Internet Speed Test, who can you believe?

[RealBlackStuff]

Yesterday I got FTTH (Fibre To The House) installed from Vodafone, with 500/50 Mbps (Down/Up).

I tested the above with my T440p (W7 Pro/SP1, Intel AC-7260 wifi-card) on the kitchen table.
The Vodafone Gigabox router is currently in the adjacent living room, ca. 5m/15ft away and there is a concrete wall between these rooms.

The router will be moved to the central downstairs hall as soon as I have a Cat6 ethernet cable.
I will also redo these tests with my T440p, but then connected via ethernet cable.

To test internet speed there are quite a few options/programs, but their results vary hugely!
From 9 different websites, I got all different results: (Down/Up Mbps)
https://www.speed.io/ Result: 24.6/51.7
https://testmy.net/ Result: 25/25.7
https://speedof.me/ Result: 33.24/34.3
https://speedsmart.net/ Result: 78.38/52.05
https://www.broadbandsearch.net/speed-test Result: 182.1/51.0
https://fast.com/ Result: 280/55
https://fiber.google.com/speedtest/ Result: 326.7/51.5
https://www.earthlink.net/resources/int ... peed-test/ Result: 393.2/51.4
https://www.speedtest.net/ Result: 402.80/51.72

Most ISPs are using the (Ookla) speedtest.net, but seeing the above, who can you believe?

Anybody have any experience with this Vodafone Gigabox router?
It appears quite basic, especially since my previous (copper wired) ISP used a FritzBox 7530 router.

[axur-delmeria]

The results of the ookla test vary depending on which server you connect to. Usually, the farther away it is, the slower it becomes, because you'll have to contend with more "hops" across the Internet's various nodes and interconnects.

[theterminator93]

Once you start getting faster connection speeds, the commercial speed test websites become less reliable.

When we test connections at the schools, we use iPerf3 and one of the listed servers that's as geographically close as possible. It's more consistent when testing connection speeds that approach or exceed 1 Gb/s.

[ajkula66]

Once you start getting faster connection speeds, the commercial speed test websites become less reliable.

This.

Yesterday I got FTTH (Fibre To The House) installed from Vodafone, with 500/50 Mbps (Down/Up).

I tested the above with my T440p (W7 Pro/SP1, Intel AC-7260 wifi-card) on the kitchen table.

What are you getting with a wired connection ?

[RealBlackStuff]

Results using WiFi-Black, Ethernet-Red (Down/Up Mbps):
https://www.speed.io/ . . . . . . . . . . . . . . . . . . . . . . Result: 24.6/51.7 146.8/51.4
https://testmy.net/ . . . . . . . . . . . . . . . . . . . . . . . . Result: 25/25.7 30.0/26.6
https://speedof.me/ . . . . . . . . . . . . . . . . . . . . . . . .Result: 33.24/34.3 36.85/36.38
https://speedsmart.net/ . . . . . . . . . . . . . . . . . . . . . Result: 78.38/52.05 78.85/52.13
https://www.broadbandsearch.net/speed-test . . . . . . .Result: 182.1/51.0 459.4/51.4
https://fast.com/ . . . . . . . . . . . . . . . . . . . . . . . . . .Result: 280/55 470/53
https://fiber.google.com/speedtest/ . . . . . . . . . . . . . Result: 326.7/51.5 477.2/51.4
https://www.earthlink.net/resources/int ... peed-test/ Result: 393.2/51.4 478.4/51.5
https://www.speedtest.net/ . . . . . . . . . . . . . . . . . . .Result: 402.80/51.72 464.84/51.51

The above were all measured using the Vodafone Gigabox Sh!tbox router.
It sends out only ONE SSID for both 2.4Ghz and 5GHz combined and has no settings to modify that!
As a result, our Amazon Firesticks are buffering on nearly all channels, with the exception of maybe Pluto, which we hardly ever use.
It also wants to force a mobile app down your throat to change wifi details using "Super wifi".
I'll replace that Gigabollox next week with my trusty Fritz!Box 7530, once I have the Vodafone login details.

[JBUK]

When you do a speedtest using Wifi you are really limited to the speed between the device and your router. There are a lot of variables at play with two band routers which are not independent of each other as the speed is limited to that of the slowest connected device. Unless you are using recent wifi technology you will not get speeds near to 500Mbps anyway. Whereas gigabyte ethernet will always be testing directly to the router and get an accurate measurement as long as the speedtest you use will actually measure up to 500.
The BT routers supplied here are the same as yours with both channels combined. Mine is still on the shelf. Your Fritzbox 7530 is hopefully better in this respect.

[ajkula66]

The above were all measured using the Vodafone Gigabox Sh!tbox router.

I just tested my wi-fi connection using some of the tests you've linked, and am getting pretty constant results (with a crap 6205N card) on most of them, so I'd say that your router could be at fault.

It sends out only ONE SSID for both 2.4Ghz and 5GHz combined and has no settings to modify that!

Yep, I've seen that nonsense before, the customers were up in arms all over the place and a firmware update was released to allow for separate SSIDs.

I'll replace that Gigabollox next week with my trusty Fritz!Box 7530, once I have the Vodafone login details.

Fingers crossed that your experience improves after that...

[RealBlackStuff]

The FritzBox 7530 router has been installed and used now for a few weeks.
Test results using the same T440p here:

Results using WiFi (Down/Up Mbps):
https://www.speed.io/ . . . . . . . . . . . . . . . . . . . . . . Result: 228.6/51.6
https://testmy.net/ . . . . . . . . . . . . . . . . . . . . . . . . Result: 26.9/23.7
https://speedof.me/ . . . . . . . . . . . . . . . . . . . . . . . .Result: 31.73/30.85
https://speedsmart.net/ . . . . . . . . . . . . . . . . . . . . . Result: 108.1/51.82
https://www.broadbandsearch.net/speed-test . . . . . . .Result: 242.8/51.6
https://fast.com/ . . . . . . . . . . . . . . . . . . . . . . . . . .Result: 230/54
https://fiber.google.com/speedtest/ . . . . . . . . . . . . . Result: 225.2/51.5
https://www.earthlink.net/resources/int ... peed-test/ Result: 228.4/50.8
https://www.speedtest.net/ . . . . . . . . . . . . . . . . . . .Result: 239.34/51.34
Couldn't be bothered wasting more time to test via Ethernet.

"We are not amused", as the saying goes.
I guess this Fritzbox 7530 is getting a bit too long in the tooth.
So I have ordered another router, Asus RT-AX86U / AX5700 which should be here next week.
Hoping for a LOT better results!

PS afterthought: since we only use mobiles (a.k.a. cell phones) there is no need for all that landline phone-stuff which is included in most Fritzbox routers, really a waste of space and electronics.

[RealBlackStuff]

My 'new' Asus RT-AX86U router has been installed downstairs in the same place and used now for a few days.
I bought it on eBay, same seller has another one: https://www.ebay.ie/itm/235378555993
As soon as I received and tested it, I installed my favourite Asuswrt-Merlin firmware on it.

I always was a fan of Asus routers and in hindsight that Fritzbox (a leftover from my previous ISP) was only mediocre, with a lot of unnecessary (phone-)ballast.
In my office upstairs I have an older Asus router, which is used as a repeater and has a printer attached via USB, so we can print via wifi.
That router was originally a cheap American T-Mobile TM-AC1900, which I converted back to its original Asus RT-AC68U with Asuswrt-Merlin firmware.
I used it for many years when we still lived in USA.

Anyway, RT-AX86U test results, using the same T440p from the same location, are below.
(For comparison I left the previous Gigabox and FritzBox 7530 figures in there.)

Results using WiFi (Down/Up Mbps): . . . . . . . . . . . . . .Vodafone Gigabox --- Fritz! 7530 --- Asus RT-AX86U
https://www.speed.io/ . . . . . . . . . . . . . . . . . . . . . . Result: 24.6/51.7 ---- 228.6/51.6 --- 400.6/51.7
https://testmy.net/ . . . . . . . . . . . . . . . . . . . . . . . . Result: 25/25.7 ------- 26.9/23.7 ---- 28/25.8
https://speedof.me/ . . . . . . . . . . . . . . . . . . . . . . . .Result: 33.24/34. ---- 31.73/30.85 --- 32.11/29.14
https://speedsmart.net/ . . . . . . . . . . . . . . . . . . . . . Result: 78.38/52.05 -- 108.1/51.82 --- 162.9/52.25
https://www.broadbandsearch.net/speed-test . . . . . . .Result: 182.1/51.0 --- 242.8/51.6 --- 398/51.8
https://fast.com/ . . . . . . . . . . . . . . . . . . . . . . . . . .Result: 280/55 -------- 230/54 ------- 310/55
https://fiber.google.com/speedtest/ . . . . . . . . . . . . . Result: 326.7/51.5 --- 225.2/51.5 --- 402.5/51.5
https://www.earthlink.net/resources/int ... peed-test/ Result: 393.2/51.4 --- 228.4/50.8 --- 414.2/51.7
https://www.speedtest.net/ . . . . . . . . . . . . . . . . . . .Result: 402.80/51.72 -- 239.34/51.34 --- 290.12/51.66

Not too shabby!

[dr_st]

A good router makes a ton of difference. For about a year I owned a TP-Link AC2100, and was too lazy to install it, because my older TP-Link W9970 seemed to have been working fine. Most of the time. At some point I bit the bullet, and lo and behold - WiFi speeds improved everywhere by ~100% and the intermittent issues with losing file sharing connectivity on the LAN went away completely.

[RealBlackStuff]

Found yet another speedtest, which gave me a surprising result.
Left is the BEFORE result, right is the AFTER result.
BEFORE: running the laptop (T440p) as-is.
Make 2 small changes (as seen below the picture), taking only a few seconds, to see the results
AFTER: major improvement!

- If you are using Windows 7/8/8.1 or 10:
It is important to verify that your current network connection is set as a private connection and not public.
Windows is limiting the performance of public connections.
If your network connection is already set up as private you can try to run these two commands
inside the Windows command prompt ("CMD", run as Administrator) to restore original performances:
netsh int tcp set glo autotuninglevel=disabled
netsh int tcp set glo autotuninglevel=normal

Before I forget, the above is the result from: https://real-debrid.com/speedtest

No idea what the settings were (because this laptop with W7/SP1 has not had any updates since 2018).
Guess I'll have to rerun the other speedtests.

Results using WiFi (Down/Up Mbps): . . . . . . . . . . . . . . . . . . BEFORE --- AFTER Real-Debrid
https://www.speed.io/ . . . . . . . . . . . . . . . . . . . . . . Result: 400.6/51.7 --- 428.3/51.5
https://testmy.net/ . . . . . . . . . . . . . . . . . . . . . . . . Result: 28/25.8 ------- 266.3/44.2
https://speedof.me/ . . . . . . . . . . . . . . . . . . . . . . . .Result: 32.11/29.14 --- 378.08/54.07
https://speedsmart.net/ . . . . . . . . . . . . . . . . . . . . . Result: 162.9/52.25 --- (not available)
https://www.broadbandsearch.net/speed-test . . . . . . .Result: 398/51.8 --- 432.0/51.4
https://fast.com/ . . . . . . . . . . . . . . . . . . . . . . . . . .Result: 310/55 ------- 520/50
https://fiber.google.com/speedtest/ . . . . . . . . . . . . . Result: 402.5/51.5 --- 445.6/51.5
https://www.earthlink.net/resources/int ... peed-test/ Result: 414.2/51.7 --- 464.2/51.6
https://www.speedtest.net/ . . . . . . . . . . . . . . . . . . .Result: 290.12/51.66 --- 452.01/51.46

[solidpro]

Here in Milton Keynes, we don't even need a router. Just a 1Gbps copper ethernet RJ45 dropped into our house. The engineer didn't know what he was doing so I blagged the combination to the cabinet and patched mine in myself - it's just a cisco switch patched out for every house in the street (there is only 14) with a fibre uplink to where-ever the 'exchange' or the next switch is. Supposed to be 1Gbps but obviously contention on a single switch, single uplink, and beyond is all variable. Of course, I use a firewall to provide security, NAT, DHCP, etc.

I worked for a big software company once that had enough money and were complacent enough about every ISP that they simply decided they would become an ISP so they could do their own internet.

[RealBlackStuff]

Been thinking about your copper cable.
If you already have fibre coming into that cabinet, I would also want fibre from there into my house!
Fibre is umpteen times faster than copper and does not care about other houses in your street/estate.
Something for a rainy sunny day!

[solidpro]

1Gbps copper is the same as 1Gbps fibre. I pay for 1Gbps.

I suspect the last 30 metres are copper due to the sharp turns around walls and around corners.

[ajkula66]

I suspect the last 30 metres are copper due to the sharp turns around walls and around corners.

That should make absolutely no difference whatsoever. Take it from someone who earned his retirement from telecom industry a couple of years ago. Fiber doesn't care about the bends any more than copper does. And yes, I've done both going back to 56Ks, ISDNs and T-1s...all the way to FiOS and HTech (10Gbps and up) with everything you can think of in between.

[solidpro]

I'm in telecoms now and copper is cheaper, easier to terminate, less problem if there is a break or fault, easier to work with, more commonly understood and meets the technical requirement of the service I'm paying for.

[ajkula66]

I'm in telecoms now and copper is cheaper,

Definitely.

easier to terminate

Debatable.

less problem if there is a break or fault

Disagree.

easier to work with,

Strongly disagree.

more commonly understood

Definitely

and meets the technical requirement of the service I'm paying for.

No dispute there.

[solidpro]

Making up a fibre optic patch cable from scratch, including terminating both ends is far simpler with copper. I woudn't even know where to start with building my own fibre optic patch cable.

[ajkula66]

Making up a fibre optic patch cable from scratch, including terminating both ends is far simpler with copper. I woudn't even know where to start with building my own fibre optic patch cable.

Well, I could teach a blind monkey to splice fiber... ...wouldn't try the same when it comes to splicing a 1200-pair copper cable...

[solidpro]

Um, what ethernet patch wire has 1200 pairs?

[ajkula66]

Um, what ethernet patch wire has 1200 pairs?

I said copper not ethernet. You won't find the word "ethernet" in any of my posts above.