Different Line Speeds on Different Devices

I recently changed from BT HH6 to BT Smart Hub 2, both with BT’s SuperFast & an estimated download speed between 59 & 73mb/s.
The HH6 always gave a download speed around 68Mb on both 5GHz wifi & ethernet with both Windows-10 laptop & iPad on wi-fi.

The Smart Hub 2 however does not match this; using Fast.com www speed tester (& also a number of others) on both 5GHz & on ethernet with my Dell / Windows-10 / Chrome it will only have 50Mb (less on 2.4GHz)
Whereas using Fast.com on my iPad & 5GHz I get 70Mb

The question is why do I have such dramatically different results using the same speed test software & the same 5GHz band on a Windows-10 laptop compared to an iPad.
It’s all working OK with no obvious line speed problems, it’s just a question with an itch that needs scratching.

I’ve asked the same question on the BT forum, but maybe a Naimite has an answer.

Different WiFi chips in each device. Different drivers, operating system overheads. Bad aerial placement. I recently bought a DAP which had really poor WiFi performance. Had to stand directly under my ap to get any decent bandwidth. Both my phone and pc where a room away and tranced it. Getting two wifi devices to perform the same is likely to not happen, also environmental differences each time you try to compare them will also effect it and your internet speed is never fixed and will also effect it.

There are many variables with Wi-fi… not least as some of said as chipsets and interoperability… one can’t reliably use Wi-fi for speed tests… in fact many speed test sites will state this.
So it would be of more note if there was a marked difference via Ethernet.

It’s worth noticing your Actual throughput will be shaped by a BRAS in the broadband aggregation layer in BT’s access network, this typically is about 10% to 15% less that your VDSL line speed… and shapes the traffic for optimum flow considering overheads and error correction in your access line. Without this your overall throughput would be less with real data flows, and cause unnecessary traffic in backhaul network… think of it like reduced speed limits on very busy motorways.

For example my line speed on my Smart Hub 2 is approx 48 Mbps, my BRAS setting is approx 42 Mbps, and speed tests with my Smart Hub2 typically turn in around 41 to 42 Mbps … but I find no difference between Ethernet OSX machines and my iPad …
I use Ubiquiti wireless access points however.
To identify my BRAS setting in the core access I use (this should be your true internet downlink access speed, so you don’t need to use ‘speed checkers’ which can be notoriously in accurate ) https://windows.mouselike.org/be/index.asp?DoAction=BrasChecker

Hi Simon, I’m very aware of this; this is not my ‘question’
My new Super Hub 2, measured on Windows/Chrome and locked on wifi band 5GHz, and also when connected with ethernet, gives aprx 50mb/s.
Whereas on iPad 13.3.1, also locked on 5GHz, shows 70mb/s.
Previous BT HH6 with the same as above including iPad all gave the same 68 to 70mb/s results.
All that’s changed is the new BT SH2 … ???

Ok so what about your BRAS setting… that will be the actual true speed… not the speed ‘guessed’ by your speed test software/web site.
If you definitely want to run a speed estimation package, run the BT wholesale one… does that still give a variance? I suspect it might…
If it does without looking at the algorithm used to estimate the speed it’s hard to tell…
It possibly even could be different max MTU sizes with your new SH2 etc… but without looking at WireShark that would be a guess … also explore possible areas, like duplex settings on Ethernet… any client /router firewall settings etc.

So for speed I suggest use the BRAS method… for speed estimation software use it as an approximation and relative guide when used on the same platform.

I saw no speed estimation difference when I went from HH6 to SH2, but I only used OSX and iOS and did run such software on Windows…so never investigated any observed differences that may have existed.

Sorry I can’t think of an obvious reason why your speed estimation software is throwing up differences… so can only suggest investigation if it bothers you.

Hi Simon, what is BRAS setting, www look up is all about lady stuff, so not helpful.

BT Wholesale on both iPad & Windows is variable, it shows 50mb/s ar some times, other times it’s the same as all other www testers and has the 50 / 70mb/s variance with iPad vs Windows

That made me laugh…
BRAS: Broadband Remote Access Server [pronounced Bee-Raz]

It shapes your network traffic based on the established current synchronisation parameters of your broadband router modem to the DSLAM/MSAN (which is in the local cabinet for VDSL) . Your access bandwidth becomes optimised for throughput (which can be 10 to 15% less than your indicated line speed because of error correction/vectoring and other technologies used) … without it if you were to connect to higher speed servers than your internet connection (most likely) your bandwidth performance would be poor and there would be much discarded traffic/data in the route to the remote server… Shaping is undertaken in most WANs at the access boundaries.

This web service can interrogate your BRAS setting if you are on a BT Wholesale based service… though sometimes a result can’t be returned for you

This won’t help either, but am I the only one who visits the Padded cell and occasionally types ‘pad’ into the browser search window expecting it to auto-complete, but when I hit retrun gets search results for ‘Padded Bras’ - with all the targetted ads/tracking cookies I often wonder what Mrs AC will think I’ve been looking at! :smiling_imp: :smiling_imp:

Might help:


1 Like

I don’t think it’ll explain your wi-fi variances, but if you find your router connection speeds get progressively worse at some stage and worry about line issues bear in mind that the ADSL filters at the sockets can go bad.

I had this issue a year or so ago and was running diagnostics with my ISP - I then found a spare ADSL filter, swapped it and problem solved.

I understand about the ADSL filter AC & thanks for the post.
Problem as I see it is the BT Smart Hub 1 on 5GHz band tested on Windows/Chrome gave 70’ish download speeds as expected, swap to Smart Hub 2 with 5GHz band locked in & its 50mb/s
iPad still shows 70mb/s from the same wifi hub on the same route & same ADSL filter

Hi again @Simon-in-Suffolk
BRAS is something new to me
Problem is it does not allow consecutive tests, 1 had to wait >1 hours between, hardly user friendly.

With Smart Hub 2 on 5GHz band (2.4GHz switched off)

Windows-10/Chrome (copy paste)
The current Downstream BRAS rate is: 73.7 Mbps
The current Upstream BRAS rate is: 20 Mbps

The current Downstream BRAS rate is: 73.7 Mbps
The current Upstream BRAS rate is: 20 Mbps

Fast.com test shows

49 Mbps (download)
Unloaded 7 ms
Loaded 164 ms
Upload 17 Mbps

69 Mbps (download)
Unloaded 6 ms
Loaded 35 ms
Upload 18 Mbps

Its very strange that the BRAS test show identical results, very strange, I’m wondering if this is an actual test or just the estimated speed for that phone line (number)
PS: 2 hours later I ran the same tests, exactly the same numbers.
Conclusion is BRAS is not an actual line test

The Fast.com (and some others from www) on Windows all show a low 50mb/s speed whereas iPad using the same selection of www test programs is at the expected BT advertised rate.

All very strange & despite being a little more IT techie than most I find this most unsatisfactory.
Still waiting a response from BT

…. YF …. confuzed .com

Hi Mike the BRAS setting is the actual setting by the shaper … everything else is an estimate such as from ‘speed test’ software or sites… and dependent on interoperability matters. and other matters not directly connected to your internet connection.
BRAS setting are dependent on the values recovered from the DSLAM / MSAN sync settings, so will be the setting for your connection irrespective of which client.
So your can be confident your actual downlink speed is around 73 to 74 Mbps… as effectively that is your shaped throughput based on the quality of your physical connection.
It would be very strange and in fact not be possible to have different values for different clients on your connection.

Hopefully it make senses now with regard to your downlink speed. BTW for uplink speed it works a slightly different way and the BRAS value is not the speed.

As far as interoperability with your windows machine and your speed test software/web sites and your home network/router… you might need to ask your speed test software provider or Microsoft support… it doesn’t look like it is a particularly access network related matter… unlikely BT or any ISP would have a view.

I came to the same conclusions Simon, but thanks anyway.
I’m satisfied my down-link speed is as BT advertised & aprx 70mb/s. However what I have trouble with is that the BT HH6 showed 70mb/s no matter what its tested with, whereas the BT SH2 shows something so completely different depending on the device thats used to run the test program.
I might understand some of this stuff, but what about Joe Public. As for the lack of response from BT … nuff said

Indeed which is why my suspicions would be at client hardware/client operating system level… I would need to look at WireShark traces between Windows client and router switchport of before and after to see what is different… in fact I am quite curious.
The other question would be is it the same for all Windows (I assume 10) clients or just certain devices… assuming you have more than one.

I checked on Lyn’s laptop, its the same as mine Windows-10 & gives the same low 50mb result.

I’ve e-mailed my son who has BT SH2 & has a selection of Windows & Apple devices, awaiting reply.

My son has a Windows10 machine… I’ll see later if that differs from the OS X machines … however it’s wifi connected. (Via my switch connected Ubiquiti wifi access points).
However the OS X and iOS machines all appear consistent.

1 Like

Just to avoid cross wires
All my testing has been over wifi with the 2.4 band switched off (locked into 5GHz)
However I get the same low 50mb result with ethernet to laptop.

It looks like the ‘problem’ is in my Dell Win-10 laptop wifi.
I’ve run a reset of everything, BT SH2, laptop, reset network adaptors, checked & then manually searched for network driver updates. Although the Intel wifi adaptor is good, its out of date & no longer supported & it’s 433Mbps so I suspect that is part of the problem.

With 2.4 band off (locked to 5GHz)
Dell Win-10 laptop still shows (aprx) 50mb on wifi, but it now shows 70mb on ethernet
HP Win-10 laptop shows 70mb on wifi
iPad shows 70mb

Why it was consistent across the board with the old BT HH6 but not with the new SmartHub2 remains to be answered.

Finally, after a lot of fiddling around, & being far to busy with this lock-down thing, I found & fixed it.

Dell had installed a program called SmartByte . The mystery was the application was installed sometime ago in 2018, but for whatever reason it was not a problem until the BT SmartHub-6 was install last month - hence me pointing the finger at BT.

SmartByte is an application that detects streaming video & gives that feed most of the available Internet connection. It works with both Intel and Realtek wireless cards, but it seems that its not working correctly as it does not release the video priority when video is not in use.

It was OK for www browsing & you would hardly notice a difference, the real problem was downloading something, e.g. an audio album & also when uploading from laptop to NAS over wireless.

So my solution is to uninstall SmartByte, I now get the expected ~70mb/s wifi downlink speed test & a wireless load speed to/from the laptop is normal.

1 Like

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.