Can't boot from SD card anymore

My board worked for 1 day. Now does not boot anymore.
Am I doing something wrong?

Apart from that, I never got the boot login screen. I use FTDI serial interface and could see boot log just fine, but never got to the first login screen.

Now it hangs at:

RDA8810 Boot_ROM V1.6
HW_CFG: 0286
SW_CFG: 0000
Load EMMC
Open eMMC

Is my board defective?

Thanks

Setting the boot jumper to internal memory gives me this output:

RDA8810 Boot_ROM V1.6
HW_CFG: 0282
SW_CFG: 0000
Load NAND
Nand Init, type = 06 (bus 8bit)
Nand Init Done
Loading Image from NAND … Done
Check Image Magic Error FFFFFFFF
Load NAND Fail
Load SD
Open SD card
mcd_Open
Card is SDHC
Inserted Card is a SD card
mcd_Open Done
Loading image from SD card …Done
Check Image … Done
Run …
Init Clock …
Clock config ver: 1.0
PMU vbuck1 = 9, vbuck3 = 10
setup ddr vtt to 1

… and hangs there.

No, the board is probably not defective. I turns out that on the 2nd boot, it tries to load up the alsa-ctrl program, and THAT is what’s defective.

Take the SD card, mount it on Linux or Mac, go into /usr/bin and remove alsa-ctl. Then sync and unmount the SD card, replace it in the OrangePi-i96 and it will probably boot up.

1 Like

Interesting. I’ll do that tomorrow when I get back home. Thanks.

THANK YOU so much !!!

I’ve tried both the Debian Stretch and Ubuntu Xenial server images that are available for the Opioid-i96 and having issues with startup. The alsa-ctl stuff appears to be removed but I failed to be able to log in using the ‘orangepi’ / ‘root’ and ‘orangepi’ password combos. I have observed resizing of the SDCard partition after the initial boot. However failing at the login prompt repeatedly seem to lead to more flakiness and I end up needing to start from scratch with another burn of the image to the SDCard. There is a long delay after password entry and then prompting for login again! Using:

OrangePi_i96_debian_stretch_server_linux3.10.62_v0.0.4.img or
OrangePi_i96_ubuntu_xenial_server_linux3.10.62_v0.0.4.img

Is there a “known good” recipe for getting to a stable state? I have years of experience setting up and using Armbian on other Orange Pi boards so am normally savvy enough to work through obstacles. Getting a “leg up” toward being able to log in to the system reliably would be greatly appreciated.

Is the initial gibberish on the serial port is actually useful boot-up info just at a different baud rate? I’m away from home, where my digital storage oscilloscope is, but intend on trying to figure that one out when I get back as I’m currently “blind” as to what is failing.

Paul,

The initial stream of text at boot time is coming at you a 921600 baud. One my boards, I get readable text from the U-boot and then the kernel and then the shell.

I just now took the OrangePi_i96_ubuntu_xenial_server_linux3.10.62_v0.0.4.img image, copied it to a SD card (dd if=… of=…) and then mounted the same flash disk, erased all /usr/bin/alsa* and /usr/sbin/alsa* and then inserted the SD card into my orangepi-i96, and it booted right up.

I had a little trouble with the debian image, I’m still chasing why…but I thought I’d let you know that Ubuntu seems ok once you clear out the alsa* sound files.

The debian image works fine as well. As usual, check the i96 board is getting adequate current, and that you’re not using a crappy SD card (like I was).

I really appreciate your response to my post. It was very helpful and after eliminating all possibility that the problem was due to the power supplies tried or various SDCards used, I found that it was the serial terminal programs that I had tried that were to blame.
You are correct that the baud rate for ALL of the boot up messages is at 921,600 baud (8N1). Somehow both ‘minicom’ and ‘screen’ are failing to display UBoot and Kernel output properly and it’s just gibberish. Perhaps it’s overrunning the rate that they can take it in. I will note that I am using one of the new MacBook Pro M1Pro’s with those apps installed using “Homebrew”. Fortunately, for whatever reason, ‘picocom’ does work. Oh, and I tried various USB-TTL Serial adapters with different chipsets. They all seem to do the same.
Now that I have it working, the frustrations with the WiFi setup corrupting itself is in front of me. So far Debian seems to be working better than Ubuntu and I actually had it stable enough to do a full ‘apt update’ + ‘apt upgrade’ while it was working. But the kernel messages related to wifi, unreliable response to ‘ping’ requests from another machine, and other nuisances make me wonder if my time is so invaluable that it’s worth pursuing this board. It seems like it has broken the will of others similarly. If the Armbian folks were working on the Orange Pi i96 then there might be more reason for hope. Over the years I’ve run many OPi boards using various versions of Armbian and it’s been quite stable.
Thanks again for your help. The certainty of what you wrote enabled my persistence to find the variable that was out of whack!
For reference to others – this worked for me: ‘picocom -b 921600 /dev/tty.usbserial-10’