More GPS Problems - Damaged hardware?

2»

Comments

  • Martin, are your loss values measured thus including the 50-75 mismatch loss?
  • Hi Jim,

    The figures were estimates for guidance purposes when planning a distribution system, but realistically a 75 to 50 Ohm 1.5:1 SWR mismatch is not going to cause any major problems.

    Any additional mismatch loss will be minimal (typically less than 0.5dB) and that's assuming the GPS antenna and KiWi GPS input, and coax is exactly 50 or 75 Ohms (which is unlikely). If you try to add impedance matching networks to improve the situation, you often end up with much greater losses, or only achieve very narrow band matching, so it's often better to just live with it.

    The losses can be greater if you have an 'unfortunate' combination of termination impedance's and length of cable which bring about further impedance transformation in the system, but as the amplifier in the GPS antenna has so much gain it will pretty much determine the GPS receiver noise figure, so any minor impedance mismatches downstream will not have a major impact.

    Of course you also need to factor in cable and connector losses. I tend to use 75 Ohm low loss foam filled foil and braid satellite TV coax
    (but if you are interested in the VLF/LF bands, not copper coated steel inner as the limited skin depth and ferromagnetic properties add extra loss) for most of my receive antennas, including the GPS distribution, as I find that for the money, it provides the best screening & lowest loss and is cheap to add to, or replace, as required. I also use a lot of compression F type connectors, as they are cheap and quick to fit if you have the right tool, they also indicate to me that the cable is carrying DC (as I have a lot of active antennas) so I don't accidentally mix up signal only (BNC) with signal & DC (F).

    I've mixed 50 and 75 Ohm components in many receive systems over the years (even professionally), and as long as you take care (and the connectors will physically mate correctly) you shouldn't experience any major problems. Transmission systems are different as you need to transfer as much power as possible with minimal losses and impedance bumps, the only factor you really need to consider in receive systems is noise figure and the subsequent Signal to Noise Ratio. Note that this last comment is a generalisation and there will be exceptions.

    Regards,

    Martin - G8JNJ
  • I have a splitter with DC pass on one port on the way, so I can get rid of the Rube Goldberg set of RF adapters I have now.

    Thing are working now that I have the 40 dB antenna outside, but it's still a bit of a mystery to me why I've had such issues with GPS vs many other users who are able to put their basic GPS antenna on a bookshelf on a small sheet of aluminum foil as a ground plane, and get good results. I did notice when I first got the KiwiSDR that I needed to carefully position the supplied antenna. Moving it an inch caused a substantial difference in results. Reflections/etc from the structure of the house? (Wood frame, asphalt shingles, shack is in top floor with a large double window facing south where I had the antenna located) Interference issues? I have several computers in the shack, who knows how much RFI is being spewed on 1575 MHz. When I first got the 40 dB high gain GPS antenna, placing it on the desk in front of the south window did not produce good results either, usually 1 sat at most. The Trimble Thunderbolt 10 MHz reference worked OK even with a modest antenna in the same location, but as others have noted they typically have a more robust GPS receiver. Moving the antenna about 10 ft to the other side of the shack helped tremendously, and of course outside is best of all as expected.
  • GPS performance can be affected by signal strength and multi-path. The latter can be quite an issue on indoor and/or window sill installs. On my day job, we did a study of GPS and ending up getting rid of a commercial module in one of our products due to its inability to handle multipath.
  • Gain is not a good an indicator for the sensitivity of an active antenna. Too much gain can cause problems too. It should be just enough to compensate for cable and connector losses.
    It is also possible that the GPS antenna is picking up unwanted signals and amplifying them, causing a degradation of SNR of the GPS signal.

    More expensive GPS antennas have band pass filters to prevent overloading, and QFH antenna design to mitigate multipath fading.

    But a cheap antenna placed outdoors is usually still better than a professional antenna indoors.
  • Ironically I'm having now a similar problem on my shiny new second kiwi.

    I use an active GPS splitter, so each device should have the same signal. The first kiwi shows a sky full of satellites, the new one shows none. It had a few in the beginning, but now, a day later, it doesn't see them anymore.
    Both kiwis have the same GPS settings.
    I've changed the cable
    Changed the port on the splitter
    Gave it it's own GPS antenna - no difference.
    Voltage is present on the antenna connector (3.20-3.25 V)

    Ideas?
  • First off close the browser and reopen it between checks. Especially as there will have been a large update for the new Kiwi.
    I've spent a while on stale browser data.

    No RP-SMA leads or adapters in there by mistake? (obviously the second antenna test should rule that out if that antenna works elsewhere).

    What I'd do. Check it is running and not in the middle of some update, using the admin interface power down new Kiwi/Beagle completely, leave it for five minutes at least (in case anything is hot). Close the browser (all tabs).
    Use a known good GPS antenna directly connected to the Kiwi, power it up and watch what happens in the first few minutes.
    The only clue we have so far is "It had a few in the beginning" recreate "the beginning" with a good antenna and work away from that.

    Stu
  • edited August 2019
    In "the beginning", the antenna was connected the same way to the splitter. The signals were already a lot weaker, but I decided to take care about that later.
    I've restarted the new kiwi several times and made sure that nothing else was keeping it busy. Both run the same version v1.309. Browser cache has been flushed and the page reloaded countless times.

    I don't know if this is related to the GPS problem, but the system load on the new kiwi is much higher than on the old one. But kiwid being the only process, which is producing significant load.
    old kiwi: %Cpu(s): 58.6 us, 29.6 sy, 0.0 ni, 11.1 id, 0.0 wa, 0.0 hi, 0.7 si, 0.0 st
    new kiwi: %Cpu(s): 51.5 us, 42.6 sy, 0.0 ni, 0.0 id, 0.0 wa, 0.0 hi, 5.9 si, 0.0 st
    Further, the CPU never seems to be running in idle state on the new kiwi.
    Without kiwid running, the CPU is >95% idle.

    Edit:
    I got one sync, and the signal strength looks good. It seems that something keeps the kiwi from acquiring new sats.
  • Are there any power settings in the interface? I do not have power on the GPS antenna.
  • edited August 2019
    AFAIK power is always on the GPS SMA ~3.3V.
    If the connector/coax is not shorted (or open circuit due to a RP-SMA lead) it should provide power to the antenna without any extra steps.
    Stu
  • No +3.3V on the open GPS SMA connector or with an active antenna or splitter attached? The current limit is about 50 mA per the GPS front-end chip spec (Skyworks SE4150L).
  • jksjks
    edited August 2019
    Linux idle time: Don't worry about that. Due to realtime requirements that Kiwi software uses a private thread package which runs flat out 100% of the time in the eyes of Linux. In fact Linux showing idle time is a bad thing. It means the Kiwi software is getting stalled because of some Linux reason (probably device related).

    If your 2nd Kiwi really has no user connections (including internal WSPR auto-run) then the GPS really should behave the same as the first unit. One thing to check is the "FPGA eCPU" load percentage on the status tab of the admin page. That number should be roughly the same for both Kiwis and below 40-50%.
  • jksjks
    edited August 2019
    Voltage is present on the antenna connector (3.20-3.25 V) Is that on the SMA connector open-circuit or at the SMA connector with a load attached? If it's with nothing attached then something isn't right. Do you have 3.3V on the "3.3G" test point as shown in the troubleshooting guide? http://www.kiwisdr.com/ks/troubleshooting.pdf (be very careful when measuring that you don't short anything with the voltmeter probe).
  • @jks
    I don't know what happened. Without users, I got GPS sync. With users I got high system load (not userspace), no GPS sync and lag on the web interface.
    I've cloned the image from the old kiwi to the new kiwi and now everything works good. Problem not understood, but solved.
  • In your top(1) summary above for the new Kiwi it shows "5.9 si". So that Kiwi is spending nearly 6% of the time servicing Linux "software interrupts" whatever that means. I'm sure that's also responsible for the 2x increase in system time. My bet would be some Linux driver gone crazy.
  • That problem appears to be specific and reproducible to the latest 4.4 kernel, version 4.4.155-ti-r155 or one of its modules.
    I'm just mentioning it in case somebody else has that problem. 4.9.147-ti-r121 runs fine as far as I can tell.
  • Resurrecting the thread with pertinent information.

    I've just installed one of these: https://www.sv1afn.com/en/gnss-gps/-7.html
    LNA / bandpass filter / multiplexer. No external power required, the bias-tee from one of the fed devices works perfectly. Each attached device thinks it has an active antenna attached, and the system as whole has gain compared to the passive splitting.

    I changed out my antenna from the as-supplied puck to this: https://www.amazon.co.uk/gp/product/B07Y7WHKCT, which is sitting on the shed roof jammed into a shaped hole I cut in an upside-down steel biscuit tin. I was not having good results from passive splitting, so I went down the active route. SV1AFN appeared to have the only LNA+bandpass+multiplex that I could see for under a hundred euro.

    Both of my Kiwis are now being fed GPS signals, very loud and clear, and getting fast fixes.
    Plus, I've got capability of another two GPS feeds from the device.
  • edited July 2020
    I went down the higher performance outside GPS antenna route a year ago (see my post(s) above). I added a second KiwiSDR and split the signal to it as well, and both are getting great signals and usually tracking 11 or 12 satellites (I have a shadow area to the north due to a row of 40 ft evergreens). I think the hockey puck may work in some cases, but it's hit or miss.
Sign In or Register to comment.