CSC Armbian for RK3318/RK3328 TV box boards


jock
 Share

22 22

Recommended Posts

4 minutes ago, Huafu said:

@jockI've tried all combination of the uncovered copper without success. Even tried and with each of the 3 spots...

I also tried UART, nothing just some weird chars right when I touch tx and Rx to their spots, but nothing when I give power to the board, with or without the reset button pressed...

Thanks for tour time and if you can't take a deeper look nevermind.

 

Tho if you have a solution related to the overscan of the first board I'll take it 🙂

Well I would try to guess what's wrong with that.

I just checked the ddrbin and it the very exact same ddrbin that is currently shipped along the current images, so all the things I said in the previous post should not apply to your case.

 

You must see something intelligible from the serial, be sure to put it at 1,500,000 bps (1.5 mbps), otherwise you only get garbage.

 

About the overscan, I never faced such issue with the board. Had to fix that on my monitor/TVs usually setting the right option in the menu (some call it "16:9", some "Just Scan", "PC mode" or other fantasy names...), but never had to handle it at board level.

It happened to me that some older TVs just don't have the feature to remove overscan :unsure:

Link to post
Share on other sites

Donate and support the project!

@jock For my RK3318 box, everything works well but recently there is always error message during starting up.

The message is "[FAILED] Failed to start Resets System Activity Logs.",

 

Here is the detail status of this service:

 

root@mybox:~# systemctl status sysstat.service
● sysstat.service - Resets System Activity Logs
     Loaded: loaded (/lib/systemd/system/sysstat.service; enabled; vendor preset: enabled)
     Active: failed (Result: exit-code) since Thu 2021-11-25 12:10:54 +07; 26min ago
       Docs: man:sa1(8)
             man:sadc(8)
             man:sar(1)
    Process: 698 ExecStart=/usr/lib/sysstat/sa1 --boot (code=exited, status=2)
   Main PID: 698 (code=exited, status=2)
        CPU: 38ms

Warning: journal has been rotated since unit was started, output may be incomplete.

 

So I just confused how to fix this one.

 

Thank you.

IMG_4873.jpg

Link to post
Share on other sites

On 11/16/2021 at 11:02 AM, fabiobassa said:

Anyway, just to answer , in my personal study case I have done on it:
voip pbx ( asterisk)
network nas
network firewall
pihole dns firewall
openvpv, wireguard and tinc vpn

home assistant
volumio
special ip cam monitoring and data acquisition

 

Hi Fabio, I would be interested to know how did you make Volumio work on a RK3328 Android box? I have a Bqeel U1 Max (RK3328) which I don't use and repurposing it for Volumio would be awesome.

Link to post
Share on other sites

Good morning, thank you for having accepted me in the forum. I'm new to these adventures I installed Armbian on my HK1 Max Circle box with the RK3318 chip everything went fine but I don't have a browser installed, how can I install it? Thanks

Edited by mibi
Link to post
Share on other sites

@kindgott 

this 3ad is more hardware operating-system oriented than user space applications oriented.

It cames by itself that if we would only try all the universe of application avaible for linux we loose the original  point of view :

working or not working on the most boards on the market


To answer to yor question , btw , I installed 2 years go  and I did step by step, not using a precompiled image. Now I see that many solutions are precompiled images and I don't like this approach

I guess google as alway can help solving the question at the actual state of art of volumio


 

Link to post
Share on other sites

Hello @jock, thanks for helping again.

  

On 11/24/2021 at 7:29 PM, jock said:

You must see something intelligible from the serial, be sure to put it at 1,500,000 bps (1.5 mbps), otherwise you only get garbage.

What is weird is that the red led turns on directly when I give power to the board and nothing else happens, the red light remains on, period.

 

Thanks for the baud rate, I didn't have the good setting, but I might also plug it the wrong way.

I've tried with and without the GND. Either way I do not use the +5V as I don't know where to connect it. I'm only using TX and RX, and I've tried with and without the GND but I guess I don't need it if I'm not using the +5V.

When I make the contact between the UART cables and the board, I got garbage, then I give power to the board and nothing is shown anymore. Is this the right way to proceed?

 

image.png.9b1d75293f7222ac2b0d67b3767c88c2.png

 

On 11/24/2021 at 7:29 PM, jock said:

About the overscan, I never faced such issue with the board. Had to fix that on my monitor/TVs usually setting the right option in the menu (some call it "16:9", some "Just Scan", "PC mode" or other fantasy names...), but never had to handle it at board level.

It happened to me that some older TVs just don't have the feature to remove overscan :unsure:

Then I have an old TV :)

 

UPDATE: related to screen resolution, I've discovered that my big TV screen is recognized thru EDID as a 7' LCD, which is totally wrong. Manufacturer is good, but EDID says:

 

  Video Data Block:
    VIC   4:  1280x720    60.000 Hz  16:9    45.000 kHz  74.250 MHz (native)
    VIC  19:  1280x720    50.000 Hz  16:9    37.500 kHz  74.250 MHz
    VIC   5:  1920x1080i  60.000 Hz  16:9    33.750 kHz  74.250 MHz
    VIC  20:  1920x1080i  50.000 Hz  16:9    28.125 kHz  74.250 MHz
    VIC   3:   720x480    59.940 Hz  16:9    31.469 kHz  27.000 MHz
    VIC  18:   720x576    50.000 Hz  16:9    31.250 kHz  27.000 MHz

Tho the specs of the monitor says its max resolution is 1360x768.

 

I'll try to find the correct EDID info from linux hardware and rewrite it using https://github.com/linuxhw/write-edid

Edited by Huafu
wrong screen resolution
Link to post
Share on other sites

@Huafu

The ground (GND) pin is the most important pin, you must absolutely connect it if you want a reliable result, otherwise you can get varying results starting from occasional garbage, total garbage or even adapter breakage.

If you have a common regular USB serial adapter, +5V is not needed at all and must be kept disconnected.

 

All the three pins (GND, TX and RX) must be connected before giving power to the board and possibly with the serial adapter disconnected from the computer too.

Once the hardware setup is done, you can connect the serial adapter to the computer and configure minicom with the right parameters: usually it is sufficient to keep defaults and change baud rate to 1.5Mbps.

 

Finally you can give power to the board.

 

Link to post
Share on other sites

Hi, after having successfully had my H96Max+ working for a few days now with a single major hiccup I would like to inquire about three items.

 

  • The single issue that I encountered was a total freeze during intense activity on the SD card while copying files transferred through wifi.  This copying is something that I could have done in a much quicker way but that I purposely did this way to stress the machine with a 2+ hours of continuous activity.  However, because I had this single freeze, that did not happen again, it is difficult for me to say if this was due to some OS or hardware instability or maybe some external factor (e.g., a power fluctuation).  It would be interesting to me to know if someone else experienced similar issues.
  • I would like to know if RK3328 supports some form of hardware watchdog timer to be used with the watchdog daemon.
  • I understand that once the machine is installed (thanks for the great work of jock and possibly others), maintaining the machine up to date and secure should be relatively easy, since most of the OS can be updated through the regular armbian repos. In my understanding the notable exceptions are the kernel packages at https://users.armbian.com/jock/rk3318/upgrade/ namely those that had to be put on hold wrt apt (linux-image-edge-rockchip64 linux-headers-edge-rockchip64 linux-dtb-edge-rockchip64). If I understand correctly, these can only get updated thanks to jock's effort because rk3318 is not part of armbian (yet). Having these packages depending on a single person best effort means on one hand that gratefulness to Jock is huge but on the other hand may also represent a potential issue. I thus wonder how difficult is it to re-build them by https://github.com/paolosabatino/armbian-build/tree/rk3318 and the armbian kernel tree, how much does the rk3318 diverge from the mainline kernel, how easily/frequently do the rk3318 patches break against newer mainline kernels and how difficult would it be to recover from a bad kernel (in case one tries to build and deploy a kernel for learning how to do it and something goes wrong).

 

Thanks for any help!

Edited by callegar
Improving clarity
Link to post
Share on other sites

@callegar

1) To me happen quite often  on 322x ( not ried on 3318 thought) , both on sd and on internal emmc/nand. It look like the ethernet o wifi is not adapting very well the speed  of the net related to the speed of writing/ reading from storage. I can confirm that this happened to me also and exactly when i tried to stress the board as you did.

2) waiting for better @jock answer

3) yes you are right , at the actual state kernel and related things are not part of official armbian neither debian , but also all of us should consider that this is very esperimental case of study so the best is auto-compile since is not very difficult.
The difficult part is patch things as soon as problems are dicovered

 

Link to post
Share on other sites

i use Armbian 21.11 - Debian Bullseye minimal - mainline kernel 5.15.2 from Jock

and i notice that

rsyslogd: file '/var/log/syslog'[7] write error - see
 https://www.rsyslog.com/solving-rsyslog-write-errors/ for help OS error: No spa
ce left on device [v8.2102.0 try https://www.rsyslog.com/e/2027 ]

log is in ram, and it's small , $ df :

/dev/zram1         49560    9120     36856  20% /var/log

but there is a cron job who clean var/log

Link to post
Share on other sites

@dam74The issue with rsyslogd is a symptom. On my system, I think I managed fixing the issue by setting also the SystemMaxFileSize option in systemd journald.  Wonder if on a system with some good 32 or 64GB of resident storage having journal log meant to be permanent on zram is the right thing, though, given that journald already has volatile storage on /run.
 

Edited by callegar
Link to post
Share on other sites

Hi @jock
I come back a bit on the led-conf3 that you provided me and which seemed to be correct however:
- There are physically on the front of my box 2 LEDs (1 red and 1 blue)
- when I run  " sudo du -a /sys | grep led | grep trigger "

      /sys/kernel/tracing/events/libata/ata_qc_complete_failed/trigger
0       /sys/kernel/tracing/events/dma_fence/dma_fence_signaled/trigger
0       /sys/kernel/tracing/events/kyber/kyber_throttled/trigger
0       /sys/kernel/tracing/events/btrfs/btrfs_failed_cluster_setup/trigger
0       /sys/kernel/tracing/events/ext4/ext4_journalled_write_end/trigger
0       /sys/kernel/tracing/events/ext4/ext4_journalled_invalidatepage/trigger
0       /sys/kernel/tracing/events/xdp/mem_return_failed/trigger
0       /sys/kernel/debug/tracing/events/libata/ata_qc_complete_failed/trigger
0       /sys/kernel/debug/tracing/events/dma_fence/dma_fence_signaled/trigger
0       /sys/kernel/debug/tracing/events/kyber/kyber_throttled/trigger
0       /sys/kernel/debug/tracing/events/btrfs/btrfs_failed_cluster_setup/trigger
0       /sys/kernel/debug/tracing/events/ext4/ext4_journalled_write_end/trigger
0       /sys/kernel/debug/tracing/events/ext4/ext4_journalled_invalidatepage/trigger
0       /sys/kernel/debug/tracing/events/xdp/mem_return_failed/trigger
0       /sys/devices/platform/gpio-leds/leds/ir/trigger
0       /sys/devices/platform/gpio-leds/leds/working/trigger
0       /sys/devices/platform/gpio-leds/leds/net/trigger
0       /sys/firmware/devicetree/base/gpio-leds/ir/linux,default-trigger
0       /sys/firmware/devicetree/base/gpio-leds/working/linux,default-trigger
0       /sys/firmware/devicetree/base/gpio-leds/net/linux,default-trigger

I therefore tried to modify the trigger on

> /sys/devices/platform/gpio-leds/leds/

Where can we find 3 leds? Either .. So I make some modifications on the 3 leds detected by the system and I find the red front led of my box (under the "net" value) I modify the value of the trigger on heartbeat and Hop the red led of my box starts flashing (cool ..!)
Unfortunately, it is impossible to modify the light state of the blue led which lights up when the box starts up, I have unsuccessfully modified the values of the triggers of the "ir" and "working" leds but nothing changes, currently the triggers in questions are on "none" and yet I still have the blue led on ..
Is there anything else to do?
Thanks in advance..

 

 

Link to post
Share on other sites

 Share

22 22