Wifi: Not Found | [On] | Disconnected [Radxa Zero 3E]

[monkeyotg]

Creating a bug report/issue

• [*] I have searched the existing open and closed issues

Required Information

• DietPi version | 9.13.2
• Distro version | bookworm
• Kernel version | 6.1.115-vendor-rk35xx #1 SMP Thu May 1 16:13:28 UTC 2025 aarch64 GNU/Linux
• SBC model | Radxa ZERO 3 (aarch64)
• Power supply used | 5V 3A
• SD card used | SanDisk Ultra 32G

Additional Information (if applicable)

• Can this issue be replicated on a fresh installation of DietPi? Yes, I am on a fresh DietPi.

Steps to reproduce

1. fresh installation with AUTO_SETUP_NET_WIFI_ENABLED=0
2. using dietpi-config to turn Wifi on (at this point shows: Not Found | [Off] | Disconnected)
3. reboot, no Wifi (at this point shows: Not Found | [On] | Disconnected)
4. change the configuration of /boot/dietpi.txt to AUTO_SETUP_NET_WIFI_ENABLED=1
5. reboot, still no Wifi

[Joulinar]

The E version don't have WiFi onboard. https://radxa.com/products/zeros/zero3e/

[monkeyotg]

additional information, I'm not sure if it's relevant

I turned Bluetooth on in dietpi-config, and got this error

dbus.service is not active, cannot reload.
invoke-rc.d: initscript dbus, action "force-reload" failed.

reboot, then the device disappeared in my LAN and no more reachable per SSH.

[monkeyotg]

Oh, sorry, just knew it. I thought it is 3W + Ethernet. Sorry for such a stupid issue.

[MichaIng]

😄 no problem, how often I am mixing things up when rotating between some SBCs here.

additional information, I'm not sure if it's relevant

I turned Bluetooth on in dietpi-config, and got this error

dbus.service is not active, cannot reload.
invoke-rc.d: initscript dbus, action "force-reload" failed.

Can be ignored. DietPi by default has no dbus installed. But some packages have hooks included which try to restart the dbus service, to register with it. Not needed for Bluetooth core functionality, other than an actual Bluetooth adapter 😉.

[monkeyotg]

@MichaIng btw, what does DietPi exactly do, when I turn Bluetooth on in dietpi-config? After doing that, I can't get Ethernet to work. Can not get IP from DHCP Server, and doesn't work with static IP either. It's so weird, I have played with the Bluetooth Configuration on another Radxa Zero 3E (You know, I didn't realize that 3E doesn't support Wireless before yesterday. 😅), that devices worked fine after rebooting. I think it's 99% unrelated to this setup, but I have absolutely run out of ideas, so...... Anyway, thanks!

[MichaIng]

https://github.com/MichaIng/DietPi/blob/dev/dietpi/func/dietpi-set_hardware#L1162

• It installs the bluez APT package
• It (tries to) enable some modules, bluetooth most importantly
• If rfkill is installed, it assures Bluetooth devices are unlocked
• It enables the Bluetooth base service: systemctl enable --now bluetooth

This should not affect Ethernet, of course. If it does, please check journal logs:

journalctl -e

The Radxa ZERO 3E has no onboard Bluetooth either, so you have an external Bluetooth adapter? Otherwise, no idea whether these drivers/services can cause any cross issues when there is actually no Bluetooth adapter. Of course they shouldn't, but the 3rd party drivers are usually incredibly crappy 😄.

[monkeyotg]

https://github.com/MichaIng/DietPi/blob/dev/dietpi/func/dietpi-set_hardware#L1162

• It installs the bluez APT package
• It (tries to) enable some modules, bluetooth most importantly
• If rfkill is installed, it assures Bluetooth devices are unlocked
• It enables the Bluetooth base service: systemctl enable --now bluetooth

This should not affect Ethernet, of course. If it does, please check journal logs:

journalctl -e The Radxa ZERO 3E has no onboard Bluetooth either, so you have an external Bluetooth adapter? Otherwise, no idea whether these drivers/services can cause any cross issues when there is actually no Bluetooth adapter. Of course they shouldn't, but the 3rd party drivers are usually incredibly crappy 😄.

Yes, I also think its irrelevant, but as I said, I have really really really no idea now. 😅 But Thanks!!!!!

[MichaIng]

But just to clarify:

• This is a Radxa ZERO 3E?
• And you enabled Bluetooth while there is no Bluetooth adapter, or did you attach one via USB or so?
• And in case you attached one, you did not enable WiFi as well, did you? Having WiFi and Ethernet enabled concurrently currently causes issues, since both are tried to be configured for default route.

[Joulinar]

The E version don't have wifi, even if enabled

[MichaIng]

Yeah I mean if a USB WiFi+BT adapter has been attached.

[Joulinar]

As I understood there was no adapter.

[monkeyotg]

But just to clarify:

• This is a Radxa ZERO 3E?
• And you enabled Bluetooth while there is no Bluetooth adapter, or did you attach one via USB or so?
• And in case you attached one, you did not enable WiFi as well, did you? Having WiFi and Ethernet enabled concurrently currently causes issues, since both are tried to be configured for default route.

no usb dongle

I didn't realize that there is no Wifi/Bluetooth support on the 3E in the first place. I thought it should be 3E=3W+Ethernet. I bought the 3E for a use case that only requires Ethernet. However one day I came across another idea that requires Wifi. So I turned Wifi on for testing, without thinking. But it turned out Wifi: Not Found | [On]. Then I did some searches and found some reports from other users (with different SBCs) saying that in some cases Bluetooth should be turned on at the same time. So I tried turning Bluetooth on, but didn't work. I thought, maybe there were some conflicts with my current settings. Why not test it on fresh DietPi? You already know the story after that. I posted this issue. Then after @Joulinar's reminder, I realized that the 3E does not support Wifi/Bluetooth for the first time.

The first try, everything worked without a problem after rebooting. But the second try, after rebooting Ethernet doesn't work. I've tried different images, under the Debian from Radxa the system can find the hardware:

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether fa:7a:ec:03:f3:21 brd ff:ff:ff:ff:ff:ff
    inet6 fe80::b6cb:cf0b:3961:c2c6/64 scope link noprefixroute 
       valid_lft forever preferred_lft forever

but can not get an IP Address from DHCP Server, and it doesn't work with a static IP either.

Troubleshootings were done with an usb Ethenet adapter and now I am waiting the micro hdmi cable.

[MichaIng]

WiFi is still turned "On". Should not have an effect, but please try to turn it off, and also turn off Bluetooth explicitly.

[monkeyotg]

WiFi is still turned "On". Should not have an effect, but please try to turn it off, and also turn off Bluetooth explicitly.

Thanks for the tips. But how can I do it? The onboard ethernet doesn't work now. And The default configuration of DietPi doesn't seem to support usb ethernet adapter by default. With AUTO_SETUP_BOOT_WAIT_FOR_NETWORK=0 I can boot the 3E, but the usb ethernet adapter does not work, I can't see usb ethernet adapter by inputing ip address, but I can find onboard eth0 without ip address. This means I have no chance to get Internet on DietPi, so there is also no chance to do apt install to try turning/returning wifi/bluetooth on and off.

Btw, I got my micro hdmi cable today, so I'm using console through a monitor now.

Maybe somehow the hardware/firmware was broken.

[MichaIng]

But how can I do it?

In the menu, select Wifi: Not Found | [On]. It should actually ask you right away whether it should be disabled, since no adapter has been found. If it does show a sub menu instead, select Disable there.

DietPi doesn't seem to support usb ethernet adapter

When multiple Ethernet adapters are present, it uses the first which has a cable attached.

If is still does not work after disabling WiFi (and Bluetooth) and rebooting, please show the output of these commands:

ip a
ip r

[monkeyotg]

ip a ip r

Thanks! But doesn't work.

• usb dongle attached

• usb dongle removed

[MichaIng]

Hmm, got it. So I don't think this is related to WiFi Bluetooth, but probably some issue with the kernel/driver for the onboard WiFi adapter. Ah wait a second, I do have a Radxa ZERO 3 here now (thanks to @StephanStS), so can test.

[monkeyotg]

Hmm, got it. So I don't think this is related to WiFi Bluetooth, but probably some issue with the kernel/driver for the onboard WiFi adapter. Ah wait a second, I do have a Radxa ZERO 3 here now (thanks to @StephanStS), so can test.

I'm not familiar with Linux. So I'm a bit confused.

some issue with the kernel/driver for the onboard WiFi adapter

If I'm on the right track. It sounds like, it is kind of “software-related” issue. This issue should be resolved by flashing a fresh DietPi or other images. However, I've tried several different images, this issue does not walk away.

[MichaIng]

No more testing for me: it broke. My first ever broken SBC, among .... 30 or so? And it broke the very first ~30 minutes of usage. Awesome ...

No idea, I guess it is the SD port, maybe the mechanical mechanism of fixating the SD card in the port. There are two metal contacts touching when the SD card is inserted. They still do, but maybe one of them has been bend somehow.

I flashed the image, it booted and connected via Ethernet, though static IP, wanted to test DHCP later. It was a bit slow while upgrading some packages, and with the cables it was in a bad position, so I lifted it, aiming to move it elsewhere. In that moment the SD card remounted read only, and then vanished completely. All I had was an in-RAM bash session with no access to anything on the SD card. It was an old SD card, so I flashed a new one, but it does not boot anymore at all, bootloader gives zero output, i.e. no bootloader at all. So I guess firmware is not detecting any SD card.

Will try to find some high resolution pictures of how the port is intended to look like tomorrow, otherwise this was the shortest SBC lifetime ever 😄.

If I'm on the right track. It sounds like, it is kind of “software-related” issue. This issue should be resolved by flashing a fresh DietPi or other images. However, I've tried several different images, this issue does not walk away.

Since a fresh image does not work either, and the adapter is detected etc, I am guessing a kernel or driver issue. We had some other cases where it failed at this stage, and was an issue with the kernel. Software does nothing special, normal Debian-standard ifupdown.

Please test with new kernel. I just put one into the repo. As said, Ethernet generally worked for me.

[monkeyotg]

My first ever broken SBC, among .... 30 or so? And it broke the very first ~30 minutes of usage. Awesome ...

Sorry for your lost. 😅

Please test with new kernel. I just put one into the repo. As said, Ethernet generally worked for me.

Could you tell me where I can find the new kernel, and how? Please forgive my ignorance. 🐸

[monkeyotg]

Will try to find some high resolution pictures of how the port is intended to look like tomorrow, otherwise this was the shortest SBC lifetime ever 😄.

I took some macro photos of the tf card slot on a health 3E, hope they help.

https://upload.adminforge.de/r/3R2BWuBcgt#qDX95TArBrhrTKFEXnxWJ/jbNaKOw6CUCH5Gg9QutoY= https://upload.adminforge.de/r/pQBnHQdmWU#85QgMHcxgqsnTPP7B5eyNNACfVcdbTPMBzjnQjyMe1U= https://upload.adminforge.de/r/JLjnbbr3Up#gm6x+MY+tvachOKgWvcKvudi6adfXQ2hXQ2vCuFuvKA= https://upload.adminforge.de/r/NyTBkfp1Fn#rYnQWD5tdK2LBGucwrGK6I2w/m/2JdDnuj8MX0JlKew= https://upload.adminforge.de/r/ss1PTv__JR#DrPfbKNHA42nk13Wa+3C6i2eJHHO+UxflX3A9CXg2mQ= https://upload.adminforge.de/r/S6r4PzjH3b#Bi2zGTLDEVr0WR3dw4mUNts1EXgyavSS6tIm0HscILk=

[MichaIng]

Could you tell me where I can find the new kernel, and how? Please forgive my ignorance. 🐸

It is on the APT server already. Just do apt update && apt upgrade.

I took some macro photos of the tf card slot on a health 3E, hope they help.

Thank you very much, that was very helpful. With daylight and the help of your photos I could verify that the SD slot looks and behaves totally as expected, nothing unusual. SD card sits very tight as well, nothing unusually bent or loose or so. Aside of the fact that the SD card was the first thing lost, when it was not able to (re)boot from it (while still showing serial console output at first), I pressed it a little gently, and felt I might have bent something. But I am not certain it was only the bottom pins moving down further as they can. When the card is inserted, one can push it further down as those 8 contacts at the bottom have more room to move. But the springs push up normally when taking out the card, so it is not like doing this would damage/bent them permanently.

So I think it is hardware damage elsewhere. But things started to fail when I wanted to move the SBC aside with serial console and Ethernet cables (and power of course) attached, holding it at SD card and PCB at the other side. Holding it at SD card probably was the trigger for the initial failure, but I don't think it caused it. As said, extracting the APT packages, even just small ones, took very long at some point, which was the reason I wanted to move the board aside and work on something else. After all, I think this was a sign for some issue already, probably some electronic damage arising while the board heated up during operation.

Let's see whether someone on the Radxa forum has in idea how to debug or what to watch out for to identify a hardware damage: https://forum.radxa.com/t/27348

[monkeyotg]

It is on the APT server already. Just do apt update && apt upgrade.

tried by using the usb adapter, then disabled the ethernet, unplugged the usb dongle, rebooted, re-enabled the ethernet, but the issue persists

perhaps the device is somehow dead 🤷🏻‍♂️

[MichaIng]

A static IP did not work in your case either, right? Maybe both our 3E have a hardware defect 😄. At least yours does boot.

If you want, I can create a mainline kernel based image for you to test. We use the vendor kernel only, because it is needed for onboard WiFi driver on 3W, but that should be irrelevant for 3E. The chip itself is generally well supported by mainline Linux, aside of some more exotic things like NPU and such.

[monkeyotg]

no, a static IP does not work either.

If you want, I can create a mainline kernel based image for you to test.

If it's not too much trouble for you, I'd be happy to be the lab rabbit 😂

[MichaIng]

Build running: https://github.com/MichaIng/DietPi/actions/runs/15429768638 You'll find the image here, once done: https://dietpi.com/downloads/images/testing/

[monkeyotg]

Build running: https://github.com/MichaIng/DietPi/actions/runs/15429768638 You'll find the image here, once done: https://dietpi.com/downloads/images/testing/

Thanks!!!!! will test it tomorrow.

[monkeyotg]

It didn't save the Ethernet. I think it would be a hardware failure. 💀 I think I could contact the seller to find the possibility of a replacement. My poor 3E, it was disabled at a very very very young age. 😅

[MichaIng]

To be absolutely sure, you could try it with an official image from Radxa. I'll do this with mine now, following the suggestion of the Radxa support guy.

Makes sense following standard procedure 😉.

[MichaIng]

I cannot believe it, the Radxa image boots 😮. Now testing the mainline image I created for you.

.... nothing. Now testing the previous vendor kernel image.

... it boots 😮. Hmm, this is the exact same issue that did not boot at all earlier, and that day I tried multiple times with different SD cards, PSUs, attached Ethernet or not, attached HDMI and not, just to rule out any effects. So that problem seems a bit random. Trying the mainline kernel image again.

... yeah now the mainline image boots as well. I also tried a few times to trigger the failure, and it seems to be the SD card slot indeed: it continues to reboot successfully if I just unplug and plug again power. But it sometimes fails when I unplug and plug again the SD card. And then, it fails every boot. As there is not much play or anything, I cannot feel or identify any difference in how I insert the SD card, but obviously there are nuances somewhat.

Before I have up with the board, I pressed the left side outer contact inside a bit, since it seemed essential to me that it touches the inner one with an SD card inserted, and on your images it looked like that one would be slightly looking more inwards, when no SD card is inserted. Maybe that changed something, though with an SD card inserted, both contacts are bent much more outside, clearly touching each other, so no idea how what I did could have made a difference.