the tone had to be a
certain value relative to the data energy, which was spread out.

First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4
Prev: [How to] wireless network: modem/router + wifi router
Next: WAP54G remote log