TP-Link Archer C50 (EU) V6.0 - 5 GHz interface not present

Installing and Using OpenWrt
27

I have been running on a C50 (CA) V6.0 for about a month now. The sysupgrade bricked my router, however when I ran the tftp back2stock recovery, I ended up running openwrt. I am a bit new a trouble shooting this, so I likely made a mistake.

I have run into a problem where it will disconnect the devices, but since I am running it as an access point the computers got really confused, and did not want to jump back on to the primary router. Seeing that rebooting solved the issue, I just use cron to reboot every night and have been running without a problem.

Just sharing my experiences on this firmware, with a different local of hardware. I also get faster throughput than stock.

thanks for the work you put into this.

2 Likes
28

Standard flashing through tplink web interface bricks the unit.
However it restores easy by using the tftp version back to normal firmware.

Reflash through TFTP method the openwrt file for tftp worked fine.

So for those coming to this firmware just do the FTFP flash of

experimental-archer-c50-eu-v6-factory-for-tftp(name tp_recovery).bin
it works

thanks goes to all involved in making this work

30

Built a snapshot for v6 using v4 profile in imagebuilder. Only replaced wpad-basic-wolfssl to wpad-mesh-wolfssl, and removed a few ipv6 packages, which you can install if you want. 2.4g, 5g, LEDs, ethernet and everything works as expected. Uploads are a bit flaky as it seems.

Use this image in sysupgrade only. Force install if you have to since the device will think itself as v4 in this build, and preferably remove all existing configs, since they may conflict with the new build.

tplink_archer-c50v6-sysupgrade_2022.02.08_13.59.48.bin

1 Like
32

Intermittent Wifi disconnect ( Wifi light stays fully on instead of blinking and the Wifi interface in Luci shows ? even after restarting Wifi using any method, and the SSID disappears from the network, reboot solves it ) issue after a few hours ( 8-12 hours ) with the same MCU messages in the logs and no error if disabling WMM

MT7628 OpenWrt 21.02 - MCU message 08 (seq 1) timed out

This sees to be a regression Wifi issue ( not present in 19.07 ( unconfirmed report ) occurs in 21.02 )

Marssl: 21.02.0rc4 - Less problematic after all package updates

subhranil: 2022.02.08 ( 21.02.1 ?) - Problem exists even after all package updates

Official Archer C50v4 21.02.2 firmware running on the v6 ( don't require 5ghz) - Problem exists even after all package updates

Note, have htmode40+ and legacy ( required for standardisation ) may not be related as reports of intermittent wifi issues without htmode40+ and legacy were still probably there ( unconfirmed as to frequency of the same etc )

33

I also have the TP-Link Archer C50 (ES) V6.0 device. I tried using a V4 image using the approach here, using the Web UI to load it, but it rejected it.

However looking at the TP-Link official image files, they are completely different for the EU and ES versions of the device. The EU image BIN file is 7930368 bytes, and the ES image BIN file is 2183572 bytes. So I think they're ripping us off in the ES regions and selling a cut-down device. Maybe it even has less flash and/or RAM, which would likely make it useless. So I'm giving up with this device. I'll use it as an access point elsewhere and look for some other router to put OpenWRT on (which I need for load balancing and monitoring) -- unless anyone has any other ideas.

34

I can update that the official Archer C50v4 firmware update 21.02.3 for the v5 too, but using for the V6 ( remember no 5ghz required just stability on the 2.4 ghz ) WITH WMM DISABLED updates fine using sysupgrade and shows no intermittent Wifi issues.

Side note: Since SQM is already installed, the so called WMM "QOS" feature may not be an ideal piece of code and there should be a way to enable "bn", ( WMM disabling just enabled "ag" at upto 20 MBPs tested ) and not try to do any QOS stuff on the Wifi interface. Funny that all OpenWRT routers with "abgn" and WMM enabled never really performed more than 20 Mbps ( even with WAN at 100 Mbps ) in six months of testing. So disabling WMM wasn't such a big deal when comparing Openwrt and Openwrt, even better at a steady 20mbps when assumedly WMM was not interfering with the speeds.

Another option is to try and test the official MTK wifi drivers:

Do not have the werewithal currently. Although it might be a good time to investigate Open Source code and developer team psychology.

Linux has been plagued with the "reputable developer syndrome" as in their code is only allowed to be built on and patched ever so slightly always tiptoeing around "developer reputation" and the like.

If we look at the angle that the original Openwrt open source driver was "booked for eternity" with a fixed amount of code, be it good intent or sheer inability to rewrite equally performant code in open source, or even employer and law forcing open source to write non-performant code so as not to appear to be copying the official code, be it closed or open, we have a problem.

It might be time for an open contest to invite corporates to donate code chunks as replacement for large sections of the code and a free for all forking culture so that main-line does not get booked by non-performant good intention or performant but commercially motivated bad intention code.

Basically the open source MTK wifi driver needs careful scrutiny and updating and this statement by Openwrt ( https://openwrt.org/toh/tp-link/archer-c50 ): "

(Aug 2020): The cause for poor 2.4 GHz range experienced with some C50 v4 and probably C50 v5 models has been found. Also reports of random reboots OpenWrt forum Bug Report 2781 Fixed 2.4 GHz poor wifi speed (Feb 2021) from 19.07.7 mt76: mt7603: add additional EEPROM chip ID

certainly doesn't mention the WMM issue and certainly isn't all the problem with the MTK Openwrt driver.

35

Did anybody try to use/build the official firmware? According to the readme the archive has everything needed to build a TP-Link firmware:
https://static.tp-link.com/resources/gpl/Archer%20C50_V6_GPL.tar.gz

I imagine building a custom TP-Link firmware with shell acces would help with figuring out how things should be for OpenWRT?

36

can you please tell which serial connector you used

37

Hello, can you tell me how to make firmware for recovery, I have a router with V6 (RU)

38

Hi all, if I understand correctly, at this point the OpenWRT status for this device:

With 19.07.8 (and 19.07.9) everything works more or less stable except for the lack of 5 GHz WiFi.
With higher versions there is a WiFi regression as reported by @idcrises but it can be worked around by disabling WMM.
Is this correct?

Did anybody run an official 22.03 build?

39

A post was split to a new topic: Archer C50 v6 - all lights blinking after OpenWrt installation

40

Thanks to @Marssl builds and device details, I've managed to add Archer C50 V6 into current master of OpenWRT.
I'm testing it since today, so far it looks pretty good, both Wi-Fi interfaces have been recognized by their model names. I'm using my Archer as a wireless repeater to the LTE router with relayd. In comparison with @Marssl builds download speed is much more stable.
For those of you who want to build an image yourself below is link to the code:

Keep in mind, that flashing device running on factory firmware require additional steps - othewise you may brick the device. Instructions here: https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=01dcd574a24864c10a9c4df4304a3ce0da5c11a9
Here is the sysupgrade (USE AT YOUR OWN RISK)
https://drive.proton.me/urls/1C7GXZ5PQ0#Zhq1YgI2zkrF

1 Like
41

Hello @remgodow

Thanks for providing your code.
I also gave it a try earlier in the year but couldn't get it stable enough so I reverted back to stock FW. I see that your changes are very similar to what i attempted except I based my attempt on branch openwrt-22.03 instead of master.

I think the wps led can be removed since this board does not have a dedicated WPS LED.

Also, since it's the first time i try to add support for a device, I don't know how i'm supposed to get some of the values.
For instance : How do you know the correct values for mac-address-increment.
Or the values for the HWID, HWREVADD and HWREV ?

Below is my attempt :

42

@Renaud11232 I think you are right about wps led, I wonder if wan_orange is usable, if not then its section could be removed as well.

I've based my code on C50 v4 DTS and commit adding support for Archer C6 v3 which has the same 5GHz chip https://github.com/openwrt/openwrt/commit/d22fb7f4fd833c420bc2ab34384862da0a6688b5,.

Initially I've replaced whole PCI configuration with C6 one. Then I figured out that 2.4GHz is on the SoC, therefore its defined in mt7628an_tplink_8m-split-uboot.dtsi. So I removed 2nd pci device, which is where my mac-address-increment comes from.
From what I remember after reading other dts files (mainly C6 v3), I believe that usually ethernet interface gets the "factory" mac from ROM, without increment.
Given the nature of TCP/IP stack I believe that mt7628an_tplink_8m-split-uboot.dtsi is bugged as both ethernet and 2.4G gets the same MAC. So my conclusion is that all interfaces should have unique MACs, and the easiest way is to increment them by 1,2... and so on.

As for the values of TPLINK_HWID, TPLINK_HWREV, TPLINK_HWREVADD - I have no clue, I've copied them from C50 v4. I think that they may be used for generating factory image and recovery image (which is not possible for our unit). They may be hidden somewhere in the oem rom or hardware datasheet.

Unfortunately my device froze an hour ago with about a day and a half uptime. I wonder if it'll occur again.

43

4 posts were split to a new topic: Archer C50 v6 (eu) only 2MB free space

44

Hi, what is the current status of C50 v6 support? You wrote you managed to commit the support to the main repo? So is it available in snapshot? Thanks!

45

No, it is not in the main repo, I've just made a fork based on master branch of openwrt and generated snapshot builds out of it.
At that time the freezing issue have been still present, so I've flashed stock on my device.
Fortunately it may not be a problem as of today, as it seems that the MT7628 bug have been fixed - https://github.com/openwrt/openwrt/issues/10074
You could try to rebase my fork on top of current master, and check yourself.

1 Like
46

I merged all your commits into a single commit and put it on top of the current master. There were no issues applying the combined set of your commits and I could also build an image without problems. Sources: https://github.com/jschwartzenberg/openwrt/tree/archerc50v6
Compiled images here: http://files.s-berg.nl/archer-c50-v6/

I did not test these yet. Maybe somebody can test these ahead of me. If this works I propose creating a PR to master and hopefully it can be merged.

Thanks a lot!!

1 Like
47

@jschwart I installed your image on a fresh v6, it flashed and booted properly. 5 GHz seems to work as well (61 Mbps up, 43 Mbps down over WLAN over my 250/50 line).

One odd thing though, maybe it was just a fluke(?): I installed luci and then configured wifi over it. After setting WPA3 and a password and applying the changes, the router crashed and I had to restart it manually. WiFi 5 GHz was enabled after that as requested in luci. But could not reproduce the problem.

EDIT.: The crash occurs when enabling the 5 GHz WiFi in LuCI. A manual reboot of the device is necessary.

1 Like
48

That's great news!! So it seems we have one known issue, but with a known workaround! Be sure to keep us informed if you run into more issues and whether things are stable further one.

@remgodow @Renaud11232 would one of you like to have the honour of creating a PR to master or shall I do it?