Speed Test Discrepancy

[TheLyppardMan]

I'm on BT's Full Fibre 500 service and if I test speeds on my laptop using an ethernet connection, the results from the Ookla speed test site indicate approx 500-503Mbps download, but if I run a speedtest from my RT-AX88U, either manually or via the spdMerlin script, I see much lower values and I'm just curious as to why that may be. I don't have QoS enabled at the moment, so that can't be the reason.

 

[Ellenswamy]

Same with me, I have Up to 1.2 and 40 up. I do a speed test on the router and get 500-600 down. But on my MacBook I get 700+ down over wifi which is expected so I just ignore what the tour says now.

 

[Paliv]

For higher speed tiers you are CPU limited by the router. I have 1200/35 and I occasionally will get that speed on an AX86U, but usually it’s lower. It’s better to test single devices, or aggregate of multiple devices if you are on a faster than gigabit plan.

 

[TheLyppardMan]

For higher speed tiers you are CPU limited by the router. I have 12000/35 and I occasionally will get that speed on an AX86U, but usually it’s lower. It’s better to test single devices, or aggregate of multiple devices if you are on a faster than gigabit plan.

Thanks for that explanation. Much appreciated.

 

[TheLyppardMan]

I'm on BT's Full Fibre 500 service and if I test speeds on my laptop using an ethernet connection, the results from the Ookla speed test site indicate approx 500-503Mbps download, but if I run a speedtest from my RT-AX88U, either manually or via the spdMerlin script, I see much lower values and I'm just curious as to why that may be. I don't have QoS enabled at the moment, so that can't be the reason.
View attachment 39649

I've changed the firmware from Merlin to the latest ASUS firmware and rebuilt everything from scratch. Speed tests from the router are now giving this:-

I have also switched on Adaptive QoS and set the speeds to 475/65. If I run a speed test from my laptop using an Ethernet cable, the speed test results on the Ookla website now remain just inside the limits I have set, so everything seems to be working as expected now.

 

[Tech9]

Some of the scripts you used with Asuswrt-Merlin impact the built-in speed test. I told you in another thread I would not use Skynet for specific (not speed test related) reasons. I'll tell you in this thread I would not use spdMerlin as well - it only stresses your ISP line on regular intervals to show you a graph. How many times a day you need to check your ISP speed and for what reason? By the way, you may not need QoS at all. Keep it disabled until you need it.

 

[TheLyppardMan]

Some of the scripts you used with Asuswrt-Merlin impact the built-in speed test. I told you in another thread I would not use Skynet for specific (not speed test related) reasons. I'll tell you in this thread I would not use spdMerlin as well - it only stresses your ISP line on regular intervals to show you a graph. How many times a day you need to check your ISP speed and for what reason? By the way, you may not need QoS at all. Keep it disabled until you need it.

Thanks Tech9. I've just disabled QoS as you suggested. Was it also your good self who suggested that the ASUS stock firmware was your preferred choice? I'm sure I did read something along those lines, which prompted me to give the stock firmware a try, and based on your advice I decided I could live without Skynet, automatic speed tests, etc. The led script was useful, but I've just turned them off with the switch on the router (which is near a bedroom, so I need them to be off at night at the very least). The only other change I have had to make is to switch on UPnP if I want the Networx network traffic graph to monitor all the network traffic, rather than just that passing to and from my laptop.

 

[Viktor Jaep]

QoS is used to prioritize network traffic in times of congestion. If there is no congestion, QoS doesn’t do anything. With 500Mb, I doubt you'll see any congestion, ever. Also, as @Tech9 mentioned, these speedtest utilities are nice-to-haves, but difficult to rely on when running from the router itself due to resource and CPU contraints. It's better to hook up an external device wired directly into the router itself (or over wifi) and running a speedtest from that, which will be much more accurate.

 

[Tech9]

Was it also your good self who suggested that the ASUS stock firmware was your preferred choice?

I only test some of the Asus routers in my routers collection when new firmware is available. Usually I pick AC68U and AC86U because of the smaller footprint and popularity. One is older ARMv7 example, the other newer so-called HND ARMv8 example. AX AiMesh I test with AX88U and AX58U. I don't use any of them on my network. My Wi-Fi system is completely different. In my experience stock 386 base Asuswrt is more stable than any available 386 Asuswrt-Merlin versions, especially if you run AiMesh setup, more so if you run mixed routers AiMesh setup with AC68U variants involved. When I test routers I make sure they are tortured well and so far Asuswrt holds better. I would skip entirely 386.4 and wait for fixed 386.5, if Asuswrt-Merlin is what you are after. Also, I would not use USB stick for scripts, but a small SSD in external USB enclosure. Read carefully what the scripts do and how. Some scripts may impact your Internet use experience. Don't install scripts just because someone else is running something on his router.

If you want to tinker with your router - Asuswrt-Merlin
If you want to have more free time in your life - Asuswrt

 

[TheLyppardMan]

Thanks Tech9. The second option seems the better choice, so I think I'll stick with that, for now at least.

 

[Tech9]

I'm sure Asus fans will jump in right away to defend their hardware and software choices, but you asked me and I'm giving you my opinion. The router is yours and what you run on it is entirely your decision. When I find something not properly working I just reset the router and put it back on the shelf. When you encounter issues after too much of tinkering it affects your entire family.