Actual performance vs performance test huge difference
I am running into a challenge with my AmplifiHD, running the latest non-beta firmware. My challenge is that in the current working from home situation I do conference calls a lot. And I have noticed significant performance challenges and 'bad connection'.
I have used the App to check on the devices page and there I see that my laptop requires around 7Mbps to run a call fluidly. What is puzzling me is that when I use the same App to check the performance (from my iPhone) it shows download 160.7 Mbps and Upload 42.5 Mbps. So there should be enough bandwidth to go around.
What I notice is that when my daughter uses her iPad to play an online game is also shows around 7Mbps in the App. When my laptop and the iPad are connected to the same mesh point, my conference call will fail. When she turns off the iPad, I can do calls perfectly. So apparently it seems that the 7 Mbps + 7 Mbps is too much to handle, but in the performance test is manages to get well above that.
If at any time I look in the App at the main page to see the number of connected devices and the total up/down speed it very rarely comes above 10Mbps, and all devices seem to compete for bandwidth that should be is large supply.
The mesh point band is set to 5GHz and connected devices should also not be limiting (new HP laptop, 2 year old iPad, AppleTv, chromebook and a bunch of iPhones).
Would be awesome if someone here can help me out. I must be overlooking something!
UI-AmpliFi last edited by
@Marc-Verdonk Based on your findings, it is doubtful that ISP bandwidth is the culprit here. You said when the iPad and your laptop are connected to the same MeshPoint, your calls will fail. If you are connected to the primary router or another MeshPoint with your laptop are you facing these same issues? What if other devices, more than just her iPad are being used?
What is the connections strength of the mesh point? (Tap on the mesh point in the app to view this data) For performance, I recommend anywhere between -55 to -60 dBm.
If you run a speed test using a 3rd party site (Fast.com or speedtest.net) while your laptop is connected to the MeshPoint, what speeds does your laptop get? If the MeshPoint overall is not providing great speeds because of interference or because of distance from the router, then it is possible that the 14Mbps both devices consume could be creating a bandwidth issue in that specific location.
@UI-AmpliFi Thanks for your response and you are definately on to something here. I noticed the challenge did not happen when my daughter was connecting to the primary router.
So this is what I tested:
- moved laptop down, connected to primary router and did speedtest. 200Mbps down and 50Mbps up.
- moved laptop back up, connected to mesh point and did speedtest. 9Mbps down, 20 Mbps up.
According to the App this mesh point has a signal level of -51dBm. which seem to be 'Good' and it operates at the 5GHz band.
But, when I look in the App on the homepage it actually says 'Primary Router' connected to "Meshpoint 1" at -71 dBm and then Meshpoint 1 connected to Meshpoint 2 at -51dBm.
So it seems that Meshpoint 1 is the culprit and that is where I am loosing the bandwidth. I will see how I can fix this as the meshpoint is actually pretty close to the Router, but it is in an area with a lot of electrical wiring.
Thanks for helping me think through this. I think I am one step closer to solving this. Please let me know if you think the above makes sense and that looking at Meshpoint 1 is the likely solution.
@Marc-Verdonk Update: I moved the meshpoints around and made sure the connected in the correct order to the primary router. Both are now 'Good' at the -50 dBm level.
Results are that my laptop connected to second mesh point now achieves 70Mbps down and 25Mbps up, which is a huge improvement over the 10/20 that it was earlier. So this should have solved the problem.
Only thing that still puzzles me is why my iPhone X, connected to the mesh point in the old situation was still able to achieve 160/42. Which is why I did not think to look at the meshpoint performance. But since this problem is now solved I will learn to live with that mystery.
Thanks for your help!