Suggestion and comments on GPS please [resolved, see list of improvements]

edited July 2018 in Problems Now Fixed

Would be interested to hear how people have achieved a good reliable GPS fix. I have tried and failed on a number of occasions, and so far tried the following.

Mag mount antenna (with 26dB LNA) mounted outside on metal plate.
Moving the antenna as far away from the house as coax will allow
Changing the antenna for another similar mag mount
Cutting off the thin coax and replacing it with thicker lower loss double screen coax.

I still can't get a reliable fix, sometimes its shows tracking 6,7,8 SATS and good 1,2 or 3

The position of the SDR makes it very hard to position the antenna with a clear view of the sky, it has around 40% clear view and other areas are obstructed by the house, trees, fence.
On occasions, it runs for while, 3.6K fixes but then stops.

I have checked the voltage on the SMA, thinking it could be an issue but its fine 3.3v

So how do others get on, what exactly do you need to do to get a reliable fix?

enclosed image shows a typical screen, tracking

It's a little disheartening with TDoA as I really wanted to be part of the system, and located close to London. Could my KIWI be faulty, of do you really have to have an antenna with 90% clear views of the sky?

Dave M0TAZ

Attachments:
https://forum.kiwisdr.com/uploads/Uploader/09/32600f840c5418dbcd90ef52b09ef0.jpg

Comments

  • It's difficult to tell, but it sort of looks like it's severely attenuated. When you spliced in the better coax how good was your splice on the antenna end? A bad impedance bump here would not be good. Does the new coax terminate in an SMA on the Kiwi end? How long is the new run? What kind of coax? Look at the chart in this document to understand GPS signal loss issues: http://freqelec.com/gps_gnss/gps_ant_issues_r1_5-07.pdf

    Can you try a different antenna? We've had a few failures of the hockey puck mag-mounts that Seeed ships. Having only 40% view of the sky shouldn't be a problem. Especially not now that we have Galileo support.

    In your image PRN29 having an RSSI of 528 is marginal. Those Galileo E1/8/9 trackings are false because the SNR was only 16 at acquisition. An RSSI in the 200s means the tracking loop is unlocked (and the "U" status symbol agrees).
  • It's difficult to tell, but it sort of looks like it's severely attenuated. When you spliced in the better coax how good was your splice on the antenna end?

    Its probably not the best, but the antenna seems to pick up more after the change than before. In the end, I took the antenna out of the plastic box and terminated the braid onto the metal can behind the patch antenna. I read the document that's helpful, but how does one select an antenna that has "good out of band filtering" they all seem to come from China for £5.99 or less.

    I did try another antenna that was factory original, but that didn't seem to help similar results.

    The new section of coax is about the same size as RG58, but its much better, double screen, foil screen and solid centre core. I have around 3 or 4 meters, and it was terminated with an SMA from the manufacturer.

    I dont have any issues with the QRP labs WSPR GPS, works just fine indoors and in the garden. I except the implementation of GPS in the KIWI needs a little more care than the QRP labs version and did hope the outside antenna would fix. I have ordered another 2 GPS antenna, one a chepo china special 26dB LNA and the other says "marine GPS with 25dB amp" to see if they help.

    The image shows the best I have had for a while, what does the PLL "kick" signify?


  • Okay, a few things here. "I except the implementation of GPS in the KIWI needs a little more care than the QRP labs version and did hope the outside antenna would fix". Just so I understand: does this mean you were able to plug the same antenna/coax into the QRP Labs WSPR device and it received better than the Kiwi?

    Your new image above looks fine actually. RSSI in the 1200-1500s for Galileo is pretty normal. The GPS RSSIs are a little low. What concerns me is the shadow map. It's only been a 6 hour run, but it's clear you only see sats above 45 deg elevation. Worse, because you're at 51 deg latitude the GPS polar "hole" is coming into effect and that is limiting what you get above the W-E line. So your actual sky view is smaller than I thought.

    But having only 212 fixes over a 6 hour period is terrible. Another issue is that when you limit your sky view to one area like this you have high "dilution of precision" (GDOP, https://en.wikipedia.org/wiki/Dilution_of_precision_(navigation)). This will cause the time/positions to be less precise and they may be so far off that they get filtered out.
    M0TAZ
  • The QRP labs have a patch antenna built onto the GPS receiver, you don't have the option to add an external antenna. This unit will fix inside the house, on the ground floor with little or no view of the sky.
    I think from what you say my view of the sky is still too limited, I don't have a very good view of the horizon in any direction. Its worst to the North as the house is in the way, better South and West as you will see from the new shadow map.

    6K fixes in 15 Hrs

    I took the patch antenna out of its plastic housing and mounted it on a metal plate, earthing the shield of the coax to the metal plate. The metal plate is around 40cm x 40cm and it is
    better, but still not reliable.

  • Your latest shadow map has the classic shape described in this post: http://forum.kiwisdr.com/discussion/1083/gps-antenna-aiming
    M0TAZ
  • OK thanks, I have learnt a lot from these posts.

    I feel the need to elevate the antenna somehow, to give it a better view of the horizon.

    That's the next part of my resolution strategy :-)
  • OK, an update for the group. The GPS issue is mainly fixed now, these are the steps I needed.

    1/ Remove GPS patch from the puck, replace coax with something double screened.
    2/ Attached the braid and shield of the puck can to a metal plate 40 x 40 cm
    3/ Elevated this in the air to provide some clearance for house/trees/fences/local clutter

    Each time I did one of the above it got better, and I hope its now mostly fixed.




    WA2ZKDKA7ULX1DQ
  • Well done! It's amazing that devices with commercial GPS solutions (cellphones etc.) work indoors as well as they do. They just have better processing software than the open source solution we use.
  • I was so happy with the puk on a metal plate I decided to pick up a decent GPS antenna from ebay. HUBER+SUHNER 1315.17.0008 1575MHz 34dB N female gps antenna antenne Aerial (ebay search terms)
    This is now coupled to a short length (circa 6m) of LMR400. That really is the best I can do now, short of re-locating the antenna on the roof. The GPS antenna was £15 delivered from China, and it would seem to be the real deal.

    WA2ZKD
  • edited August 2018
    My project for tmrw is to use a Spectracom GSG-6 GPS simulator/generator to map actual signal strength at the Kiwi GPS input to reported RSSI. Report to follow


    https://spectracom.com/products-services/gnss-simulation/gpsgnss-simulators?gclid=CjwKCAjw2MTbBRASEiwAdYIpseYAdL4X4Yesk-UGnJ_EYn9Hags3PRNiwBCbjQr8x5DRX_YVIZQ5WxoCTQEQAvD_BwE
    M0TAZ
Sign In or Register to comment.