[time-nuts] Re: Composite sky GPS - with an added digression

[Lux, Jim]

On 6/30/23 6:40 PM, glen english LIST via time-nuts wrote:
> Hi Jim
>
> Seems this question of mine is well covered in  the ether.
>
> Yes, might not be as bad as I think. I'll write a program to compute 
> various scenarios and also will be interested to see how the fixes get 
> affected  with loss of signal over some AZEL patch of the sky.
>
> The other thing- these narrowband antennas, over 24 MHz ish, I would 
> suspect they are NOT phase  coherent, and that a combined pair of 
> signals over identical sky would be anythign but flat frequency response.

Oh, they're probably phase coherent "enough" - and it may not make much 
difference.

Say your antennas are separated by a meter - that's 5 wavelengths at L 
band (20cm). If they were ideal, you'd get a series of fans to form 
grating lobes - the fans would be long in the direction crossing the 
line between antennas, and narrow in the direction parallel to the 
baseline between antennas.

Non ideal antennas make the grating lobes wiggle or be wavy along their 
long axis. Of course what YOU care about are the nulls (and to a lesser 
extent the phase smoothness as you traverse the pattern).  I've looked a 
lot of these kind of models and your saving grace is that the nulls are 
deep only when the signals from the antennas are equal strength, which 
doesn't really happen much.

If you were trying to do real time kinematic surveying to millimeters, 
and are depending on smooth phase response - yeah, probably not going to 
work.  Those folks obsess about apparent phase center displacements of 
millimeters over a hemisphere.  Good multiband choke ring or artichoke 
antennas are where it's at.

Or, if you need precise position calculation, then the "multiple 
receivers and post process" is probably a better approach, because that 
can explicitly address that the antennas are not co located.





>
> Yes, - the LNA. some of these have 40dB gain, golly. I have some 1500 
> MHz ferrite isolators, also. probably unnecessary as you say depends 
> on the S12 for the LNA Â  and thus the match the antenas see into the 
> splitter is likely important. Have spec-an to 50 GHz so.....

>
> Time now I think to set up a couple of receivers and get some raw 
> data., and compare with single receiver and combined antenna. Patch 
> antennas generally suck I find for all the usual reasons, I am a fan 
> of quadrafilar helicies.... but patches are a dime a dozen and 3what 
> is in use,  so that's what I will use for the tests and report back.

Indeed - sounds like fun.

I'll bet if you get raw observables out of your receiver, you could 
probably sort them (based on the look direction to each SV) into which 
antenna is more likely, and then run them through any of the online 
processors with some excision of "observations likely to be wonky"

(Which is a lot of work, but..)