Amplifi HD wifi capping at 450 Mbps

  • I would expect the tests at would be near 900+ Mbps.

    When I plug my Mac's Ethernet directly into the ISP router or into one of the 2 Amplifi HD routers, shows 950 Mbps, as I would expect.

    I have 2 Amplifi HD Mesh Routers that are both wired into my ISP's router. It seems that one Amplifi Router is in Bridge Mode and the other has Ethernet Backbone enabled, which I think is the right configuration.

    One odd thing that I notice is that when I do the performance test from the Amplifi iOS app, the results are reported around 200 Mbps, which is different from what I see when the Mac is plugged into the Ethernet port of the Amplifi Router that has the Ethernet Backbone enabled.

    I don't think the issue is with my ISP or with the ISP router, as I am seeing the expected speeds when over Ethernet.

    I've tried the wifi speed tests with a number of devices, including a 2019 MacBook Pro, so I don't think the issue is there either.

  • Hi @fmsf - is your wireless client operating with 3 MIMO chains on a clear 80 MHz channel?

    The best theoretical single wireless client speed in one direction in ideal conditions with no channel contention is 1300 Mbps (MCS9) using 3 MIMO chains on a crystal clear 80 MHz channel

    If your wireless client is only using 2 MIMO chains, then the max theoretical speed drops by 33% to 867 Mbps

    A 40 MHz channel width would cut theoretical WiFI speeds in half compared to an 80 MHz channel

    Channel contention with other clients or mesh points, interference, signal quality, airtime fairness, latency, response frames, etc., can all cause drops in throughout, from slight amounts to big amounts

    So a lot also depends on the environment you are testing in

    I believe all iOS devices still max out at 2 MIMO chains, but I am not familiar with the latest Apple gear

    It sounds like most of your testing might be on 2 MIMO chain capable clients

    If you are absolutely sure you are using a clear 80 MHz channel then you would need to scan the environment to rule out what might be causing the additional drop in throughout

    Or you might have a marginal 80 MHz channel width, or dropping down to a 40 MHz channel width, which could be putting you in that 450 Mbps range when testing on a 2 MIMO client maybe?

    Also note that AmpliFi HD's do not have MU-MIMO capability to clients, and WiFi-5 is SU-MIMO only on the up-link to access points, so any other wireless clients in the environment doing any communication at all requires the access point to stop and service requests from clients other than the one doing the speed testing

    Even if a client appears inactive, some communication still happens in the background

    While the AmpliFi HD is technically capable of 1300 Mbps, the client and the environment need to match

    And, since the HD's only have 1 Gbps Ethernet WAN/LAN ports, there is no technical way for them to ever process that speed, effectively capping them to 1000 Mbps

    So if you do happen to have a 3 MIMO chain capable client operating at the highest speed, the environment might only accommodate a 40 MHz channel width or a degraded 80 MHz channel width at those speeds, which also places you in that 450 Mbps range considering packet overhead

  • Hi, it is a clear 80 MHz channel connecting 802.11ac. I'm not sure about the number of MIMO chains, and I can't seem to find the spec. That may be the issue with these 2 machines. I'm not sure how to verify the number of chains. Any thoughts?

    The issue with the Amplifi iOS app performance test has nothing to do with that though, I think.

  • Hi @fmsf - if you you go into the AmpliFi app, select the Devices page, click on a client, then select Client Details, you will see the number of MIMO chains in use and the Tx & Rx Link Rates that the client has negotiated with the router on that channel

    The reported Link Rates will be the maximum theoretically achievable in your environment at that moment based on the signal quality available between the client and the router with no other clients in contention for bandwidth in proximity on the channel in use

    If you run a reliable bandwidth test, you should be able to achieve close to the Link Rates (minus packet overhead and latency) as long as no other clients contend or interfere with the channel in use for the test

    If results are significantly lower and your testing client is the only one communicating on that channel, then you can look at enabling the Switch Buffers in the web UI, and testing with the A-MSDU option on and off in the web UI, which has caused some problems with Apple device testing in the past

    One other thing to investigate is if the MTU setting that AmpliFi is using is optimal for your ISP

    If the MTU is causing packet fragmentation throughput will be degraded

    Your Ethernet clients are also able to take advantage of HW NAT if it is enabled, which I believe doesn't apply to WiFi clients

  • Hi @Derek-Saville - thanks.

    For the 13" MacBook Pro, the iOS app shows just 2x2 MIMO chains, so that explains that one. But for the iMac, it shows 3x3, and the Tx and Rx bitrates are both 975 Mbps. These are both connecting with 802.11ac, but shows 400 Mbps for the 2x2 MacBook and the 3x3 iMac.

    In the web UI, I don't see any options for Switch Buffers or to see what the MTU setting that Amplifi is using. It seems that my ISP expects 1500. I'm using Bridge Mode, which often causes the UI to not show everything.

    I do have the A-MSDU option enabled. I'll try it with that off.

    When connecting via Ethernet, I get 950 Mbps, so the throughput into the ISP is fine.

  • Disabling the A-MSDU option doesn't impact the speed. Still around 450 Mbps.

  • Hi @fmsf - unfortunately I no longer have an HD running to see what options are available in the web UI with the latest firmware release

    Some options only show up in Beta firmware, and the Switch Buffers may have been permanently implemented now
    (in Bridge Mode HW NAT would not exist since no NAT is taking place)

    One thing I would recommend trying is taking your 2nd HD RAMP offline and running the WiFi speed tests with only the bridged router, just to be sure the the RAMP itself isn't causing any interference

    Otherwise, after running some tests, generate Support Info files and e-mail them to @UI-Brett (BRETT.BENNETT@UBNT.COM) along with the details of your ISP and modem

    They might be able to see something and/or compare the telemetry with results from other users on the same ISP setup

  • Hi @fmsf - do you have IPv6 enabled or disabled?

    Some ISP's require IPv6 for higher bandwidth service, so you might check that as well, just in case...

  • Seeing as the Ethernet connected Mac gets 950 Mbps, I don't think there's a problem with IPv6.

    I also found this Gear Primer review that shows the AmplifiHD getting 482 Mbps with a 3x3 MacBook Pro, so maybe this is just the real-world limit of these routers. That would be too bad.

  • Hi @UI-Brett - what are the highest AC WiFi throughput results you have achieved on an HD in the lab with 2x2 & 3x3 MIMO on 40 MHz and 80 MHz channels?

  • @Derek-Saville I do not know lab results, but real world throughput is likely not going to exceed 500Mbps.

  • Same problem. Rarely see speeds over 280 Mbps on 5G even in the same room as router. On the 2G SSID the speeds have dropped as low as 11 Mbps I don't even have mesh points. Been trying to resolve with tech support for over a month. They are now telling me that Gigabyte speeds on AmplifiHD are only possible with a wired connection. Why would I use a WIRELESS router if I need to be directly connected by a cable? Plus I already have an additional port on the modem if I wanted to do that.

Log in to reply