Multiple access points using same antena

Ok using cisco br350 access points and cisco antenas.
Does anyone know if it would be possible to share one aerial between 2
access points ?
Would I need to create a magic box of tricks to seperate the frequencies ?
or just splice them in ?

On Tue, 20 Jul 2004 11:56:14 +0100, "Mr_Flibble"
<(E-Mail Removed)> wrote:

>Ok using cisco br350 access points and cisco antenas.
>Does anyone know if it would be possible to share one aerial between 2
>access points ?

Yes, but you probably don't want to do it.

You need what is called a -3dB splitter, power divider, or Wilkinson
combiner. It's a passive device with little (-0.5dB/port) loss.
They're more commonly used to attach two antennas to one access point
rather than two access points to one antenna.
http://www.hyperlinktech.com/web/sc2402n.php
http://www.qsl.net/n9zia/wireless/pics/powersp.gif
http://www.hdcom.com/2.4ghzdivcomb.html

>Would I need to create a magic box of tricks to seperate the frequencies ?
>or just splice them in ?

One does not "splice" frequencies. One either mixes or adds. If
mixing, you get the multiples of the sums and differences between
frequencies. The FCC and the UK Post get rather irritated when this
happens as it always splatters into services on other frequencies
creating interference. Mixing is a bad idea. Summing is what you're
trying to do. In order for that to work, everything in the system
must be linear so as not to create a mixer. Well, that isn't going to
happen. Even the slightest amount of corrosion is going to create a
mixer. The combiner will give you some isolate between xmit ports,
but you still run the risk of mixing in the output stages of the
radio. The radios with PIN diode diversity switches are particularly
bad as the diodes tend to rectify and cause a mix.

There's also a big problem at the receiver. The power
divider/combiner splits the receive signal equally between the two
radios. Therefore, you lose -3dB plus the combiner loss (-0.5dB) of
receive sensitivity.

In addition, most receivers cannot handle the overload from the
adjacent xmitter. The direct sequence spread spectrum signal is about
22MHz wide. However, the power spectra continues well beyond that.
The splitter/combiner will offer some isolation (-15dB?), but xmitter
will still deliver some crud on the receive signal. If the receiver
isn't desensitized by simple overload, then the increase in the noise
level will do the same thing.

To make it work, you would need two 22Mhz band separation filters to
give better isolation. I would use a diplexer (hi/lo pass filter)
instead of a splitter/combiner.
http://www.rflinx.com/Filters/2400_UHQ_Filter.htm
The basic idea is to get as much seperation as possible between radios
and radio frequencies. That could easily cost more than a 2nd antenna
and coax.


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

"Jeff Liebermann" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> On Tue, 20 Jul 2004 11:56:14 +0100, "Mr_Flibble"
> <(E-Mail Removed)> wrote:
>
> >Ok using cisco br350 access points and cisco antenas.
> >Does anyone know if it would be possible to share one aerial between 2
> >access points ?
>
> Yes, but you probably don't want to do it.
>
> You need what is called a -3dB splitter, power divider, or Wilkinson
> combiner. It's a passive device with little (-0.5dB/port) loss.
> They're more commonly used to attach two antennas to one access point
> rather than two access points to one antenna.
> http://www.hyperlinktech.com/web/sc2402n.php
> http://www.qsl.net/n9zia/wireless/pics/powersp.gif
> http://www.hdcom.com/2.4ghzdivcomb.html
>
> >Would I need to create a magic box of tricks to seperate the frequencies
?
> >or just splice them in ?
>
> One does not "splice" frequencies. One either mixes or adds. If
> mixing, you get the multiples of the sums and differences between
> frequencies. The FCC and the UK Post get rather irritated when this
> happens as it always splatters into services on other frequencies
> creating interference. Mixing is a bad idea. Summing is what you're
> trying to do. In order for that to work, everything in the system
> must be linear so as not to create a mixer. Well, that isn't going to
> happen. Even the slightest amount of corrosion is going to create a
> mixer. The combiner will give you some isolate between xmit ports,
> but you still run the risk of mixing in the output stages of the
> radio. The radios with PIN diode diversity switches are particularly
> bad as the diodes tend to rectify and cause a mix.
>
> There's also a big problem at the receiver. The power
> divider/combiner splits the receive signal equally between the two
> radios. Therefore, you lose -3dB plus the combiner loss (-0.5dB) of
> receive sensitivity.
>
> In addition, most receivers cannot handle the overload from the
> adjacent xmitter. The direct sequence spread spectrum signal is about
> 22MHz wide. However, the power spectra continues well beyond that.
> The splitter/combiner will offer some isolation (-15dB?), but xmitter
> will still deliver some crud on the receive signal. If the receiver
> isn't desensitized by simple overload, then the increase in the noise
> level will do the same thing.
>
> To make it work, you would need two 22Mhz band separation filters to
> give better isolation. I would use a diplexer (hi/lo pass filter)
> instead of a splitter/combiner.
> http://www.rflinx.com/Filters/2400_UHQ_Filter.htm
> The basic idea is to get as much seperation as possible between radios
> and radio frequencies. That could easily cost more than a 2nd antenna
> and coax.
>
Thank you for the info, I was always wondering if it was possible however it
may actually cost less than the antena and coax.
(even though I would prefer the second antena approach) the place were this
equipment is charges us about £4000 a year per antena and we would need more
engineering plans drawn up.

On Thu, 22 Jul 2004 09:04:28 +0100, "Mr_Flibble"
<(E-Mail Removed)> wrote:

>Thank you for the info, I was always wondering if it was possible however it
>may actually cost less than the antena and coax.
>(even though I would prefer the second antena approach) the place were this
>equipment is charges us about £4000 a year per antena and we would need more
>engineering plans drawn up.

If you're serious about doing this, then I suggest a band pass filter
approach. You will not need a power splitter or combiner if done
correctly. A "T" connector, multiple of 1/2 wave (electrical) coax
runs, and a pair of 22MHz BW cavity band pass filters for channels 1
and 11. Each cavity filter would go to an access point. If you can
get about -60dB isolation between filters, it will probably work.
Make an effort to avoid corrosion and dissimilar metals as any
non-linearity will create intermodulation products and spurious
signals. Verify that you're not belching garbage with a spectrum
analyzer.

This sounds like a topic more approriate to WISP (Wireless ISP)
discussions. I suggest you join the ISP-Wireless mailing list at:
http://isp-wireless.com



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