Orange Pi Zero 3 ethernet hardware issues 看全部

This post was finally edited by ag123 at 2023-11-28 21:04

hi

I've a Orange Pi Zero 3 2GB board.
I've tried the Debian Bookworm distributions from
http://www.orangepi.org/html/har ... ange-Pi-Zero-3.html
the image used is actually released in Aug 2023.

However, I'm encountering a lot of ethernet issues, ethernet can  *dropped* (become disabled) by simply rebooting or simply leave the board there like say for an hour (in fact less than 45 minutes) ethernet dropped and become unresponsive, the ethernet leds goes dark as well. When ethernet is dropped/disconnected, I found that it cannot be restored using software methods e.g. ip link set end0 up. The only way to restore the connection is to disconnect and re-connect the ethernet cable. Sometimes that is not enough, and I've to power cycle the board, disconnect and re-connect usb power and reboot. It is looking a lot like a hardware problem.

As I'm not sure if there is a way to solve the problem, I tried DietPi which just recently create a test image.
https://github.com/MichaIng/DietPi/issues/6594

the results of my tests are in this comment.
https://github.com/MichaIng/Diet ... ecomment-1827683215

Unfortunately, the same ethernet problems link dropped, become not responsive (led go off), cannot be restored by ip link set end0 up happen in OrangePi distributed Debian Bookworm images (August 23) and more recently the DietPi Debian Bookworm test images published and discussed in the above link.

Can Orange Pi revert about this? I've bought several boards as I wished to use this as a wifi hotspot and the problems is very disappointing as the link would drop very frequently like less than an hour. Note that otherwise, this is a very good board as I've been able to achieve exceed 100 Mbps (100 mega bits per second) throughput through the on board Wifi and Ethernet chip.

edit: currently still testing
https://github.com/MichaIng/Diet ... ecomment-1827683215

last disconnect is cause by a loose cable, fixed by adjusting the cable a little, but this isn't the only case.

  • Sofa ag123
  • 2023-11-28 22:47:07
This post was finally edited by ag123 at 2023-11-29 14:56

follow-up from previous:

down again ! this time nothing moves, ethernet led goes off

  1.    [17967.972930] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  2.    [17967.973328] br0: port 1(eth0) entered disabled state
  3.    root@DietPi:~# ip link
  4.    2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq master br0 state DOWN mode DEFAULT group default qlen 1000
  5.    root@DietPi:~# uptime
  6.    22:27:39 up  5:01,  2 users,  load average: 1.07, 1.04, 1.00
  7.    root@DietPi:~# ip link set eth0 up
  8.    root@DietPi:~# ip link
  9.    2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq master br0 state DOWN mode DEFAULT group default qlen 1000

   software means cannot restore link.  
   This time, simply adjusting the cable, disconnect / connect didn't help.
   poweroff
   disconnect / reconnect power (usb) cable. Reboot and login again over serial

  1.    root@DietPi:~# ip link
  2.    2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master br0 state UP mode DEFAULT group default qlen 1000


   Takes a full power cycle to recover. initially, the ethernet link did not come back on.
   board is warm. takes a while like a minute, then finally reconnect ethernet cable works.

   Ethernet down again less than 10 minutes after power cycle, board is warm


tests, observed 2 cases of ethernet disconnect:   
- loose cable, fixed by adjusting cable  
- ethernet lost connection after some time, possibly due to heating or some other hardware issues.  
  requires a full power cycle to recover and did not recover initially, only after a minute that it is successful.
  it is uncertain if it has anything to do with the hub/switch which it is connected to.




  • Bench ag123
  • 2023-11-29 17:20:26
This post was finally edited by ag123 at 2023-11-30 15:10

follow-up from previous:

ethernet link disconnects recovers on its own if ethernet set to run at 100Mbps (this is good news)
  1. ethtool -s eth0 speed 100


There have been several link down events at 100 Mbps.
But that this time, the link recovers on its own.
Note that the link down events are rather frequent
Thus far for one of the situation, recovery needs adjusting the ethernet cable,
possibly due to a loose ethernet cable.
  1. [ 5152.517613] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  2. [ 5152.518017] br0: port 1(eth0) entered disabled state
  3. [ 5154.565859] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  4. [ 5154.566162] br0: port 1(eth0) entered blocking state
  5. [ 5154.566192] br0: port 1(eth0) entered listening state
  6. [ 5155.589608] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  7. [ 5155.589987] br0: port 1(eth0) entered disabled state
  8. [ 5158.661851] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  9. [ 5158.662126] br0: port 1(eth0) entered blocking state
  10. [ 5158.662154] br0: port 1(eth0) entered listening state
  11. [ 5173.733061] br0: port 1(eth0) entered learning state
  12. [ 5188.837080] br0: port 1(eth0) entered forwarding state
  13. [ 5188.837122] br0: topology change detected, propagating


Disconnects is frequent, some 25 disconnects and reconnects in 4 hours.
But that so far at 100 Mbps, the kernel / system manages to recover on its own.
Wifi (UWE5622) remains up and connectable.
  1. [ 6979.333601] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  2. [ 6981.381844] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  3. [ 6982.405594] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  4. [ 6985.477839] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  5. [ 6986.501592] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  6. [ 6987.525845] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  7. [ 8434.437556] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  8. [ 8476.421803] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  9. [ 8980.229593] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  10. [ 8983.301834] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  11. [ 8984.325575] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  12. [ 8986.373819] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
  13. [10947.333595] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  14. [10949.381835] dwmac-sun8i 5020000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx


Unfortunately even with this ethernet reduce to run at 100 Mbps network link still goes down twice in 6 hours, requiring at least
a cable adjustment (e.g. unplug/replug cable). Note that nothing moves, the link just goes off on its own.


In the last test, link disconnects after 18 hours 23 minutes at 100 Mbps requiring a
complete power cycle, led goes off. Note that nothing moves, the board is left overnight
to monitor stability. Adjusting the cable (disconnecting and reconnecting) did not restore
the connection. I did not successfully restore the ethernet connection even after a power
cycle, even tried power cycling the connected upstream ethernet switch.
  1. [77171.726686] dwmac-sun8i 5020000.ethernet eth0: Link is Down
  2. [77171.727097] br0: port 1(eth0) entered disabled state
  3. root@DietPi:~# uptime
  4. 12:59:20 up 1 day,  2:30,  1 user,  load average: 1.00, 1.00, 1.00
  5. root@DietPi:~# ip link
  6. 2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq master br0 state DOWN mode DEFAULT group default qlen 1000

note that it is uncertain if this is a single piece defect or that it could be an
issue with the ethernet PHY.

got ethernet back after an hour,  startup from cold, setting ethernet speeds to 100Mbps
  1. ethtool -s eth0 speed 100


  • Floor ag123
  • 2023-12-3 11:10:34
This post was finally edited by ag123 at 2023-12-3 12:17


I did one thing, changed the ethernet cable. Maybe I should have done that earlier :rofl:

now it looks different:
I’ve just completed an Android apps update, Orange Pi Zero 3 is my Wifi hotspot, Orange Pi’s official ubuntu images released in Oct 23:

orangepi@orangepizero3:~$ vnstat

  1. <div>                      rx      /      tx      /     total
  2. eth0:
  3.          today    413.90 MiB  /    5.22 MiB  /  419.12 MiB
  4. wlan0:
  5.          today      4.24 MiB  /  417.69 MiB  /  421.93 MiB</div>


But that I only managed to configure a 2.4Ghz 802.11g wifi hotspot in the official images with hostapd. that is 40 Mbps.
Throughput is decent though today. i’m running at 100 Mbps at the ethernet end.

1 link drop, recovers on its own - it’s a new cable ! uptime 1 hour

after replacing the cable with a new cable, link has been up 20 hours without a new link down event.
the hardware issue is a defective cable.

Aw, this became an exceptionally nice post. In idea I must put in writing in this way additionally – taking time and actual effort to make a very good article… but what can I say… I procrastinate alot and also by no means appear to go done.        Packers and Movers Chandigarh
1234.. 6NextPage