Debian booting problem

[krystian.s]

Hi,
today I received my nanoPi 2 board. I looks cool but I have some problem with booting debian image on it.
I download image from Google Drive with support to AP mode.
Next I prepare microSD card and write image using OSX system and also tried to do that on Windows.
Then connect TV and plug in to supply. On screen I see FriendlyARM logo which disappear, then blue led is blinking and stays blue. After that nothing is done. I can't see anything on TV and also can' discover WiFi for nanopi2 ap.

In attachment I include logs from UART

[FATechsupport]

Thank you very much for choosing our boards and sorry for causing you this trouble. Please check my comments:

1. make sure your power supply is 5V/2A and it is stable

2. could you please help check the resistance of R452 (please check the attached file): is it /100k/1%? If yes could you find a new resistor whose resistance is /115K/1% and replace that R452 with this new resistor? And then try again.

NanoPi2-ResistorR452I.png

If the above two still don't work for you please check the following point 3.

3. Please try an alternative TF card. We listed some on our wiki site. Quite a few customers used Kingston cards and didn't succeed. We haven't found the root cause yet.

[FATechsupport]

Here is the resistor's position. It didn't seem to be posted in the previous post

NanoPi2's R452

NanoPi2-ResistorR452I.png (127.12 KiB) Viewed 20786 times

[krystian.s]

I checked value of this resistor and my multimeter gives me 50k. But I measured it in board so value can be affected by other elements. Are this two resistors connected in parallel?

[FATechsupport]

Hi Krystian, you only need to change R452. In the second picture there is only one resistor marked in red. That is the one you just need to change.

[krystian.s]

Today I tried to use new sd card - SanDisk Ultra 10 class 64GB and also booting failed.
While booting I measured VDD_5V and I have 4,86V it is correct value? Samsung chip is very hot

In my market I can't find resistor with 115k and 1% tolerance value. I can buy 120k resistor. It will be good resistor for replacing?

[FATechsupport]

I would suggest you use our recommended resistor

[davef]

Alternative to increasing R452 is to decrease the value of R453. However, the most helpful piece of information would be ... what is the acceptable voltage limits on VDD_ARM_AP.

Then measure that voltage before deciding to make any changes. You will need a fairly good-quality DVM. If you decide to change the resistor make sure you have done a really good job or your 1.2V supply just might go up to 5V. If you don't solder your new R1 properly and it appears as an open circuit ... just try plugging say 10M Ohms into the formulae found beside this supply.

Changing 0402 chips (if it is, I haven't got the hardware yet) requires a fair bit of skill.

For the over-heating problem refer to the posting on the other FA forum:
http://www.friendlyarm.net/forum/topic/6521

[krystian.s]

I checked nanopi2 circuit and found that VDD_ARM_AP power signal represents VDDI_ARM on S5P4418 processor.
On 87 page of Samsung datasheet we can find that description for VDDI_ARM signal: "1.0V to 1.3V for CPU".
According to nanopi2 circuit we can see that VDD_ARM_AP is obtained by MP2143DJ step down switcher. On board we have the same resistors R452 and R453 with 100kΩ and 1% tolerance. Using equation we have 1,2V.
In MP2143DJ datasheet I found this attention: "The feedback resistor R1 must account for both stability and dynamic response, and thus can not be too large or too small. Choose an R1 value between 120kΩ and 200kΩ." So maybe value of R1 is too small and has some affects for this power supply?

I will replace this two resistors by 150k and 130k, combination of this resistors will give me about 1,29V.
It will be challenge for me because this resistors are very small

On Monday probably I will check current value of this power supply

[davef]

On page 43-2 you will find 1.14V min 1.2 V 1.365V max. Also the MP2143DJ has a tolerance of +/-1.5% within the voltage range of 2.5V to 5.5V PLUS a temperature variation of +/-2% over -40 to +85C (quite impressive). So, there could be about +/-2.5% over typical operating conditions.

The 120K you suggested brings you to the 1.365V max (taking tolerances into account), hence the suggested 115K by FAtech support.

I would be more inclined to put a large value in parallel with R543 as this would be a safer approach.

I doubt if R1 being 100K would make a measurable difference. Was a 130K easier to find?

Good luck doing this at home. At work I have a SMD re-work station and microscope for placing 0402 chips.

[davef]

FAtechsupport,

Just out interest which type of errors seem to be caused by a low VDD_ARM_AP? (As seen from his log.txt file)

1) hw-breakpoint: Debug register access (0xee113e95) caused undefined instruction on CPU 1

2) Fail, noack i2c.2 addr [0x70] data[0x70], trans[ 0: 1]

3) Unable to handle kernel NULL pointer dereference at virtual address 000001f1

[halherta]

I also had the Debian booting problem. I'm using an official 16GB Sandisk ultra microSD. I board kept rebooting even with a 5V 2A adapter.

I overcame the problem by zeroing out the microSD before writing the image (.img) file. Under Linux (running on the PC) this can be done with:

Code: Select all

 sudo dd if=/dev/zero of=/dev/sdX bs=1M 

where '/dev/sdX' is the drive label.

Once the microSD was zeroed out, wrote the image to it and the random reboots disappeared. Hope someone finds this helpful.

[FATechsupport]

R452 influences VDD_ARM. After that change (from /100k/1% to /115K/1%) the voltage jumps from 1.2V to 1.29V.

[krystian.s]

Hi,
I measure current voltage of VDD_ARM and I have 1,059V so according to datasheet it is too small. I order 200k (for R452) and 180k (for R453) resistors which will give me about 1,266V.

[davef]

Your measurement indicates a precise value (1.059V) I wonder how accurate it is? Which DVM are you using? Might be a good idea to measure the other voltages ... if they are closer to the stated values then maybe you can trust this reading. There are four more supplies using the same device.

Your reading indicates that either the load on this chip is affecting its output voltage or the IC is not what the datasheet claims it is.

I would also suggest that you put a DSO on this supply output and see if there are transients when the unit is booting up and also any excessive ripple on this supply. VDD_ARM_AP has quite a bit of filtering on it.

Out of interest did you try:

sudo dd if-/dev/zero of=/dev/sdx bs=1M as suggested by halherta?

[FATechsupport]

R452 influences VDD_ARM. After that change (from /100k/1% to /115K/1%) the voltage jumps from 1.2V to 1.29V.

One more detail: 1.2V doesn't provide enough power in some transient condition so we change the resistor

[krystian.s]

Hi again
Today I received 115k resistors. I replaced R452 100k resistor with 115k and still can't boot device.
Now I have on my multimeter 1,135V instead 1,059V. Also now board restarts itself after 5sek waiting (have this value from logs).
I will check with 120k resistor

[FATechsupport]

Hi again
Today I received 115k resistors. I replaced R452 100k resistor with 115k and still can't boot device.
Now I have on my multimeter 1,135V instead 1,059V. Also now board restarts itself after 5sek waiting (have this value from logs).
I will check with 120k resistor

Hi Krystian could you help us collect its system messages and post them here? We will take a look

[davef]

After a month or two of proper operation the "Debian booting problem", when using the HDMI display, has returned.

After a few seconds one gets a boot message that says it is going to reboot in 5 seconds. Looking further up the boot messages one sees this :

Code: Select all

[    0.000000] Kernel command line: console=ttyAMA0,115200n8 root=/dev/mmcblk0p2
 rootfstype=ext4 init=/sbin/init systemd.show_status=false g_ether.host_addr=82:
cf:ce:fa:44:18 lcd=HDMI720P60

�[    2.000000] VFS: Cannot open root device "mmcblk0p2" or unknown-block(0,0): 
error -6                                                                        
[    2.008000] Please append a correct "root=" boot option; here are the availab
le partitions:                                                                  
[    2.016000] Kernel panic - not syncing: VFS: Unable to mount root fs on unkno
wn-block(0,0)

Yet, when using the LCD display it boots properly.

This time I decided to watch VCC_ARM_AP during boot-up (with the HDMI display) and don't see any sag or transients on this supply output.

Both of my NanoPi2 had the factory modified R452 and VDD_ARM_AP read 1.28V. I am using a recommended microSD card.

Have there been any subsequent modifications to the NanoPi2 to eliminate this problem?

[FATechsupport]

Hi Dave,

No we haven't made any subsequent changes. Could you please post your complete system log starting from system boot?

[davef]

It is not really a .zip file. Rename to HDMI boot problem.txt

I only showed the first error message, which seemed to indicate a microSD card problem. (mmcblk0p2)

[FATechsupport]

Hi Dave, please get the latest Debian image files. In our latest image we added "rootwait" and fixed this issue.

[davef]

Thank you. I won't be able to try it out until Next Tuesday.

[davef]

I still see 27th Feb (as date created) for both Debian files on Google Drive. Am I looking in the right place and at the right files?

[FATechsupport]

Yes Dave that is the correct place

[quchi]

Looks like i'm having a different booting problem
Random boot - never a dull moment - never twice the same.. but some patterns
Most of the time as read-only
Always the 1st time after i created the sd image - Xorg cannot start
Some other times i get the desktop but nothing on it
- just black
However the menu works and i can bring up the terminal - still read only
Rarely maybe 1 in 5 i get a desktop and can connected to wifi
Very frustrating
The goog drive download files showing Feb 25 not 27th like mentioned above
Got the link from wiki
http://wiki.friendlyarm.com/wiki/index.php/NanoPi_2

[FATechsupport]

Looks like i'm having a different booting problem
Random boot - never a dull moment - never twice the same.. but some patterns
Most of the time as read-only
Always the 1st time after i created the sd image - Xorg cannot start
Some other times i get the desktop but nothing on it
- just black
However the menu works and i can bring up the terminal - still read only
Rarely maybe 1 in 5 i get a desktop and can connected to wifi
Very frustrating
The goog drive download files showing Feb 25 not 27th like mentioned above
Got the link from wiki
http://wiki.friendlyarm.com/wiki/index.php/NanoPi_2

Hi is this issue the one you showed me in your IMG_4959 and IMG_4960 pictures?

[krystian.s]

Hi again,
I don't write here quite long time from my last post.
In that time I received new board with fixed resistor. Yesterday I have some time to check if this new board boots correctly.

Now for testing I used 64GB SanDisk card with 10Class and Debian image with AP mode (Feb 27). For power supply I used custom DC/DC converter with max current of 3A and very stable 5V (measured: 5,02V with load).

When I plug in HDMI cable with TV and after that try to boot - board can't setup and run correctly, It constantly reboots (I see some logs on TV screen and also Friendly ARM logo) after displaying it runs again and again...
So I decided to unplug HDMI cable from board. After that I see that board starts working. I see it in my WiFi networks. Also when board was working I plug in it to my TV and see Debian screen.

I try to repeat this behavior and see that it is repeatable (not always but mostly).
Summarize: If I want to boot Debian properly I need to unplug HDMI cable from board (I don't test If I should unplug HDMI cable from board or from TV) and after successful booting process I plug my TV.

Friendly ARM - Could you fix that issue?