[time-nuts] GPS failed... now fixed

[skipp Isaham]

Hello Again Everyone,  

I have good news, regarding the previously reported failed GPS story has been mostly 
resolved... and here is a partial report, more info to follow if anyone is interested. I also want 
to thank everyone who had previously replied, I read all the replies and hope to be able to 
respond to any questions, in more detail as things calm down a bit. 

Different GPS receivers (two diferent models and three different brands) at a mountain top
radio site in the San Francisco East Bay Area, all failed to lock, while having been in active 
service for some years.  It took some time to pool the proper resources and get back to the 
site.

Identified RFI source is a Mimosa Airspan B5c Backhaul radio (4.9-6.2GHz) spewing out a wide
offending signal that drifted around about 3 MHz above GPS L1 frequency 1575.42 MHz.

The newer Mimosa Dish was placed after the GPS antennas. The Mimosa radio was right next to 
our GPS Antennas.

Testing included Teeing into the GPS antennas, I/we use a Mini Circuits Splitter with DC block to a 
R&S ZVL, that revealed RFI at -34dB blasting into the GPS antennas. The frequency of the RFI was 
1579.3 +/- 500KHz.and drifting around 3MHz away the GPS Center Frequency. This RFI overloaded 
the GPS LNA's. even the SAW filter GPS's were overloaded when within 5 feet of the “M” radio.

A SAW filtered GPS was waved around the area, overloading it when placed next to the “M” radio.
At this point the RFI source was proven, so the GPS antennas were moved to a location away
from all other electronics about 12 feet away from the “dirty Mimosa”, 10 feet away from other P2P 
equipment.

All extra (original) coax was pulled back into the vault, and tied down in a large single loop to avoid 
changing impedance of the coax. The GPS's then locked, and the temp antennas were moved to a 
ladder next to the “M” for more testing of the SAW filter GPS's in this high RF environment.

The SAW filter antennas offered about 8 to 10 dB of attenuation from the offending “M” at only 3 MHz
away from designed center freq. In other words They were able to lock the Zypher at about 5 feet from
the “M”. The non filtered GPS's would begin to lock at about 6 to 8 feet away from the “M”. The existing
VIC100 GPS on MPLS equipment has filtering, and performed about as well as other filtered GPS's.

I/we can recommend either of the PCTEL GPS's listed below be purchased for systems in future design.
The two listed have a wide voltage range, so one GPS model could be used for both 3.3v and 5V equipment.
Additionally they could combine one GPS antenna with a high freq splutter, and DC block one leg to
eliminate the need for two GPS antennas out in a harsh environment.

The MPLS Panasonic GPS's do have a good out of band filter, Two of the three Filtered GPS's
performed about the same as the Panasonic VIC100, they are only 5volt.

Airspan Mimosa Tech support was contacted, to report the High level RFI found at the site, we are
waiting for a reply.  It most likely has a defect, the Mimosa radio was tape labeled “Kills GPS”, to identify
the bad radio, and alert others with GPS problems so they will become aware of the interference source.
The RFI was visible even on the temp GPS antennas +60 feet away.

Mimosa spec sheet:
https://mimosa.co/uploads/datasheets/Mimosa-by-Airspan-B5c-Datasheet_DS-0008-09.pdf

Best performing, and specs for the GPS's during testing at the site

$187 GPS-TMG-HR-26NCM1575 MHz 3.5 dBic
Timing Antenna, 26dB Amplifier radome shape
sheds water and ice, while eliminating problems
associated with bird perching.N Female,≥ 65 dB
@ 1559 MHz ≥ 65 dB @ 1625 MHz,  3.3- 12.0 V

$96Â  3971D-HR-DH-WHigh Rejection dual
SAW, Permanent Mount GPS Antenna: Small,
TNC Female, 28 dB gain, Out of band rejection
>50 dBc @ ≥ 40 MHz, 2.7 to 5.5 V

CCAH32ST01 Existing MPLS Panasonic GPS:
TNC, or N connector, Gain 38dB5V +/- 0.5 V,
Out of band rejection 60dB (Typical) at 1575.42 +/- 50 MHz

The High Gain +40dB model tested would be a good pick on extra long coax runs

*** In addition to a R&S ZVL, a $40 (Amazon) SDR with common free PC software was extremely
helpful.  A used MINI-CIRCUITS ZN2PD2-50-S+ 500-5000 MHz 2 Ways DC Pass Power
Splitter from Ebay (about $25 plus S&H, used) with a DC block allows both the ZVL and
SDR to work from the same antenna/source.

The original three GPS receivers are now "back on line", and we know a bit more about
the site environment.

thank you

cheers,
skipp

skipp025 at yahoo.com
skipp025 at jah who dot calm
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Mimosa_Airspan_and_GPS_Antennas.jpg
Type: image/jpeg
Size: 44248 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20220715/6a68fa7d/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Prox_Testing.jpg
Type: image/jpeg
Size: 28663 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20220715/6a68fa7d/attachment-0001.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ZVL_display_GPS_M2.jpg
Type: image/jpeg
Size: 32226 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20220715/6a68fa7d/attachment-0002.jpg>