I agree also that the second antenna is for diversity but none of the
manufactures can answer why it is needed if spread spectrum is suppose to be
so innately fade resistant, even though the spread spectrum Wifi is using
now with CCK isn't really truly random anymore.

Russ N7QR
russm at ieee dot org

"Arthur" wrote in message
news
On Thu, 7 Oct 2004 12:00:52 +0100, Daniel Kelly (AKA Jack)
wrote:

I'm not a radio engineer so I'm not too sure what's going on. If I was
forced to make an educated guess then I'd say the problem was caused by
interference caused by reflected, out-of-phase signals. For example,
it's
conceivable that the receiver could be picking up two 'copies' of the
transmission from the TX - one 'copy' is the direct signal, the other
'copy'
is a signal that has bounced off some surface - these two 'copies' would
cancel eachother out when they're 180 degrees out of phase.

Yout assumption is absolutely correct


So... this brings me to the question I posed in the subject line: why do
WiFi basestations use two antennas? For example, my 3Com basestation
has
two little whip antennas sticking out the back. Are these two antennas
used
to combat the problem that I am experiencing? If so, how can I
implement
two antennas on my system?

It's a technique called 'space diversity' that is also used in many radio
systems such as radio microphones and cellular radio. Two separate
receive aerials are used, on the principle that they are unlikely to both
be in a null at the same instant. There are various techniques to
implement this which range from simply coupling the two antennas together
to having entirely separate receivers and just selecting the best signal.
In your case it will complicate things considerably I'm afraid.

-Arthur

On Thu, 7 Oct 2004 22:53:55 -0700, "m" wrote:

I agree also that the second antenna is for diversity but none of the
manufactures can answer why it is needed if spread spectrum is suppose to be
so innately fade resistant, even though the spread spectrum Wifi is using
now with CCK isn't really truly random anymore.

With very short direct and reflected paths typical in indoor
installations, you are not going to have multiple nulls within the
spread frequency range. In order for a spread spectrum system to work
properly in a multipath environment, there should be several multipath
nulls within the spread range. If there is only one wide dip covering
your spread range, it can take out most of your signal.

If the spread range is from say 2400 to 2430 MHz, the corresponding
wavelengths would be 125 - 123.5 mm so the spread in wavelengths is
1.5 mm so you would need at least a path difference 40 wavelengths or
5 m to get a phase difference of 180 degrees that would cancel the
signal at one end of the spread range but add up in the other end.

With a larger frequency spread or larger path distance between the
direct and reflected signal, the phase difference would be several
times 360 degrees and there would be multiple nulls (at 180, 540 etc
degrees) and multiple peaks, within the spread range, which the spread
spectrum modulation could easily tolerate.

Paul OH3LWR

I agree also that the second antenna is for diversity but none of the
manufactures can answer why it is needed if spread spectrum is suppose to be
so innately fade resistant, even though the spread spectrum Wifi is using
now with CCK isn't really truly random anymore.

Russ N7QR
russm at ieee dot org

"Arthur" wrote in message
news
On Thu, 7 Oct 2004 12:00:52 +0100, Daniel Kelly (AKA Jack)
wrote:

I'm not a radio engineer so I'm not too sure what's going on. If I was
forced to make an educated guess then I'd say the problem was caused by
interference caused by reflected, out-of-phase signals. For example,
it's
conceivable that the receiver could be picking up two 'copies' of the
transmission from the TX - one 'copy' is the direct signal, the other
'copy'
is a signal that has bounced off some surface - these two 'copies' would
cancel eachother out when they're 180 degrees out of phase.

Yout assumption is absolutely correct


So... this brings me to the question I posed in the subject line: why do
WiFi basestations use two antennas? For example, my 3Com basestation
has
two little whip antennas sticking out the back. Are these two antennas
used
to combat the problem that I am experiencing? If so, how can I
implement
two antennas on my system?

It's a technique called 'space diversity' that is also used in many radio
systems such as radio microphones and cellular radio. Two separate
receive aerials are used, on the principle that they are unlikely to both
be in a null at the same instant. There are various techniques to
implement this which range from simply coupling the two antennas together
to having entirely separate receivers and just selecting the best signal.
In your case it will complicate things considerably I'm afraid.

-Arthur

I agree also that the second antenna is for diversity but none of the
manufactures can answer why it is needed if spread spectrum is suppose to be
so innately fade resistant, even though the spread spectrum Wifi is using
now with CCK isn't really truly random anymore.

Russ N7QR
russm at ieee dot org

"Arthur" wrote in message
news
On Thu, 7 Oct 2004 12:00:52 +0100, Daniel Kelly (AKA Jack)
wrote:

I'm not a radio engineer so I'm not too sure what's going on. If I was
forced to make an educated guess then I'd say the problem was caused by
interference caused by reflected, out-of-phase signals. For example,
it's
conceivable that the receiver could be picking up two 'copies' of the
transmission from the TX - one 'copy' is the direct signal, the other
'copy'
is a signal that has bounced off some surface - these two 'copies' would
cancel eachother out when they're 180 degrees out of phase.

Yout assumption is absolutely correct


So... this brings me to the question I posed in the subject line: why do
WiFi basestations use two antennas? For example, my 3Com basestation
has
two little whip antennas sticking out the back. Are these two antennas
used
to combat the problem that I am experiencing? If so, how can I
implement
two antennas on my system?

It's a technique called 'space diversity' that is also used in many radio
systems such as radio microphones and cellular radio. Two separate
receive aerials are used, on the principle that they are unlikely to both
be in a null at the same instant. There are various techniques to
implement this which range from simply coupling the two antennas together
to having entirely separate receivers and just selecting the best signal.
In your case it will complicate things considerably I'm afraid.

-Arthur

On Thu, 7 Oct 2004 22:53:55 -0700, "m" wrote:

I agree also that the second antenna is for diversity but none of the
manufactures can answer why it is needed if spread spectrum is suppose to be
so innately fade resistant, even though the spread spectrum Wifi is using
now with CCK isn't really truly random anymore.

With very short direct and reflected paths typical in indoor
installations, you are not going to have multiple nulls within the
spread frequency range. In order for a spread spectrum system to work
properly in a multipath environment, there should be several multipath
nulls within the spread range. If there is only one wide dip covering
your spread range, it can take out most of your signal.

If the spread range is from say 2400 to 2430 MHz, the corresponding
wavelengths would be 125 - 123.5 mm so the spread in wavelengths is
1.5 mm so you would need at least a path difference 40 wavelengths or
5 m to get a phase difference of 180 degrees that would cancel the
signal at one end of the spread range but add up in the other end.

With a larger frequency spread or larger path distance between the
direct and reflected signal, the phase difference would be several
times 360 degrees and there would be multiple nulls (at 180, 540 etc
degrees) and multiple peaks, within the spread range, which the spread
spectrum modulation could easily tolerate.

Paul OH3LWR

On Thu, 7 Oct 2004 22:53:55 -0700, "m" wrote:

I agree also that the second antenna is for diversity but none of the
manufactures can answer why it is needed if spread spectrum is suppose to be
so innately fade resistant, even though the spread spectrum Wifi is using
now with CCK isn't really truly random anymore.

With very short direct and reflected paths typical in indoor
installations, you are not going to have multiple nulls within the
spread frequency range. In order for a spread spectrum system to work
properly in a multipath environment, there should be several multipath
nulls within the spread range. If there is only one wide dip covering
your spread range, it can take out most of your signal.

If the spread range is from say 2400 to 2430 MHz, the corresponding
wavelengths would be 125 - 123.5 mm so the spread in wavelengths is
1.5 mm so you would need at least a path difference 40 wavelengths or
5 m to get a phase difference of 180 degrees that would cancel the
signal at one end of the spread range but add up in the other end.

With a larger frequency spread or larger path distance between the
direct and reflected signal, the phase difference would be several
times 360 degrees and there would be multiple nulls (at 180, 540 etc
degrees) and multiple peaks, within the spread range, which the spread
spectrum modulation could easily tolerate.

Paul OH3LWR