Jitter issues

[Adrian4]

Hi,

I have an office site with about 20 or so MR44's. They have been complaining for a long time about VOIP issues that has proven very hard to solve as things generally look ok whenever you try to investigate and the problems are extremely intermittent but regular enough to cause complaints.


While trying to troubleshoot, I was sat int eh office about 7pm last night with everyone gone so I could make config changes. While I was doing this I was making lots and lots speed tests using speed.cloudfare.com and one or two other tests. I did this moving around all over the office connecting to different AP's.

I notices that I was consistently getting Jitter of between 20 and 40. Results below 30 were the least common and it was around 40 most of the time. My up and download speeds were consistent every time.


There was no one else in the office, the office in on large site we own which means there are no neighbours - Our WiFi is the only WiFi in range.

I even switched SSID to one we have that doesn't go through firewalls or anything, just straight out to the internet and that as giving me the same results!!!!

Where is this Jitter coming from??!?! I dont understand 😞

[Adrian4]

I'm now at a different site on the other side of the country, so different internet circuits and I'm still getting about 40ms Jitter consistently!!! I have a intel AX201 with fully up to date drivers.

How can it be so consistent no matter which site I am at !?!

[alemabrahao]

Have tried creating a traffic shaping rule for Voip on the SSID?

[Adrian4]

yea, although its a bit of a grey area. First of all the VOIP us called "Fuze" which isnt common enough to be in the standard list so I had to rely on their tech docs regarding how they tag etc - turns out they tag wrong lol so put in a rule which I think would work but its difficult to tell because its not Meraki all the way to the gate - we have some old Avaya switches (and Forcepoint firewalls) and I'm not sure if they honour the tags.

However, the speed test I was running were just on web based tools - no VOIP involved and I was getting a consistent 30-40ms jitter no matter where I was or how much network traffic there was.

[BrandonS]

Kind of a wild guess, but I saw similar with a new customer recently that has Meraki.  They had 8 MR's in an open office area that were about 30-50' away from each other.  They had inconsistent performance and poor Zoom calls quite often.  It was kind of obvious to me there was probably co-channel interference, but I verified it with a spectrum analyzer and turned the power down on all MR's to the minimum and haven't heard a complaint since. I also turned off most of the 2.4GHz radios for good measure.

All that to say, maybe co-channel interference if you have too many MR's near each other with power set to auto or too high?

[Adrian4]

Hiya, the work I was doing when I noticed it was manually setting the channels so there is no interference 😛

[PhilipDAth]

Were you testing using a battery powered device?  If so, this is pretty normal and expected behaviour due to client power saving.

If you are testing with a Windows device you can disable WiFi power saving with these commands:

powercfg /SETDCVALUEINDEX SCHEME_CURRENT 19cbb8fa-5279-450e-9fac-8a3d5fedd0c1 12bbebe6-58d6-4636-95bb-3217ef867c1a 0
powercfg /SETACVALUEINDEX SCHEME_CURRENT 19cbb8fa-5279-450e-9fac-8a3d5fedd0c1 12bbebe6-58d6-4636-95bb-3217ef867c1a 0

[Adrian4]

it was indeed a laptop on battery! this is really interesting because I went home and tried on my home network and got about the same result - and then I tested on my personal laptop at home and got the same result and was about just just give up lol

I will play around with settings and see what happens, thanks!

[Adrian4]

so gave it a go, as well as turning off a power saving option in driver options - no effect.

Annoyingly - On my home network, running on wall power and connected via eth cable no wifi - the jitter is all over the place. I'm just running the test over and over and getting a different result each time. Nothing else is on the network using any bandwidth. I have a 300Mbs fiber circuit.

My best test was 7ms, but most around 50 and some as high as 100. There just isnt anything I can see that is changing between tests yet the values are so far apart!