[time-nuts] time-nuts Digest, Vol 105, Issue 72

[Jim Lux]

On 4/21/13 10:37 PM, Stewart Cobb wrote:
> Date: Sat, 20 Apr 2013 12:52:03 -0700
>> From: "Tom Van Baak" <tvb at LeapSecond.com>
>> To: "Discussion of precise time and frequency measurement"
>>          <time-nuts at febo.com>
>> Subject: Re: [time-nuts] pin-wheel antenna
>> Message-ID: <EB790CDDE52944E1A03608CA6CFCCCA8 at pc52>
>> Content-Type: text/plain;       charset="iso-8859-1"
>>
>>>>>> There's a very nice picture of a pinwheel from Novatel on the back
>> cover
>>>>>> of the March issue of GPS world..
>>
>
>
>> It's a thing of mysterious beauty. And the GPS World photo saves me from
>> the temptation to break open my own pinwheel antenna just to see what's
>> hidden inside.
>>
>> /tvb
>>
>
> In a previous job, I was actually encouraged to dismantle failed GPS
> antennas.  Attached please find scans of an earlier Novatel pinwheel
> antenna, model 702L (I think), covering L1, L2, and the OmniStar band (1535
> to 1559 MHz).  Note the two different lengths of the topside slots, one set
> each for L1 and L2.  The termination resistor on the backside spiral feed
> measures 133 ohms.  Its presence confused me at first, but it's only
> terminating the "wrong" circular polarization -- the correct polarization
> reinforces itself toward the connector.  Scanned on a desktop scanner (not
> a camera) at 150 dpi.  The substrate is a typical Teflon low-loss material
> and measures 0.063 inches thick, and the diameter measures about 5.628
> inches.  If anyone wants to read dimensions off the scan and model this on
> HFSS or something similar, I'd love to see the results.
>
> The more complex flame-like structures at the tips of the slots in the
> photos of the current model are probably attempts to broadband the basic
> antenna to cover L5, Glonass, and/or the new Galileo bands.  The operating
> principle of my older one is obvious almost by inspection; not so the new
> one, which probably could not have been designed without computer modeling.
>

The one you scanned is a pretty straightforward spiral (cavity backed, I 
assume, when in the package), with two different length spiral arms for 
L1/L2. And slot fed, as noted in the literature.  I've seen this 
strategy on linear arrays, but hadn't seen it for a spiral, but of 
course, it's a clever way to do it.

I think the new one is a bit more clever in terms of broadbanding; 
perhaps like a fan dipole or bowtie, with the arms of non-constant 
width.  The spiral pitch is also a lot steeper.  It's sort of half way 
between the one you have scanned and a conventional crossed dipole.