I finally returned the WAX220 last week and got a NWA50AX Pro now. So far no outages on my phones. I am using the same config that I was using before.
The discussion there is to disable 11ax on 2.4GHz, are you disabling 2.4GHz only or both to get rid of the issue on Apple devices?
I disabled it both 5ghz and 2.4ghz, only this way it worked for me
whelp, it just happened again with the new AP. guess its either apple or my config. trying to set things up fresh one more time, not very optimistic though.
I have at home 2x Netgear R7800 and 1x WAX220
Big difference: by default, r7800 supports only 32 clients (per band)
I have 150+ IoT devices already at home (sockets, switches etc.) so without workaround it was not possible to use r7800 unless I split them amongst the 2 (but then I would have to create 2 different SSID).
With WAX220, although I bought it laptops (2.5Gbps) I decided to deciate it to 2.4GHz IoT devices - I am not sure how many it can serve but 150+ does not make WAX220 suffering.
However with Macbooks connected WAX makes me seek with is upload spead (sometimes 10Mbps) while the same conditions and r7800 can go up to 500Mbps.
It seems that Netgear is aware of the disconnecting Apple devices issue, and is able to provide beta firmware to test their fix. (https://community.netgear.com/t5/Business-Wireless/WAX220-WIFI-Disconnects-on-Apple-devices/m-p/2376777/highlight/true#M12150)
Stable release is scheduled for May 24th.
Hopefully it will give insight how to fix this in OpenWrt as well?
2 Likes
They have just released the new firmware that targets the Apple devices issue!
Here is the link https://kb.netgear.com/000066208/WAX220-Firmware-Version-1-0-3-4
2 Likes
GPL source is here:
https://www.downloads.netgear.com/files/GPL/WAX220-V1.0.3.4_gpl_src.zip
I am having this issue too. I do not have a 2.5bg switch with or without PoE. Is there any development on this? Has someone found a fix?
You mean the apple devices problem?
Or the eth disconnections problem?
I'm assuming its an issue with the link negotiations for eth0. I did a speed test with iPerf3 to the AP using its own wifi and speeds were as expected, but if I run a speedtest to my router or anything over the APs ethernet connection, the upload caps out at roughly 400Mbps.
There seems to be more to this issue than meets the eye. I used ethtool to force gigabit on the port and now the AP is getting full gigabit duplex speeds to and from. If I do a speed test to the AP its self, I get expected speeds of about 1.4Gbps, but not if I do a iPerf speed test to my NAS over the wifi or a computer running iPerf cabled to the switch, but I get full gigabit from the same computer on the switch to the AP!
iPerf3 Breakdown:
Wired computer to AP: full gigabit up and down as expected.
Wireless client to AP: speeds as expected (1.3Gbps) up and down.
Wireless client to wired computer over AP: full gigabit down, 360Mbps Up. Not as expected.
Seems an issue with full duplex.
It has been reported also for zyxel nwa50ax pro.
The culprit seems to be the ETH adapter in the access point which Is unable to negotiate a correct speed/duplex.
Perhaps you could reflash the vendor firmware and check if the issue persist. If It doesn't, stay on originale firmware until openwrt fixes it
So this could be a hardware issue? I am getting a second WAX220 so I'll test the factory firmware to see if the issues exists there. Otherwise, the only fix seems to be plug the AP into a 2.5Gb switch?
I had back in the days a similar issue, or my issue was that I was not able to browse etc after sleep on my Mac things. Solution for this was adding to wireless config.
option multicast_to_unicast_all '1'
But now it’s not necessary anymore and has been fixed. At least for ipq807x.
Maybe you have already tried this, so my apologies for repeating 
Well,
In the nwa50ax pro I posted all my troubles and questions about that "bug".
They told me that it's not a bug but an hardware problem of some devices or of my eth cable (which is a cat 5e placed there 5 years ago and flawlessy served a tplink eap245).
I think it's a bug but of my opinion, in the end, doesn't matter.
Anyway I returned my nwa50ax pro and buyed a wax220. And I'm keeping the default firmware in it. At least until someone finds a fix or I'll have a 2.5gb switch
Here is the patch from the vendor firmware
$ cat package/tw/autobuild/NTGR_WAX220/driver_mt7986/patches/0106-mtk-fix-iphone-disconnect.patch
Index: mt_wifi/mt_wifi/embedded/wlan_config/config_he.c
===================================================================
--- mt_wifi.orig/mt_wifi/embedded/wlan_config/config_he.c
+++ mt_wifi/mt_wifi/embedded/wlan_config/config_he.c
@@ -22,27 +22,27 @@ static VOID he_cfg_mu_edca_init(struct m
{
/*AC_BK*/
mu_edca->mu_ac_rec[ACI_AC_BK].acm = 0;
- mu_edca->mu_ac_rec[ACI_AC_BK].aifsn = 0;
- mu_edca->mu_ac_rec[ACI_AC_BK].ecw_min = 15;
- mu_edca->mu_ac_rec[ACI_AC_BK].ecw_max = 15;
+ mu_edca->mu_ac_rec[ACI_AC_BK].aifsn = 15;
+ mu_edca->mu_ac_rec[ACI_AC_BK].ecw_min = 10;
+ mu_edca->mu_ac_rec[ACI_AC_BK].ecw_max = 10;
mu_edca->mu_ac_rec[ACI_AC_BK].mu_edca_timer = 255;
/*AC_BE*/
mu_edca->mu_ac_rec[ACI_AC_BE].acm = 0;
- mu_edca->mu_ac_rec[ACI_AC_BE].aifsn = 0;
- mu_edca->mu_ac_rec[ACI_AC_BE].ecw_min = 15;
- mu_edca->mu_ac_rec[ACI_AC_BE].ecw_max = 15;
+ mu_edca->mu_ac_rec[ACI_AC_BE].aifsn = 15;
+ mu_edca->mu_ac_rec[ACI_AC_BE].ecw_min = 10;
+ mu_edca->mu_ac_rec[ACI_AC_BE].ecw_max = 10;
mu_edca->mu_ac_rec[ACI_AC_BE].mu_edca_timer = 255;
/*AC_VI*/
mu_edca->mu_ac_rec[ACI_AC_VI].acm = 0;
- mu_edca->mu_ac_rec[ACI_AC_VI].aifsn = 0;
- mu_edca->mu_ac_rec[ACI_AC_VI].ecw_min = 15;
- mu_edca->mu_ac_rec[ACI_AC_VI].ecw_max = 15;
+ mu_edca->mu_ac_rec[ACI_AC_VI].aifsn = 15;
+ mu_edca->mu_ac_rec[ACI_AC_VI].ecw_min = 10;
+ mu_edca->mu_ac_rec[ACI_AC_VI].ecw_max = 10;
mu_edca->mu_ac_rec[ACI_AC_VI].mu_edca_timer = 255;
/*AC_VO*/
mu_edca->mu_ac_rec[ACI_AC_VI].acm = 0;
- mu_edca->mu_ac_rec[ACI_AC_VO].aifsn = 0;
- mu_edca->mu_ac_rec[ACI_AC_VO].ecw_min = 15;
- mu_edca->mu_ac_rec[ACI_AC_VO].ecw_max = 15;
+ mu_edca->mu_ac_rec[ACI_AC_VO].aifsn = 15;
+ mu_edca->mu_ac_rec[ACI_AC_VO].ecw_min = 10;
+ mu_edca->mu_ac_rec[ACI_AC_VO].ecw_max = 10;
mu_edca->mu_ac_rec[ACI_AC_VO].mu_edca_timer = 255;
}
I am just going to manually apply in my tree as there are so many hoops to jump through for submitting a PR for an openwrt patch, ill leave it to someone who does it day in and day out.
edit: maybe this is the correct place? https://github.com/openwrt/openwrt/blob/main/package/network/config/wifi-scripts/files/lib/netifd/wireless/mac80211.sh#L467
edit2: Heres a patch for 23.05.03
--- mac80211.sh 2024-07-05 14:27:14.207087744 +0100
+++ package/kernel/mac80211/files/lib/netifd/wireless/mac80211.sh 2024-07-05 13:59:31.758567460 +0100
@@ -465,25 +465,25 @@
append base_cfg "he_mu_edca_qos_info_q_ack=0" "$N"
append base_cfg "he_mu_edca_qos_info_queue_request=0" "$N"
append base_cfg "he_mu_edca_qos_info_txop_request=0" "$N"
- append base_cfg "he_mu_edca_ac_be_aifsn=8" "$N"
+ append base_cfg "he_mu_edca_ac_be_aifsn=15" "$N"
append base_cfg "he_mu_edca_ac_be_aci=0" "$N"
- append base_cfg "he_mu_edca_ac_be_ecwmin=9" "$N"
+ append base_cfg "he_mu_edca_ac_be_ecwmin=10" "$N"
append base_cfg "he_mu_edca_ac_be_ecwmax=10" "$N"
append base_cfg "he_mu_edca_ac_be_timer=255" "$N"
append base_cfg "he_mu_edca_ac_bk_aifsn=15" "$N"
append base_cfg "he_mu_edca_ac_bk_aci=1" "$N"
- append base_cfg "he_mu_edca_ac_bk_ecwmin=9" "$N"
+ append base_cfg "he_mu_edca_ac_bk_ecwmin=10" "$N"
append base_cfg "he_mu_edca_ac_bk_ecwmax=10" "$N"
append base_cfg "he_mu_edca_ac_bk_timer=255" "$N"
- append base_cfg "he_mu_edca_ac_vi_ecwmin=5" "$N"
- append base_cfg "he_mu_edca_ac_vi_ecwmax=7" "$N"
- append base_cfg "he_mu_edca_ac_vi_aifsn=5" "$N"
+ append base_cfg "he_mu_edca_ac_vi_ecwmin=10" "$N"
+ append base_cfg "he_mu_edca_ac_vi_ecwmax=10" "$N"
+ append base_cfg "he_mu_edca_ac_vi_aifsn=15" "$N"
append base_cfg "he_mu_edca_ac_vi_aci=2" "$N"
append base_cfg "he_mu_edca_ac_vi_timer=255" "$N"
- append base_cfg "he_mu_edca_ac_vo_aifsn=5" "$N"
+ append base_cfg "he_mu_edca_ac_vo_aifsn=15" "$N"
append base_cfg "he_mu_edca_ac_vo_aci=3" "$N"
- append base_cfg "he_mu_edca_ac_vo_ecwmin=5" "$N"
- append base_cfg "he_mu_edca_ac_vo_ecwmax=7" "$N"
+ append base_cfg "he_mu_edca_ac_vo_ecwmin=10" "$N"
+ append base_cfg "he_mu_edca_ac_vo_ecwmax=10" "$N"
append base_cfg "he_mu_edca_ac_vo_timer=255" "$N"
fi
3 Likes
I just got another one of these today and tested the stock firmware and can confirm, that this issue does not exist on the stock firmware. It only appears when running OpenWRT.
Anytime I try and execute ethtool -s eth0 speed 1000 duplex full autoneg on I lose my eth0 link.
Look here:
You'll find my two cents with all the links.
It's a post on the Zyxel AP but the problem Is the same.
I have both AP (WAX220 and NWA50AX Pro) and the problem I have is that my devices connected to the 2.4GHz network disconnect randomly, while on the 5GHz everything is ok (only random packet losses).
Could these problems be linked to the one reported by you?