EAP-650 1.0.13 firmware randomly dropping 2.4Ghz client after some time.

I have two EAP650s running the latest 1.0.13 firmware. I have been experiencing an issue with both units where after some random period of time, my 2.4Ghz clients (mostly IOT type stuff) stop working. They still show up in the list of clients, but the devices themselves don't work. If I restart the AP, the devices re-connect and work correctly. It's happening with both APs. I have a 2.4Ghz only SSID that is unique to each AP that these devices are connected to.
Here you can see it happen:
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
Is there any fix that's planned in the pipeline? It's been over a year now and this issue still persists, unbelievable.
- Copy Link
- Report Inappropriate Content
@FooFoo9 I am back on 1.0.91 that I got again from support and its been stable for me. It sounds like they are investigating how to migrate those changes in 1.0.91 into the 1.1.0 releases because they said their had been a lot of changes.
Very frustrating though. I had been looking at maybe upgrading to a competitors Wifi 7 product but firmware there is just as bad for iot stuff right now.
- Copy Link
- Report Inappropriate Content
I notice it mostly happens at 6am for some reason. I checked your one of your pictures and it also happened at 6am. So as a temporary workaround I set a reboot schedule for the EAP at 6am everyday. Hope it helps with support team in troubleshooting.
Another screenshot
- Copy Link
- Report Inappropriate Content
This is definitely not a time related issue, I have these drops at all times of the day. Even with 1.0.91 there are days where I have to reboot the device multiple times a day.
BUT meanwhile I suspect some client devices triggering this issue. So it might make sense for you to check if there are any devices that have scheduled activities/tasks at 6am. E.g. I noticed more than once that this issue occurs exactly when another household member arrives at home with her Apple devices.
- Copy Link
- Report Inappropriate Content
So support had me update to 1.1.0 and change the 2.4G Beacon interval to 50 from 100. Its been 2 days and no issues so far but I also don't really understand the downside of doing this.
- Copy Link
- Report Inappropriate Content
- Copy Link
- Report Inappropriate Content
I am having exactly this problem! All my IoT devices (40+) are from TP-Link. About 10 disconnect from my 2.4g band and won't reconnect unless I reboot my EAP650. I can't believe this hasn't been resolved yet.
@mackworth Is the custom 1.0.91 firmware still working for you? How do I request it from TP-Link Support?
- Copy Link
- Report Inappropriate Content
Hey guys,
I finally found the solution/root cause for this issue (at least for me). I confirmed this by undoing my changes and the issue occurred again very quickly. The 1.0.91 beta that some of us got from support also doesn't solve the issue, it's just less frequent.
Here's the TL;DNR step-by-step guide:
-
Upgrade the EAP650 to latest official firmware (in my case: 1.1.0 Build 20240830 Rel. 50826)
-
Go to: Site -> Settings -> Wireless Networks -> WLAN and click the Edit button for your affected SSID. The change is not required for other SSIDs. In my case, it is only my IoT-WiFi.
-
Scroll down to Multicast/Broadcast management. Uncheck "Multicast-to-Unicast conversion" for IPv4 and IPv6. This is enabled by default.
-
Hit "Apply".
-
Wait for the configuration to be applied by Omada controller. (Should not take more than a minute)
-
REBOOT YOUR AP!
Had no more issues since. After ~10 days I re-enabled this particular setting and the issue reoccured after just 3 days. Unchecked it and it is again running stable for 10 days now. Before, I had this issue even with the beta 1.0.91 every few hours on some days.
Here is some explanation for this issue (the long version):
I have 3 SSIDs running, separated to VLANs. A firewall routes between those subnets. To make IoT devices work, a mDNS responder is in place to forward mDNS packets between user net and IoT net. I noticed multiple times, that the issue occured exactly that moment when some Apple devices came back home and connected to WiFi. Of course it hasn't always been this particular situation. Most of the time, the issue occured "randomly", but these moments gave the needed hint for research. While searching the internet for information about multicast-to-unicast conversion in general and mDNS, I discovered this analysis by a network professional on Apple and Apple's mDNS usage:
https://youtu.be/miRV8qDOKBE?si=xBU9nFK4YpOVUkc9
After that, I tried disabling this particular feature and voilá, everything running perfectly fine now. I think the multicast-to-unicast conversion feature is broken on this device series. This might not be the fault of TP-Link: I had a Synology WRX560 before (only for a few weeks), which had the exact same issue. I remember that it had the same WiFi chipset as the EAP650 has. So this is possibly some chipset series issue. I did not have to disable the multicast-to-unicast conversion on my other SSIDs, which only have "proper" devices with both 2.4/5GHz abilities and a little more compute power than the little IoT bas***ds ;-)
Hope this helps you, too! That was quite a journey.
- Copy Link
- Report Inappropriate Content
@Arne17 Thank you very much for this write-up and solution! I have the exact same setup with a separate IoT network and mDNS. I'm trying this now and will let you know if it works.
- Copy Link
- Report Inappropriate Content
@Arne17 Unfortunately, this did not solve the problem for me. If anything, it caused the issue to manifest itself even quicker. Switching to WPA2 and PMF = Off seems to be working for me so far. I'll continue to monitor and report back.
- Copy Link
- Report Inappropriate Content

Information
Helpful: 1
Views: 3890
Replies: 56
Voters 1
