I’m gonna guess this will be the same issue. These APs also use Qualcomm Chipsets.
https://community.ui.com/questions/U6-Pro-dropping-connections/e3e7c21a-022b-4e8d-89dc-43a454f378e5
But from what I remember they never really acknowledged the issue
@robmaltsystems If it’s a Qualcomm driver issue, then MikroTik should roll back to the working driver version until a fix is available. They have the power to negotiate with Qualcomm as a manufacturer.
The real problem is MikroTik’s complete lack of acknowledgment of this issue. It’s unacceptable that paying customers had to reverse-engineer this problem and find our own workaround (staying on 7.14.3) while they remain silent.
MikroTik could easily revert to working drivers temporarily or at minimum communicate transparently about the issue. Instead, they’re shipping broken firmware and forcing customers to downgrade just to get basic WiFi functionality.
I’m quite sure the driver wasn’t updated without a reason. Most likely, it was to support new devices that only work with version 7.15 or later. If it were as simple as just reverting the change, MikroTik probably would have done that already.
That is most likely the reason.
Regarding Chromecast devices, we do have news. Qualcomm firmware with fixes is being tested internally now.
Great news! I’m looking forward to this.
That’s good news!
Also Intel issues seem to be fixed now with this new update. So hopefully next update will resolve most wifi issues, fingers crossed.
I hope so. These frequent disconnects (“MAC@wifi disconnected, too strong signal, signal strength 158”) are one of the reasons why I cant use wap ax as a daily driver.
In my experience the value of the signal strength is incorrect (signal should be less than -90dBm or something around). And is therefor a valid reason to disconnect. Why the conclusion that you can’t use it?
It continues to strike me as odd there is such a difference in user experience.
I’m connected to my wAP AX (capsman managed) since 06h13 this morning. That’s 7h already. Zero disconnects (and using Azure Virtual Desktop, I would know within the millisecond if the connection borks).
My laptop has Intel AX201 card (Wifi 6, 160Mhz).
I don't know what the actual signal is. It says 158. And when I look at the registration table for this client it says -83. So what? I reported this "too strong signal" very very months ago to support and some weeks ago I provided additional supout.rif and got response that Mikrotik is already working on a fix together with Qualcomm. So now we are here awaiting a new driver fixing a variety of different symptoms/issues.
Why not just drop a link to 7.21_ab85 for those interested in testing. you will soon have an answer!
Caution, this is pre-release that most likely contains other bugs. I can’t guarantee that with this version something will not burst into flames or something. https://mt.lv/nightly-build DO NOT INSTALL THIS if you use it for something important.
No thank you. I better wait for 7.19.2. 
Cheers Normis, now testing on Hap_AX2 and Cap_AX WiFi results are as expected.
====== RESULTS SUMMARY ======
Bufferbloat Grade,A+
====== RESULTS SUMMARY ======
Mean Unloaded Latency (ms),25.2
Increase In Mean Latency During Download Test (ms),2.52
Increase In Mean During Upload Test (ms),0
Download speed (Mbps),256.782
Upload speed (Mbps),25.749
====== LATENCY TEST DETAIL ======
Unloaded - Median Latency (ms),24.81
Unloaded - 95th %ile Latency (ms),33.5
Unloaded - Mean Latency (ms),25.2
During Download - Median Latency (ms),24.71
During Download - 95th %ile Latency (ms),49.83
During Download - Mean Latency (ms),27.73
During Upload - Median Latency (ms),22.98
During Upload - 95th %ile Latency (ms),34.57
During Upload - Mean Latency (ms),23.88
===== BUFFERBLOAT IMPACT ======
Web Browsing,PASS,PASS
Audio Calls,PASS,PASS
4K Video Streaming,PASS,PASS
Video Conferencing,PASS,PASS
Low Latency Gaming,PASS,FAIL
First comes 7.20 beta then 7.20 rc and
then we wait for the new driver in 7.20 stable
It’s on the wifi-qcom part which is a separate package so they can easily make it available in 7.19 chain.
According to Normis already in 7.19.2: hap ax3 random wireless disconnects - #316 by normis
Did you have different bufferbloat test results with previous driver?