Kiwi not coming up after update (remote site) [fixed with software reload, root cause unknown]

Comments

• KA7U

  November 2018

  My power supplies, 5vdc 2.5 amp analog supplies, fail to start the KiwiSDR from a power up. I find it necessary to pull the power plug from the KiwiSDR and plug it back in. Then the KiwiSDR will start. I speculate that the power supply does not come up to the required voltage quick enough for the KiwiSDR to start. I haven't worked at a correction for this, but I suspect if the power supply was on and a switch in the 5vdc line went on and off, the KiwiSDR would start. I'll be watching this thread for comments as I am interested.
  Ron
  KA7U
• n6gn

  November 2018

  I wonder if a monster capacitor on the 5VDC line would cure the problem. You have a 10,000 uF or more at 5V that you could stick across to see? That should keep it up for a few milliseconds.
  Glenn n6gn
• KA7U

  November 2018

  Glen, so the power goes off. The capacitor discharges to the KiwiSDR. The power comes on, the capacitor charges, the KiwiSDR fails. Repeat and it will fail again. If the power supply comes on, then a switch connects the 5vdc to the KiwiSDR it should start just like it does when the power plug is inserted into the KiwiSDR after the power supply is online.
  Ron
  KA7U
• Powernumpty

  November 2018 edited November 2018

  Ron I've noticed the same thing and was considering having a cap at the Kiwi end fed via a diode from the 5V so it is at ~4.3V shortly after the 5V is up. Then have a double pole relay short the diode to the 5V power rail and connect that to the Kiwi, my thinking is that the 5V rail does not rise cleanly enough for the Kiwi and the BBG does not like it. With a local cap it should go to 4.xV very quickly then the supply only has to bring up the last volt or so.
  They do warn against having any voltages on BeagleBone pins before the supply, I wonder what voltages (if any) are induced by the cape, antennas, leads and network before the power is applied.
  Stu
  
  After posting that it occurred to what may be needed is disconnecting the negative and positive rails from the BBG fully to reset, not just the cycling 5V supply.
  I wonder if the induced voltages from atmospheric antennas hold the BBG in a locked condition against station ground/ supply 0V and just bringing up the supply rail does not clear it?
• G8JNJ

  November 2018

  I've also had this problem and wondered about building a simple power up 'pulse' circuit to strap across the KiWi reset button.
  
  We had a similar issue where I used to work, with a product in a test rig that required a quick button press to turn it on a few seconds after the power had been applied and had stabilised. Several solutions were proposed involving external microcontrollers, but in the end a simple 4000 series CMOS RC timing circuit with associated diode logic driving an opto isolator did the trick quite nicely.
  
  Regards,
  
  Martin - G8JNJ
• UR5VIB

  November 2018

  > I have a remote system for switching the Kiwi's powersupply off/on.
  > I tried that several times. It never comes online now.
  
  Similarly to yours, in my setup I have an external watchdog device that pings KiwiSDR every five minutes and if there is no reply (several consecutive replies) to that, the relay breaks +5V wire to the Kiwi barrel power connector for 15 seconds, then sets it back.
  
  Also, I'm going to connect the BeagleBone serial console to some external device that can be accessed over an uplink. By looking at serial console output you can see what's going on during the BeagleBone boot procedure.
• n6gn

  November 2018

  A big capacitor only slows down transitions. But the whole thing operates in the context of the Beagle's brown-out/reset environment, as I understand it. As power comes up, at some point that algorithm decides to try to make it all run. What we don't want to have happen is for the voltage sag of starting is such that it can't complete the process because V drops below the hysteresis between 'run' and 'brown-out'.
  I don't know what those levels are or whether they are big enough, but I thought that making a stiffer supply might help. It might not. We probably need to understand this environment better. The problem is clearly a common one.
  
  Glenn n6gn
• WA2ZKD

  November 2018

  I have a ballsy linear PSU on a UPS, never see any problems on either kiwi
• n6gn

  November 2018

  This behavior persists when run from a battery... It does not occur with other KiwiSDR boards run from the same power sources.
• G8JNJ

  November 2018

  Whilst messing about with my Broken Beagles I noticed that the DC volts at the power supply had to be at around 5.4V to provide 5v at the Beagle.
  
  On first inspection this seems to be associated with the RF filter on the DC input of the KiWi board, which is causing a voltage drop to occur (before it is passed to the Beagle) when the KiWi is drawing it's approx 600mA average running current.
  
  Further experimentation is required to find the optimum input voltage in order to ensure a clean start up.
  
  I'd suggest checking the DC volts on the Beagle header when it's up and running, to see what's actually being applied to the board.
  
  Regards,
  
  Martin - G8JNJ

  
• Powernumpty

  November 2018

  I know they lock up at about 6.7V running which is quite a bit higher than the specs would suggest. (and recovered)
  
  Please note: This overvoltage test was done by a certified idiot, please do not attempt to reproduce the results at home.