Beaglebone sometimes doesn't start after a power loss

Hello,

I’ve noticed that the beaglebone black sometimes does not startup after a short power loss, even when pressing the power button, it does not start.

To reproduce with a regulated power supply (tested with BBB rev B and C):

  1. Output of power supply 5V
  2. connect power supply to BBB
  3. BBB starts up normally
  4. down-regulate the output slowly
  5. at 3.5 V, the beaglebone turns off
  6. go back to 5V
  7. press the power button, the beaglebone won’t startup
  8. go above 5.1V and press the power button => now the beaglebone starts

I’ve read the datasheet of the PMIC (TPS65217C) but I cannot find out why the BBB starts at 5.1V but not at 5V, has anyone an idea?

Many thanks,
Rico

Sorry to push… Gerald, maybe you have an idea?

Sorry to push… Gerald, maybe you have an idea?

Hello,

I’ve noticed that the beaglebone black sometimes does not startup after a short power loss, even when pressing the power button, it does not start.

To reproduce with a regulated power supply (tested with BBB rev B and C):

  1. Output of power supply 5V
  2. connect power supply to BBB
  3. BBB starts up normally
  4. down-regulate the output slowly
  5. at 3.5 V, the beaglebone turns off
  6. go back to 5V
  7. press the power button, the beaglebone won’t startup
  8. go above 5.1V and press the power button => now the beaglebone starts

Are you measuring the voltage at the BBB terminals or are you reading the voltage from your power supply? Do you use thin wires to connect from your power supply to the BBB? My thinking is there is a voltage drop across the wires of 100mV or more.

Regards,
John

John, this happens to me too, but powering via USB. Once in w blue moon, there seems to be some sort of race condition, or something where the BBB is not getting enough power.

So in order to thwart this I’ve stopped using shutdown now -r, and instead use shutdown now -h. Then manually remove / reapply the USB cable.

Unplug it and plug it back in. It was designed to work that way. Temporary power loss can confuse the PMIC and put it in an bad state.

Gerald

@ John:
Thanks for your answer. I measured with a multimeter directly at the BBB. The BBB normally even starts with only 4.5V so its definitely something wrong with the PMIC.

@Gerald:
Thank you. The problem is, we are using the BBB in an industrial product that has to run 24/7. Replugging the device is therefore no option. Powering the BBB with at least 5.1V seems to resolve the problem but I wondered if there is a reason for that.

Well, I didn’t design the board for that application. There is a way to take care of that if you are willing to fix it. It is not something that I will put in the design just for your application.

Gerald

Hello Gerald,

What way would that be to fix it? I thoght the problem is the PMIC itself and not the BBB design.

Add a battery to the board to keep it up long enough to shutdown if the failure is long or to not be impacted by momentary losses of power, which can also corrupt the filesystem on either the eMMC or the SD card.

Gerald

Hi Rico,

Any reason that you can’t use a cape to solve the problem? The PowerCape will prevent brown-outs and transients from bringing the BB down. Your application/script can monitor DC power good and battery voltage/current and then do a clean shutdown if necessary. The cape can also restart the system when DC power returns. It’s my cape so feel free to ping me with any questions about it…

-Ron

That is a great suggestion.

Gerald