Purple Randomly rebooting

Comments

11 comments

  • Avatar
    Firewalla

    The reboot reason is likely /var/log/syslog

    are you using router mode?

    0
    Comment actions Permalink
  • Avatar
    Firewalla

    I created a ticket for you, we can check the system out 

    0
    Comment actions Permalink
  • Avatar
    Stephen Ball

    If anyone comes across this as they have the problem, after back and fourth with Firewalla supporter we tracked it down to a faulty power adapter. Turning on wifi obviously makes it draw more power, switched with another 5V/3A power adapter and it works perfectly.

    Thanks!

    0
    Comment actions Permalink
  • Avatar
    Dave Taylor

    I've experienced a similar problem with the same solution suggested by the support folks, still waiting to see if it reboots again but so far so good.  It's really odd because I used a 65W USB-C PD adapter which should be able to provide more than enough power, I've run my laptop off that and that draws far more than the Firewalla does.  Once I switched it from the PD port to a standard USB port the reboots seem to have stopped, at least so far.

    Is anyone else powering their Purple from a USB-C PD port rather than USB A?  If so, have you experienced reboots?

    0
    Comment actions Permalink
  • Avatar
    Dave Taylor

    Oh, and as a followup to that, what happens if you power it from a Mac USB-C port?  They're especially picky about what they will and won't allow.

    0
    Comment actions Permalink
  • Avatar
    Stephen Ball

    In the end, I used the official Raspberry Pi 4 plug and it's working fine, although I also ended up turning off the wifi as it is even more short-ranged than I realised so I got a cheap AP and added it to a VLAN 😂

     

     

    0
    Comment actions Permalink
  • Avatar
    Dave Taylor

    Following up on my earlier post, it's now been ~2 weeks with no more reboots after shifting from a USB-C -> USB-C to USB-A -> USB-C cable on the same power supply.  Given that the only change was moving from a smart USB-C cable to a dumb USB-A one, my suspicion is that the Purple may not have the correct USB-C circuitry set up, e.g. the CC resistors configured incorrectly.  Thus my question about whether anyone else is running it from a USB-C rather than USB-A power source, and in particular a Mac which is quite picky about USB-C compliance.

    0
    Comment actions Permalink
  • Avatar
    Dave Taylor

    Just noticed that my earlier post, about a possible hardware bug in the Purple being the cause of reboots, has been marked "pending approval" for five months now, so I assume it'll never appear.  This is a repost:

    It's now been several months with no more reboots after shifting from a USB-C to USB-A port on the same power supply.  The Purple may not have the correct USB-C circuitry set up, for example the CC resistors configured incorrectly.

    So: Has anyone powered it from a USB-C power source, and in particular a Mac which is more picky about USB-C compliance?

     

    0
    Comment actions Permalink
  • Avatar
    Firewalla

    Some how the system (zendesk) marked your previous post as spam. I just released it. 

    As of USB-C to USB-C, what is on the other side of USB-C? The reboot usually means, the powering side not able to supply 3Amps of current. 

    0
    Comment actions Permalink
  • Avatar
    Dave Taylor

    Thanks.  So it's the same power brick in both cases, either a USB-A port with 2.4A current (12W) or a USB-C PD port with 65W power delivery.  This is why I don't think it's the power brick, because it's running fine off a 12W port but not off a 65W port on the same device.

    I've run into this before on another embedded device, what had happened was that they hadn't set the USB-C CC (Configuration Channel) resistors correctly on the device, so the device was telling the power source not to supply the full USB-C power level and it was crashing and rebooting under load.  When it was fed from a USB-A source it just pushed out the full 2.4A regardless and it worked fine.

    This is why I suggested powering it off a Mac as a test, they're very finicky about compliance with the spec and so the fact that you can't power it off a Mac would be a quick indication of a problem, e.g. if the pulldown resistors have been omitted entirely which was the case in this particular device.

    0
    Comment actions Permalink
  • Avatar
    Dave Taylor

    As a followup comment, the Raspberry Pi folks found their own way of doing it wrong, https://hackaday.com/2019/07/16/exploring-the-raspberry-pi-4-usb-c-issue-in-depth/, although their way is slightly different than in the embedded device I'm referring to which simply omitted the CC resistors entirely.  Point is, if powering it off a USB-A port works but USB-C doesn't, even though the USB-C port can deliver quite a bit more power than the USB-A one can, then the first thing to look for is a hardware issue in the device's USB-C handling.

    0
    Comment actions Permalink

Please sign in to leave a comment.