Best aimesh settings for stable roaming teams, messenger calls

[benganb]

I have problem with disconnected calls on Teams and Messenger when moving PC and phone between the two nodes during a call. Should roaming work without disconcerting?
AC86U as main and AC68u as node, use the latest FW, ethernet backhaul. They are located with large overlap. Use the same SSID on both bands.
Whats the best setup of Aimesh?

 

[sbsnb]

Use the same SSID on both bands.

You mean you have 2.4 GHz and 5 GHz with the same SSID? Have you tried giving them different names? I have experienced issues with devices hopping back and forth when the bands have the same SSID. Unfortunately, it seems that switching bands is not seamless, at least on my devices.

 

[L&LD]

Your signature indicates that you're on 386 RC2. I suggest you flash RMerlin v386.3_0 release final today.

Don't use the same SSID between the two radio bands. Don't use Smart Connect. Use a fixed Control Channel (you will need to test which one works the best in your environment).

You may need to do a full reset to factory defaults on both routers and do a minimal and manual configuration if the above changes still give you the same issues.

Even with all the above properly executed, the devices are the ones that decide how seamless they will roam. What devices do this to you?

 

[benganb]

You mean you have 2.4 GHz and 5 GHz with the same SSID? Have you tried giving them different names? I have experienced issues with devices hopping back and forth when the bands have the same SSID. Unfortunately, it seems that switching bands is not seamless, at least on my devices.

Try to avoid different names, then I need to change settings on every device. But ok maybe should test a day when I'm alone and have teams meetings.

 

[benganb]

Your signature indicates that you're on 386 RC2. I suggest you flash RMerlin v386.3_0 release final today.

Don't use the same SSID between the two radio bands. Don't use Smart Connect. Use a fixed Control Channel (you will need to test which one works the best in your environment).

You may need to do a full reset to factory defaults on both routers and do a minimal and manual configuration if the above changes still give you the same issues.

Even with all the above properly executed, the devices are the ones that decide how seamless they will roam. What devices do this to you?

Old signature.....
Smart connect is disabled and I use fixed channels.
Devices which struggle is PC Windows 10 and Samsung Galaxy phone.

 

[sbsnb]

Try to avoid different names, then I need to change settings on every device. But ok maybe should test a day when I'm alone and have teams meetings.

You only need to change the devices that you want to connect to the "other" band. If most of your devices connect to 5 GHz, leave that name the same and change the other to add _24G or something.

 

[CaptnDanLKW]

I use SmartConnect (therefore Same SSIDs) and AI Mesh 2.0 with Merlin 386.3 and both Facetime and Teams calls roam seamlessly, meaning I don't *perceive* any interruption. Not sure able FB Messenger.

Some apps have robust error correction and resiliency in the application layer and don't lean heavily on network or transport layer. In my experience, on the professional tab of both bands, you want to keep Roaming Assist DISABLED. This RA setting is like a bouncer, it will kick the station off, essentially severing your Layer 2 link, which drops your IP, and forces a complete session reestablishment.

RA doing its think is akin to toggling airplane mode. No apps can survive that.

That's my $0.02, as always YMMV

 

[binary_lifestyle]

I experienced the same annoying issue with call drops for 10-15 seconds and disabling NAT acceleration finally enabled smooth roaming for me without call drops:

Hello @HBTA, I experienced the same issue with 10-20 seconds delay after switching between AiMesh nodes (RT-AC68U + RP-AC55). The interesting part is that ping monitoring utilities (on both iOS, Android and just ping on Windows) worked perfectly during the node switch, at most one packet 100ms delay, and also Wi-Fi indicator transition from weak signal to strong signal instantly, but all the application traffic got stuck during this time, causing calls loss. Also what I noticed (maybe just my feelings and not the real case), that if browsing Web via browser during this time, jumping around the pages on the website you have opened already got stuck, however going to the new tab and opening website you did not have opened with the old AiMesh node works fine without any delays. This made me thinking that there is something happening with TCP-connections.

Solution that worked for me:
After trying many different things from this forum, I found this:

Disabling NAT acceleration mentioned there did the trick and now transition works perfectly without any hiccups during calls. This gave instant result unlike Wi-Fi options tuning like channels, Airtime Fairness and Universal Beamforming toggling.

Uneducated guess which can be wrong:
It seems to be that NAT acceleration for TCP and UDP is doing some optimizations that are failing during transitioning from one AiMesh node to another. And ping uses ICMP protocol, which does not have ports and is processed differently by NAT, maybe that is why ping works good during node switch unlike TCP.

 

[Lynx]

So @beganb did you resolve this?