It may have something to do with the fact, that IPQ-5010 has only 2.4GHz WiFi built in SoC and that 5GHz radio is added by attaching separate radio chip to PCIe interface. Previous AX hardware (at least some of devices) was built around IPQ-6010 which has both 2.4GHz and 5GHz radios built in.
When routerboot initializes hardware, it initializes it in certain order and very likely built-in hardware comes before any external one. Which of built-in gets first … might be design decission by SoC designer.
I get the explanation (and already suspected as such) but setting the order of those interfaces is something which is done by startup script.
It was already a mess when they changed from legacy to AX and now it has changed again within AX. Double mess.
Search this place when initial AX devices came out and how many posts there where about ordering of radios.
A tiny bit of consistency would be helpful for all …
Plus note …
It effectively does use 160MHz channel.
AX211 card in my laptop can use it just fine.
Did some rudimentary testing from laptop (using Btest windows app) over wap AX over Hex Refresh (yes, got me that too to play with) towards RB5009.
I saw speeds around and just over 900Mbps on download and 600-ish on upload.
Already connected to it all afternoon, zero drops (and Azure Remote Desktop would throw me out straigth away if a hickup happened, never happened all afternoon)
i also hope support for 160mhz is the reason for that improvement in 5Ghz Max data rate to 2400 Mbit/s, i think 160mhz channel support will make it an interesting device for some situations
but i have an Audience its third radio also suppports 160mhz channel (also is an external/aditional radio), i tested it and i can see consistently links at 1500 - 1700 Mbit/s datarates but throughput on that radio never surpases 500-600 Mbit/s, cpu usage stuck on 25% showing only one cpu core loaded at 100%
tested with traditional wireless package and also with newer wifi-qcom-ac, same result
i hope MikroTik does not have the same problem with ax devices, Otherwise there will be no point in supporting 160MHz channels nor devices like Chateau PRO ax
When you have the chance please test if the load is distributed across multiple CPU cores
thank you
From what I can see, it is.
Not equally distributed (sometimes up to 10% difference) but it is effectively being distributed.
I don’t think so … script says
:local ifcId [/interface wifi find where default-name=wifi1]
set $ifcId configuration.mode=ap channel.band=2ghz-ax disabled=no
…
So the script knows that wifi1 is 2GHz radio (… where default-name=wifi1 …), it doesn’t set default name.
Based on how linux kernel works, default device names are assigned according to order of being discovered. And discovery is performed by device drivers, so order of device driver loading does affect interface naming if there are devices of same kind handled by different driver (think of non-renamed ethernet interface names if there are different boards installed, e.g. one intel-based and one broadcom-based … if e.g. intel driver is loaded first, then eth0 will be intel interface and eth1 will be broadcom one and vice versa when order of driver loading is reverse). But if there are multiple interfaces driven by same device driver, then interface naming depends on hardware enumeration, often based on PCI bus number … SoC devices (or on-board devices in case of traditional computers) will most often have lower PCI bus number than those on add-on boards.
In case of RB, where only one wireless/wifi driver can be loaded at a time, interface naming doesn’t depend on order of driver loading, it depends on hardware enumeration …
What MT might have done (but they didn’t think this could be important) is to add interface renaming in the startup script. Whis is not really sticky as it’s possible to upload custom one so perhaps not really a sollution.
I guess we’ll simply have to live with the fact that even default wlan/wifi interface name doesn’t mean anything.
I was indeed referring to renaming the interfaces, bad wording from my side.
And yes, I’m also thinking we simply may have to live with it.
Anyhow, I launched a support ticket, we’ll see what they say about it if they consider changing it or not.
A “No” I already have, a “yes” I can get 
lol, very educational debate 
Support confirmed: is due to different chipset and it is not going to change.
So another thing to keep in mind when users complain about radio not functioning if wrong one got the wrong settings.
Overall I’m not so much impressed by the wap ax. The shorter range in the 2.4ghz band is a major drawback for me. The 180° antenna also doesn’t cover the area behind the device well. I’d prefer a 360° coverage and a stronger 2.4ghz signal. With cap ac I have a useable signal outdoors - with wap ax there is no signal at all. It would be nice to have all the features of wifi-qcom-ac (like vlan-id via datapath or the qos-classifier setting). In the end I don’t see any advantage over my cap ac in my environment. Cosmetic (?) bugs like this one are then the icing on the cake why I most probably will return the wap ax.
Soo… Good thing I ordered 3 of them… one to each room i want to cover with wifi…
Glad I held off now, the device sounds like a bit of a wap,wap,waaaaaaaaaaap. Sorry couldn’t help myself
You are waiting for WiFi 7 ? 
Why didn’t you go with the cap?
I think of wAPs as outdoor units.
WAP has seemingly a much higher WAF.
Bingo… My wife was so happy yesterday when she saw how wap ax will look like.
I do have 2 cAP ax and one ax2 in kids room but I wanted to replace them with wAP ax because it will look better on TV cabinet or shelf.
I have 3 of them, one in the living room, one in master bedroom and one in kids room so i don’t think there will be any problem with coverage. I don’t care about signal outdoor TBH.
I replaced AX3 with wAP AX in my office / lab setup.
It’s a lot more discreet and I wasn’t using any of the other ether ports on that AX3.
Haven’t used fixed ethernet cable anymore for my laptop since then
This !
cAP is huge and sticks out like dogs balls.
It’s just that dogs balls are more pleasant to look at.