RT-AX58U - increase in speed even for AC-only client

[eldridchapman]

Hi All,

I purchased an Asus RT-AX58U router to replace the RT-AC66U B1 as main router. The AC66U in turn became an Access Point at another part of my house. My internet is a 1GBPS connection.

What I was surprised was the significant increase of download and upload speed I observed from my AC-only client. On AC66U, I get roughly around 300mbps upload and download speed for the 5ghz band on my MacBook Pro 2017 (which support only AC and not AX).

On the new AX58U, the fastest I hit was 600mbps download and 800mbps upload on 5ghz band. I repeated the tests with latest firmware on both devices and it gave me similar results (+/- 150mbps or so). I saw similar gains on my iPhone XS as well. Do anyone know why the speed for AC clients improved with AX58U compared to AC66U B1?

 

[Deleted member 22229]

Better hardware much faster processor.

 

[L&LD]

The benefits of an AX router are not just for AC clients. The improvements are real, even if yours are exemplary.

 

[zackattack784]

Are you using 160mhz channels on the AX58U?

 

[eldridchapman]

Thanks everyone for your replies. I certainly understand that wifi speed is now dependant on the number of radios and hardware instead of just meeting the AC/AX standards.

Are you using 160mhz channels on the AX58U?

Yes, I have configured 160mhz band on the AX58U. Though my Macbook can only use 80mhz max.

 

[thiggins]

It's good that you saw a performance improvement. But moving from an AC to AX router doesn't always guarantee improved performance, as these test results show.
https://www.smallnetbuilder.com/wir...ifi6-peek-asus-rt-ax88u-netgear-rax80?start=2

It's particularly curious that you saw an improvement since you moved from a three-stream AC router to a two-stream AX.

What link rate was your MBP reporting when connected to each router?

If you were using internet based speed tests, it's possible that improved routing throughput WAN <> LAN contributed to the higher throughput.

 

[eldridchapman]

Thanks for your response. I was using Speedtest.net for testing my speed. All tests were conducted to the same host server (the fastest local server i found).

To respond to your query, I repeated the testing over the same time period (multiple times each router, best result taken) and maintained approx. same distance to both routers. However, the AC66U no longer have a wall between itself and my macbook as I've relocated it in my home. Pls find the results below:

AC66U B1

Macbook PHY rate: https://imgur.com/veovd6S

AX58U

Macbook PHY rate: https://imgur.com/9bDbmVF

PS: When i averaged all the results, I did notice the AC66U managed to match the download speed of AX58U (500+ mbps) but the AC66U's upload speed never exceeded 300mbps (while AX58U's average upload was 500mbps). Looked like a hardware-side limitation there.

 

[thiggins]

Thanks for posting the additional info. Your PHY rate screenshots show the same Tx rate, which would apply to uplink traffic.

Without running a test totally on the LAN side, it's impossible to determine whether the uplink throughput gain is coming from Wi-Fi or routing.

But at least you got improvement!

 

[K-2SO]

Not necessarily AC vs AX effect. I had similar jump going from AC68U to AC86U. Newer radio tech perhaps. And faster CPU too.

 

[thiggins]

Not necessarily AC vs AX effect. I had similar jump going from AC68U to AC86U. Newer radio tech perhaps. And faster CPU too.

Yeah, I keep hearing the argument that faster main CPU = improved Wi-Fi. But I don't think the two are necessarily related.

WiFi SoCs handle all the Wi-Fi to Ethernet conversion, so the main CPU load should be minimal. I agree that faster processing in the radio SoC, could improve performance. But then again, a beefier CPU (and more on-board memory) could have all it can do to handle the increased processing load imposed by more complex airtime and protocol management.

AC68U was three stream, AC86U is four (on 5 GHz). More streams definitely helps performance.

 

[K-2SO]

But I don't think the two are necessarily related.

They are in some cases. Specific firmware options disable NAT acceleration. Fully or partially. AC68U can process about 250Mbps. AC86U is good to about 400Mbps. In plain vanilla setup CPU does't affect WiFi so much. The fact is newer tech with more streams helps. Not related to AC vs AX support, IMHO.

 

[K-2SO]

BTW how is this Macbook getting >866Mbps link speed with 2x2 router and AC? Where the 3rd stream came from?

 

[thiggins]

BTW how is this Macbook getting >866Mbps link speed with 2x2 router and AC?

AC68U is three streams and so is MBP adapter. Link rate for MCS8 with 0.8uS guard interval is 1053 Mbps

 

[K-2SO]

AC68U is three streams and so is MBP adapter. Link rate for MCS8 with 0.8uS guard interval is 1053 Mbps

OK, but RT-AX58U is a 2x2 router. What I'm missing in this picture? NSS3 on AX58U, how?

 

[L&LD]

Should be possible?

https://docs.google.com/spreadsheets/d/1Ckfrx8lHCQuYKEiu6F-XUFAygobckSdFLTwGj1EpE_Y/edit#gid=0

 

[K-2SO]

Should be possible?

Doesn't look like it. Where the 3rd stream came from on AC?
RT-AX58U specs - https://www.asus.com/ca-en/Networking/RT-AX58U/specifications/ - "802.11ac : up to 867 Mbps"

 

[L&LD]

See H69.

 

[K-2SO]

See H69.

H69 is 867Mbps. Only TurboQAM can get higher with 2x2. AX58U doesn't have that on AC. See Asus specs.

 

[thiggins]

OK, but RT-AX58U is a 2x2 router. What I'm missing in this picture? NSS3 on AX58U, how?

Sorry, I was mixing up your AC68U to AX86U info with the OP's AC66U to AX58U.

That's because the AX58U 5 GHz radio is 4x4. The FCC photo shows four 5 GHz RF chains. ASUS appears to be playing games with the specs, claiming it's a 2x2 router.

I have an Intel AX200 associated with an AX58U right now. AX58U has 160MHz enabled and AX200 is showing 2.4Gbps "speed" in Windows Wireless properties.

 

[L&LD]

I am setting up an RT-AX58U myself and that is what I get too. AX200 adaptor shows a 2.4Gbps connection rate.