[Date Prev][Date Next][Thread Prev][Thread Next] - [Date Index][Thread Index][Author Index]

Re: variable phase delay / phased arrays



At 5 GHz a one wavelength delay is 200 ps so I'd expect that 5 ps accuracy would
be required for beamforming. Isn't this a bit fast for DSP?

73,

John
KD6OZH

----- Original Message -----
From: "George Kadziolka" <gkadziol@kaztek.com>
To: <Dquagliana@aol.com>; <amsat-bb@AMSAT.Org>
Sent: Friday, 14 November, 2003 06:02 UTC
Subject: RE: [amsat-bb] variable phase delay / phased arrays


> Doug Wrote:
>
> > Question to the group:
> >
> > Suppose I have an "n by n" patch array at my ground
> > station.  Suppose I want to receive the satellite as
> > it moves across the sky using my "n by n" patch array.
> >
> > Instead of phasing the LO to the mixer for each patch,
> > could one instead use a single LO that is fed
> > to each of the mixers, and do all of the phase
> > processing on the signal output from the mixer (say at
> > audio with DSPs) instead of transforming the phase
> > of the LO? Wouldn't this be equivalent?
>
>
> In short ... Yes!
>
> What you are doing is beamforming, or spatial filtering.  Each patch
> represents a sample in space.  As a plane wave comes in, a line of constant
> phase passes each patch at a different time.  If you put in appropriate time
> delays after each patch, you could effectively line up the phases from each
> patch and add up the contributions to maximize the signal from the direction
> that the plane wave came from (ie steer the beam).
>
> For a single frequency, the time delay is equivalent to a phase shift.  When
> we are talking about bandwidths as narrow as several 10's or even 100's of
> kHz at a couple of Gigs, you can safely use the single frequncy
> approximation and deal with beamforming as the application of appropriate
> phase weights to each channel (ie patch, or spatial sample).
>
> Now that we're dealing with the phase of the signal, we can apply the phase
> shift anywhere in the signal chain.  As you pointed out, we could make our
> job easier by applying the shift at as low a bandwidth as we need to acquire
> our signal of interest (need to be at least wide enough to be able to
> compensate for Doppler after the beamformer).  You can mix all the way down
> to baseband, as long as you do it coherently (ie you must maintain the phase
> relationship between the spatial samples ... er, I mean patches).  A common
> quadrature LO (or single LO with an I/Q splitting mixer)would do the trick.
> From then on (after A/D conversion), a DSP could easily steer the beam.  Of
> course, you would need n^2 quadrature receive channels for an n x n array.
> Nowadays, you can print the receiver on the back of a patch, distribute your
> LO to all patches, and run the digitized receive signals to a DSP.
>
> This opens up way more possibilities like calculating beam pointing errors
> and updating the phase coefficients to automatically track the satellite
> signal in a closed loop, and without the need for motorized AZ-EL mounts.
> One can even track several satellites at once.  But that could be another
> thread ...
>
> Cheers,
>
> George
> VE3LQN
>
>
> ----
> Sent via amsat-bb@amsat.org. Opinions expressed are those of the author.
> Not an AMSAT member? Join now to support the amateur satellite program!
> To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org
>


----
Sent via amsat-bb@amsat.org. Opinions expressed are those of the author.
Not an AMSAT member? Join now to support the amateur satellite program!
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org



AMSAT Top AMSAT Home