My original mtd layout (/proc/mtd)
dev: size erasesize name
mtd0: 00200000 00020000 "bootloader"
mtd1: 00200000 00020000 "uenv"
mtd2: 00200000 00020000 "dsd"
mtd3: 00366832 00020000 "kernel"
mtd4: 026f0000 00020000 "rootfs"
mtd5: 04000000 00020000 "tclinux"
mtd6: 00002100 00020000 "kernel_slave"
mtd7: 00010000 00020000 "rootfs_slave"
mtd8: 04000000 00020000 "tclinux_slave"
mtd9: 13800000 00020000 "system"
I found a workaround to get 160mhz channel to work in 802.11s: Configure the Mesh Points as AX instead of BE.
Unexpectedly, the mesh connects using EHT (Extremely High Throughput), which is only available in 802.11be.
Let me repeat that. To get 160mhz BE mesh to work, you configure it as 160mhz AX mesh. Nuts!
Great find! I am getting my 2 units today and this is the exact config I was looking to do!
Any chance of getting an invite to the discord too?
@kvic discovered a new MediaTek patch and command line switch tool. I added both to my build and nothing blew up (yet).
If you want to try, it's in my github repo.
How is the WiFi strength/coverage vs a Netgear R7800 if familiar with it? Looking at a couple of these, one as a router and one as a wired AP (and backup for the router), at $40 they seem like great value. Itās OK if wifi7 doesnāt work for a while as long as thereās a path towards stable wifi6.
I replaced two R7800s with two W1700Ks. 
Mini review:
The R7800 is almost a decade old but still perfectly adequate as a wifi router or AP. However, the w1700k is meaningful superior in every way.
No idea how it will hold up in ten years, but I betcha it will still be competitive in five.
Any chance of updating the āFlashing Methodā section of post 1? Thanks!
flashing instructions updated in Post 1.
Iāve got 2 of these incoming. Very exciting! One may become a router or stay as an AP.
Given they donāt have USB ports, anyone concerned about potential flash wear if running something that produces a lot of logging (e.g. AdGuard Home)? I normally prefer just sticking a cheap USB flash drive in the routers and send logs there, to protect the internal flash.
Insane value that the OpenWrt devs and in particular those working on this device are unlocking! 
I'm also curious how these perform compared to an MX4300 for router and dumb AP scenarios. I snagged a few of these and I'm wanting to flash them when I have some free time.
I currently use qosmios NSS build, taking advantage of the NSS accelerated PPPoE and NSS accelerated SQM. They both have quad core 1.3Ghz processorsand plenty of RAM for my needs, so I imagine they are pretty similar for non-accelerated performance, is there already support for hw offload upstream and/or built-in?
When I do get these flashed and switch over to them, should I configure the radios as AX for now if I have few or no wifi 7 devices? From what I can tell WiFi 7 is still fairly immature in the whole Linux ecosystem compared to 5 and 6, but configuring as 7 should allow older generations to connect and use their best capabilities, anybody have experience with this or do I just need to just try it out and see?
For adguard home and Pi-hole I use my home lab server running proxmox and these in LXCs.
Not a problem. AdGuard Home logs, just like device logs, are saved in RAM.
MX4300 doesn't support 160mhz channels, so this should be faster even in AX.
Definitely try running everything in BE. Only drop to AX if you have problems with backwards compatibility.
@KOA Can you retest power consumption now that 10G ports appear working?
The device's radio0 could not connect after a firmware upgrade. There were no error messages, but when turned on, the phone could not search for the 2.4GHz signal. However, radio1's 5GHz is functioning normally. The scanning function also could not detect the 2.4GHz signal. But in the WiFi interface, the signal noise is displayed normally, and the transmission power can also be seen.
Thanks but I was looking for power consumption with both 10G ports populated at 10G. I get the feeling it is 3w/port or higher due to the active cooling fan.
