FREQUENCY DEPENDANCY ON BEAMFORMING

hi all,
      i am caught up in a frequency and beam forming doubt.

Say in a typical OFDM COMMUNICATION system, assuming following stats

FFT size = 512 left_guard carriers = 41 right_guard_carriers = 40 and
one dc carrier

so, overall the data carriers are 512-40-41-1 = 432

In the 432 data carriers, dividing them into 18 subchannels
(SUBCHANNEL 1 TO SUBCHANNEL 18) each of 24 subcarriers (18*24 = 432).

And now my question is,
Assume in the Nth frame, my data is placed in the first subchannel
(SUBCHANNEL 0) when i (mobile subscriber station or customer premise
equipment) am transmitting data to BASE STATION.

in the (N+1) th Frame, my data is placed is placed in 18th subchannel
(SUBCHANNEL 18) when i am transmitting data to the BASE STATION.

so based on the Nth frame's first subchannel data, if i calculate the
beamforming weights (i mean the location signature) (ASSUMING I HAVE
UNIFORM LINEAR ARRAY ANTENNAS) at BASE STATION

and for the next (N+1)th frame, if i calculate beamforming weights (i
mean the location signature) (ASSUMING I HAVE UNIFORM LINEAR ARRAY
ANTENNAS) at BASE STATION..

THE QUESTION is..HAS FREQUENCY GO TO DO WITH CALCULATING SIGNATURE
LOCATION..WILL MY BEAMWEIGHTS VARY for Nth and (N+1)th frame..

i will put answer as NO..beamweights should be same for both cases..is
there anywhere the frequency dependancy factor creeps in..i saw..lot
many..papers on frequency dependant and frequency
independant..beamforming

can anyone explain whats the difference between those two and where
does these two fit in.with respect to narrowband and broadband
communications

thanks and regards
particle (filter) reddy

"PARTICLEREDDY (STRAYDOG)" <particlereddy@gmail.com> wrote in message 
news:195e299f-984d-4463-bceb-ac343e11bb6f@l28g2000prd.googlegroups.com...
> hi all,
>      i am caught up in a frequency and beam forming doubt.
>
> Say in a typical OFDM COMMUNICATION system, assuming following stats
>
> FFT size = 512 left_guard carriers = 41 right_guard_carriers = 40 and
> one dc carrier
>
> so, overall the data carriers are 512-40-41-1 = 432
>
> In the 432 data carriers, dividing them into 18 subchannels
> (SUBCHANNEL 1 TO SUBCHANNEL 18) each of 24 subcarriers (18*24 = 432).
>
> And now my question is,
> Assume in the Nth frame, my data is placed in the first subchannel
> (SUBCHANNEL 0) when i (mobile subscriber station or customer premise
> equipment) am transmitting data to BASE STATION.
>
> in the (N+1) th Frame, my data is placed is placed in 18th subchannel
> (SUBCHANNEL 18) when i am transmitting data to the BASE STATION.
>
> so based on the Nth frame's first subchannel data, if i calculate the
> beamforming weights (i mean the location signature) (ASSUMING I HAVE
> UNIFORM LINEAR ARRAY ANTENNAS) at BASE STATION
>
> and for the next (N+1)th frame, if i calculate beamforming weights (i
> mean the location signature) (ASSUMING I HAVE UNIFORM LINEAR ARRAY
> ANTENNAS) at BASE STATION..
>
> THE QUESTION is..HAS FREQUENCY GO TO DO WITH CALCULATING SIGNATURE
> LOCATION..WILL MY BEAMWEIGHTS VARY for Nth and (N+1)th frame..
>
> i will put answer as NO..beamweights should be same for both cases..is
> there anywhere the frequency dependancy factor creeps in..i saw..lot
> many..papers on frequency dependant and frequency
> independant..beamforming
>
> can anyone explain whats the difference between those two and where
> does these two fit in.with respect to narrowband and broadband
> communications
>
> thanks and regards
> particle (filter) reddy

I'm assuming that you're concerned with beamforming and beam patterns.  I 
don't fully understand the "signature" comment.  That would seem to have to 
do with signal processing and not so much beamforming unless you have an 
adaptive beamformer.  But, if you did then you'd probably not be asking 
about the weights staying put.

Beams are dependent on frequency.  The question to consider is: "how much?". 
Some wideband beamformers have been implemented to work on sets of frequency 
octaves and maybe that's "good enough" for some applications.  If it is, 
this implies that relatively wideband signals or sets of signals can work 
with the same beamformer / weights.  Big IF.

I would forget about all the signal characteristics (frames, etc.) and focus 
on center frequency.  Either one beamformer (se of weights) will work 
adequately over the intended range of center frequencies or it won't.  With 
the information provided, only you can tell.

I would do this:
Calculate the beam pattern at the highest frequency and at the lowest 
frequency and decide if the patterns at the extremes are adequate.  If they 
are then OK.  If they aren't then you would appear to need multiple sets of 
weights.  If that works in your implementation then that's good.  If not 
then there's more engineering to do.

Fred