GPS antenna alternatives

I bought one from via eBay for $48 and with its 40 dB gain will work with a long feedline and can be split to run other devices. It's not in the permanent mounting spot yet.



  • This is the Trimble antenna I'm using.

    It's powered by a Trimble Thunderbolt, but the antenna is passively split four ways (introducing a further 6dB loss) and connects to the KiWi's via 20m of 75 ohm Sat TV coax.

    The antenna itself is mounted on the ridge of the house approx 20ft AGL. It's slightly obscured to the North by a mounting pole for a dual band collinear, but still works OK.


    Martin - G8JNJ

  • edited March 2018
    My shadow to the north is caused by a large hill  Edit: or so I thought, looking at a sat constellation map, there's none up there from me!

  • I'll be posting the full story of the shadow soon
  • Yes, I've noticed this "too perfect" shadow on other Kiwis. But I'll let you tell the story..

  • Just to add to this dormant thread, yesterday I installed a PCTEL GPS-TMG-40N antenna with low-loss feeder and it certainly does the trick. I went for the 40 dB version so there's plenty of signal in hand to feed a 4-way passive splitter when I get one. I'm somewhat screened by tall trees in the south-west quadrant, so don't get many low-elevation signals from there.

    The KiwiSDR GPS software only seems to allow a maximum of four Galileo satellites in the 12 channels. Is this intentional? While it all works fine, it would be nice if the limit could be raised - either by allowing more than four Galileo sats, or by allowing more than 12 channels. Also, surely there isn't much point in scanning for QRZZ satellites in Region 1?

  • 24 hours from now the map will be interesting
  • jksjks
    edited August 2018
    This business of the 12 GPS channels being split between 4 for Galileo and 8 for Navstar/QZSS is not permanent. It has to do with being out of FPGA memory (in the RX4WF4 configuration at least). To be more precise it has to do with a the hurried way in which Galileo PRN lookup is currently implemented. A more efficient way of using FPGA memory to store the Galileo PRN tables would allow all 12 channels to support all sat types. Someday I'll get around to doing that.

    Galileo is different than Navstar or QZSS in that you have to use a lookup table instead of a simple LFSR to do the PRN generation.

    Yes, QZSS acquisition could be bypassed in many parts of the world if you assumed the Kiwi was fixed and used some of the other initial geolocation information. I've actually seen some of the Swedish Kiwis acquire QZSS sats. I guess the geometry works out.
  • I found this type of cheap "marine" style GPS antenna to work much better than the cheapo GPS puk. Could I ask if you are posting about the GPS antenna you are using, then a link to the website would be helpful TA

  • I wonder if this is still a patch-type antenna but with a bit of a ground plane. Nice that the coax looks to be better than RG174.
  • This is what my shadow map looks like after 18 hours. As expected, there's not much low-elevation stuff in some directions because of trees. In other directions, where the optical horizon is 20+ km away, satellites are being received at elevations as low as 3°. The 'polar hole' is clearly visible.

  • It's worth noting that four more Galileo satellites (23-26) were launched on Wednesday 25 July 2018 and are currently being commissioned.


    Martin - G8JNJ
  • I also ordered one of these (ebay description including spelling mistakes) for £15 - HUBER+SUHNER 1315.17.0008 1575MHz 34dB N female gps antenna antenne Aerial
    If it lives up to expectation (and gets to me from China) then it should allow for some decent coax and further elevation.

    I can't get the image upload to work (says uploading wait....) so you can see a picture here.

  • "I wonder if this is still a patch-type antenna but with a bit of a ground plane. Nice that the coax looks to be better than RG174."

    Most of the cheap Chinese 'marine' GPS antennas I looked at were basically patch types and perhaps had an extended ground plane.

    I'd certainly be interested to see how well one performs in comparison to the patch antenna supplied with the KiWi.


    Martin - G8JNJ
  • When the four recently-launched Galileo satellites are available, the constellation will be nearly complete with 22 satellites in 24 slots (but no orbiting spares yet). At the moment the A plane has working satellites in all 8 slots, the B plane has 3/8 (becoming 7/8 with the four new ones), and the C plane is 7/8 already. The next launch slots aren't until 2020-2022 so it will take until then to finish the job with 8 working satellites and two orbiting spares in each plane.
  • I never really thought about GPS reception being able to model the local horizon, I suppose the big data guys can correlate shadow to reported location making GPS spoofing harder.
    I've ordered a higher gain antenna to test.
  • jksjks
    edited August 2018
    It's worth noting that four more Galileo satellites (23-26) were launched on Wednesday 25 July 2018 and are currently being commissioned.
    I'm a couple days into getting each GPS channel to allow all satellite types. Very difficult (for me at least) as expected. The old inefficient way would have required 12 memory blocks and we only have 5 available (4RX4WF config). The new scheme requires only 3. But I had to learn some new Verilog and employ some tricky sequential/producer, parallel/consumer logic with pipelining issues thrown in. It isn't working yet..
  • Okay, got it. v1.216 is out with Galileo available on all 12 channels. Another small improvement towards maximizing the number of Kiwis available for TDoA.
  • jksjks
    edited August 2018
    Here's an example of all sats being available on all channels on a Kiwi in Iceland. As a bonus QZSS is being received (!) The signal is weak and only 9 degrees above the horizon. It only participated in this one GPS solution before sync was lost. But it was there!


  • Related question - I'd upgraded/run with the PCTEL 40dB antenna for a while to great effect; now suddenly I'm getting poor performance where there is almost nothing. A few sats show some RSSI about mid-green-bar for 10-15 seconds then they drop off the list. Cable should be fine. Are there any diagnostics besides the GPS admin page that I can run remotely?
  • Is the dc side OK, connectors tight? You may be able to check the antenna voltage with a t-piece, maybe wait until it is bad. At work support issues "GPS loss" is about 90% cable related.
    Also check the connectors are not just position critical, wiggle 'em.
  • Check the current draw of the PCTEL antenna and also its specified operating voltage.

    You may have to power it from a separate bias tee if the KiWi can't deliver sufficient voltage or current.


    Martin - G8JNJ
  • I just checked the KiWi GPS power circuit.

    It supplies 3.3v and will current limit at 50mA.

    So if your antenna requires a 5v supply or draws anything close to 50mA, that may be the problem.


    Martin - G8JNJ
  • Good observation 30ma at 5V so probably marginal at 3.3

    DC Voltage:
    3.3-9.0 V (operating)
    ?28.0 V (survivability)
    DC Current:
    < 30 mA @ 5 V
  • @W1EUJ: Use something like,gal,qzs&up=1 in case there happens to be a bad configuration of sats overhead. Right now you seems to have a good assortment spread over the sky (GDOP number is low). So that's not the problem.
  • edited August 2018
    Looking through a selection of PCTEL 40dB antenna specs they say typically less than 40mA, but can be up to 60mA.

    I think this may be the problem, if the current limit (required to prevent damage if the GPS antenna or coax becomes short circuit) in the chip used in the KIWi is getting close to its operating point, it may start to shut down or reduce the voltage being fed to the antenna.

    Try searching on your favourite web auction site for a 'satellite tv dc injector' they look like this

    One side connects to your KiWi the other side with the DC present goes to the GPS antenna feed. The DC supply cable goes to a suitable power supply.

    You may need some adaptors to convert from F connectors to whatever you are currently using, but the whole lot shouldn't cost more than about $10 USD if you shop around.


    Martin - G8JNJ
  • Also coax and connector losses, the PCTEL one I was pulling spec from was GPSL1-TMG-SPI-40NCB.
    "Bias Tee" is a productive search term on Ebay BTW.
Sign In or Register to comment.