Dead KiwiSDR troubleshooting [fixed with Beagle replacement]

edited December 2023 in Problems Now Fixed

One of my Kiwis is completely dead, with no lights coming on regardless of what power supply I try. I separated the boards and tried powering the BBG via a 5V microUSB connector from a phone charger. Absolutely nothing, no lights etc. So I assume the next logical step would be to buy a new BBG board, and then flash it with the correct firmware? I misplaced the original microSD card for that, I assume I can download it somewhere?

I also connected the power supply to the KiwiSDR board by itself...there are no lights but I don't think that board has any either. I have read about the common mode choke possibly getting fried, there is a rectangular component labeled 501Y which I assume is it, but it didn't get warm when attached to the power supply. I assume I should try measuring the voltage at the pins on the other side of the board.

Comments

  • This doesn't sound good.

    Have you got any idea how it may have happened ?

    Don't wory you can confess if you have to, we have all done silly things that we wish we hadn't...

    Does anything look, or smell burnt ?

    Can you actually measure the 5v supply on the appropriate pin of the Beagle header ?

    The KiWi image can be downloaded to an SD card. I'm pretty sure there are instructions posted soewhere on this forum, so make use of the search box.

    Likewise with the KiWi board if you apply 5v to the DC connector can you measure the 5v supply rail anywhere else on the board. Be aware it is routed out to the Beagle.

    The original design review document can be found here.

    http://kiwisdr.com/docs/KiwiSDR/KiwiSDR.design.review.pdf

    and the circuit can be found here

    http://www.kiwisdr.com/docs/KiwiSDR/kiwi.schematic.pdf

    The KiWi input filter is part T301 just behind the DC input socket on the KiWi board.

    Good luck,

    Martin

  • jksjks
    edited November 2023

    Rather than the design review document (which by now is horribly out-of-date) what you want is the troubleshooting guide: http://www.kiwisdr.com/ks/troubleshooting.pdf

    There is also a recent forum thread about how to bypass the 5V input choke: https://forum.kiwisdr.com/index.php?p=/discussion/2978/kiwi-1-how-to-fix-a-shorted-5v-input-cmc-t301/p1

  • Hi John,

    Ah that is a useful reference, I wasn't aware of that document.

    Regards,

    Martin

  • jks & Martin,

    Thank you for the useful information. The Kiwi in question died a few months ago during the summer; I was not present at the time so there is a good chance it was lightning related event. At least there is no charring evident on either board.

    On the Kiwi board by itself, with the barrel connector powered up, I am seeing 5.5V on the '5VE' pin, but nothing on the other ones in the line up (5V & 3.3) - I need to pause while I read the troubleshooting guide, then will take a closer look at the BBG board.

    Bruce

  • OK, I've checked all of the voltages on the Kiwi board that are shown in the troubleshooting guide. I am not seeing voltage present on ANY of the test points, except for the 5VE and the 5V IN by the barrel connector. Shall I assume the board is toast at this point? Or is there one specific component that would merit closer scrutiny, i.e. if replaced all the downstream points might be OK? My surface mount skills are non-existent but if limited to one component perhaps I could give it a try.

    Separately I'll mention that on this board (#5222), there is no T301, but there is a component in that slot labeled 501Y; I am assuming that is also a common mode filter.

  • jksjks
    edited November 2023

    We've had a number of cases where the Beagle gets blown up by (presumably) a very bad transient on the 5V input, but the Kiwi board survives. In essence the Beagle acts as a moderately priced fuse. So well worth it to replace the Beagle and see if it is fixed.

    This of course is what you want given that the Kiwi board is much more valuable. KiwiSDR 2 has a TVS on the 5V input in response to this.

  • T301 is the schematic part reference. 501Y is the CMC part value. It's all the same part.

  • John,

    OK, I guess I misunderstood, so you are saying that the voltages on the Kiwi test points are supplied by the Beaglebone board and not through the barrel connector on the Kiwi? If so I'll order a BBG immediately on Ebay.

    Bruce

  • in the past when I experienced such, I'd remove the BB from the Kiwi and see if the BB could be coaxed to life on its own.. sometimes you'll need to use a Debian image to test it on it's own. I had a number of BBG left over from Kiwi's that I upgraded with BBAI and ended up giving the BBG to guys who lost their's while the Kiwi cape survived.

  • Yes. Kiwi DC jack goes direct to 5VE on header. Down to BBG. BBG power controller switches it and sends back up to Kiwi board as 5V and 3.3V. So BBG is completely in the power path.

  • Thanks for confirming the BBG is in the power path. When separated from the Kiwi board the BBG certainly appears to be dead, with not hint of LED illumination when powered via the microUSB connector. Ordered a replacement today. Definitely worth a shot.

  • Quick update on this - the new BBG from Ebay ($32 including tax and shipping) arrived today. I flashed a 32GB microSD card with the Kiwi software, then went through the recommended steps to flash that to the BBG and....drum roll....success! That is, after the BBG flash transfer it properly booted up and began updating the software. As I write this, it is displaying the "software update in progress" message. The next big step will be to see if the Kiwi board is still alive once the software update is completed. Fingers crossed.

  • After a few hours runtime I am fairly confident now that the Kiwi board is functioning normally, confirming that the BBG board took the "hit" from the storm in August. Whew! Aside from this, I am looking forward to the release of the v2 KiwiSDR, especially if has better sensitivity for GPS reception.

  • jksjks
    edited December 2023

    The better sensitivity won't be available for some time (assuming it's even possible). It requires significant rework of the FPGA code which is already short on space. Until then best to invest in a good antenna, feed line and ground plane.

  • Thanks for the information, it does sound like there is a LOT more going in with the GPS signal processing than the standard PPS type of output that can be had with the inexpensive standalone modules that'll lock onto a dozen satellites almost immediately. The app on my smartphone will acquire 15-25 satellites indoors...but I surmise with things like TDOA on the Kiwi there is a much deeper level of data that has to be extracted.

Sign In or Register to comment.