Hikey lemaker bricked or some other way to recover? hsi_thermal T > 50


#1

After some fairly intense abuse, it appears I bricked the board, but it’s unclear. It’s stuck on the penguin default boot screen with an error that says:

hsi_thermal f7030700.tsensor: THERMAL ALARM: T > 50
of_dma_request_slave_channel: dma-names property of node ‘/smb/uart@f7113000’ missing or empty

Someone please confirm if there is a way around flashing it. Sadly I didn’t back anything up and all the work is sitting on the ROM. It would be even sadder if it was fried, but seems unlikely.

I had an iPhone 6S+ in one of the USB ports, and it was powered by a 12V @ 2A-ish power, with non-stop compiling, recompiling, Bluetooth / wifi radios both on and humming.

Looking around for the error on-line I see that it’s definitely temperature related. Hopefully there is an easy way to reset the board.

Thanks -


#2

Hi @noah,

If it was bootloader being corrupted, this is the link for flashing it again.

https://github.com/96boards/documentation/wiki/HiKeyGettingStarted#section-41

And none of the 96boards are design to use the USB port for charging other devices,
so please avoid connecting USB connector on 96boards for charging.

I hope your HiKey starts working again.


#3

I was fairly successful on the flashing until the last step:

sudo fastboot flash nvme nvme.img

At this step it halted with

<waiting for device>

Starting the process over, the computer is not recognizing the device at all. When I put the jumpers back to normal position, no lights are coming on.

It appears dead… Anything that can be done to fix it?


#4

Regarding charging other devices via USB:

Did someone not think that through on the specification? It would be pretty commonplace to attach a smart phone into the USB port, which automatically will start to draw 1 to 2.4 amps, depending on the device. Seems silly to not be able to attach your smart phone to the board.


#5

Having a standards compliant phone, such as an iPhone, attached to the board is unlikely to be the cause of the failure (although I no particular idea about what might be). The phone should not have been drawing more than 500mA (and at only 5v).

It is more that charging from communication-only ports is not very effective at refilling the battery!


#6

Hi @noah,

If you were able to perform until here:


$ sudo fastboot flash ptable ptable-linux-8g.img (if 8GB board) (ptable-linux-4g.img for 4GB board)
$ sudo fastboot reboot
$ sudo fastboot flash fastboot fip.bin

and it stopped at the next step:


$ sudo fastboot flash nvme nvme.img

not responding, it is really odd.

If you repeat the procedure and does it still go successfully until “sudo fastboot flash fastboot fip.bin”?

I am starting to suspect the board being damaged.


#7

Hi @noah,

For the charging iPhone6+, yes the AC adapter is 12V 2A and total is 24W so it looks it is not problem for charging iPhone and iPad which draw the current in fastmode from 1A (5W) to 2.4A (12W).

However, the 5V line on USB connector is made by DC-DC converter on the HiKey from 12V and this 12V DC- 5V DC converter is limited to maximum 2.6A. Furthermore, 3.3V is made from this limited 5V and all the components using 5V and 3.3V on the Hikey are consuming the same 5V 2.5A power.

Regular USB charging limit is 5V 0.5A which is 2.5W, so this would not be the problem.

There are USB charger cable for iPhone and that the USB connector has D- pin with 2.7V which indicate the iPhone to use fastmode charging on regular USB connector as described bellow.
http://www.epanorama.net/newepa/2010/08/18/apple-charger-secrets/

But if you try to charge iPhone with fastmode which is Apple’s extension, will draws 1A (5W) to 2.4A (12W),
then it could overload the DC-DC converter on the HiKey, since USB specification only specifies 0.5A on the USB connector,
and the HiKey’s design is completely complaint only on the USB specification and do not look like supporting fastmode on Apple.

It is safer not to charge iPhone and iPod on regular USB connector.


#8

Thanks for the thorough explanation Akira. It all makes sense and I appreciate the support.

At this point, the board does not respond at all via USB tty, nor do the LEDs light up, despite the jumper settings. It’s a nice decoration for the desk at the moment.

What else could be done to try to fix a board in this state?


#9

I would like to mention important point about the charging iPhone.

The genuine lightning cable from Apple with Hikey do not draw more than 0.5A, 5V, (2.5W) from the USB connector.
The design of HiKey is compliant with the USB specification.

I know that there are many third party’s charging only cables for or dongles that advertise which could charge faster by enabling fastmode of charging on regular USB connector.

Overloading USB connector on the Raspberry Pi is also common by searching on google.


#10

Hi @noah,

Would you try again from flushing the bootloader?
Does it success to the step where you were able to do last time with
"sudo fastboot flash fastboot fip.bin"?

If it fails the procedure, I am afraid that I am also running out of idea…


#11

My Ubuntu machine does not recognize it now on any USB port. It does not show up in dmesg as a new device. I assume it is dead, dead, dead.


#12

Sounds like it is probably dead. However just to make sure we cover absolutely all possible avenues… are all of the following statements true:

  • Nothing is connected to the USB type A sockets (or any other sockets except USB micro-B and power inlet)
  • Host PC is connected to USB micro-B on the board
  • There is a link connecting pins 1-2 of the boot selection header
  • There is a link connecting pin 3-4 of the boot selection header
  • There is no link connecting pin 5-6 of the boot selection header
  • Host PC observes no USB activity when power is removed, wait 10 seconds and reapplied

The above sequence causes the SoC to boot from a mask programmed ROM which contains a USB download program. If the USB does not enumerate then the board and/or SoC is damaged and software-only recovery will not be possible.