TBB Speed Test.

I like the TBB speed test as it tells you what you are getting not just your maximum speed as others do, for quite a while i have been getting 25 - 28 Mbps but for the last few days i have been getting 35+ Mbps (see below) as i cannot see TT suddenly giving me a speed boost for nothing i am wondering if the way the speed test runs has changed.

This is the TT 38Mbps max package.

https://www.thinkbroadband.com/_assets/speedtest/but...

Edited by deleted (Thu 21-Jun-18 15:57:27)

And the connection speeds of your router?

Nothing changed in terms of how the test does its testing

There's 101 reasons that could make your speed tests jump up.
DLM changes, vectoring on the cabinet, crosstalker being on holiday or ceasing their line.

We really need to see xDSL statistics from the modem.

There's nothing Talktalk could do to incase speeds like that even if they wanted to. They have no control of the DSLAM or DLM, and they do nothing that could have limited the speeds in any way.

Edited by j0hn83 (Thu 21-Jun-18 17:06:21)

In reply to a post by MrSaffron:

And the connection speeds of your router?

Nothing changed in terms of how the test does its testing

This is my speed its been the same for a few months -

Link Rate: 9994/39999 Kbps

So entirely possible to get those sort of figures from a speed test

In reply to a post by MrSaffron:

So entirely possible to get those sort of figures from a speed test

Thanks, even better today, a post about talk talk speeds that isn't negative has got to be a rare sight.

http://www.thinkbroadband.com/_assets/speedtest/butt...

I have just upgraded to max speed BB (80Mb) and some speed tests show 77/15 but the TBB test are much lower at around 38/12.

When I was on a 40/10 package the TBB test always showed a slower speed than other test sites.

Any thoughts ?

And your modem is synced at what speed?

http://tbb.st/1533744670804811855 is your test and it looks far from healthy, which explains the slower speeds

A good test would look like http://tbb.st/1533657179522923755

My reasonably new Netgear modem/router connected directly to my NET5 shows a connection speed of 75.89/20.00

Speed test from the speedtest.btwholesale web site and from the further diagnostics button shows

Ipprofile of 73.46/20
Download speed of 106.75, up 8.08

Max achievable down 73.46, up 20.

Interesting how BT show a download speed of 106, much faster than the modem sync !

In reply to a post by lmschuffer:

My reasonably new Netgear modem/router connected directly to my NET5 shows a connection speed of 75.89/20.00

Speed test from the speedtest.btwholesale web site and from the further diagnostics button shows

Ipprofile of 73.46/20
Download speed of 106.75, up 8.08

Max achievable down 73.46, up 20.

Interesting how BT show a download speed of 106, much faster than the modem sync !

Looks to me like you're using Chrome as your browser as this is a known bug for Chrome to give much higher results from the BT wholesale website.

The speed test result is obviously wrong, due to using a browser that does not work well with the test. It would appear that IE11 and Edge do work.

Tested using both IE and Chrome, opened one at a time, the outcome here is that chrome is faster on all sites and the same at speedtest.net
-------------------------
BTwholesale speed test

IE 72.08/14.54
Ipprofile 73.46/20.

Chrome 115.79/12.22 ping 37.50

Ipprofile 76.98/20.
--------------------
Broadbandspeedcchecker

IE 33.09/14.19 ( painfully slow to complete !)

Chrome 72.86/19.26
-----------------------
TBB

IE 19.01/10.8

Chrome 35.0/11.1
-------------------------
Speedtest.net

IE 73.94/7.99

Chrome 73.28/17.99
------------------------------------
You can see why I use Chrome, and it's noticeably quicker on normal browsing.

More confusing than ever.

Following a cabinet move the BT BB availability checker displays the speeds available from the old cabinet 1200 meters away rather than the new 170 meters away.

Even though the modem sync is up my BB does not seem as quick at 80 than it did at 38, the problem now is which test to belive before contacting the ISP. (again)

I had not heard the BT and chrome problem before so I googled "does the chrome browser report incorrect broadband speed tests"

Found a few replys from 6/8 years back and then.

https://community.plus.net/t5/Fibre-Broadband/speedt...

Its a Known problem especially with Firefox

Mozilla support says its because BT test still using Flash not HTML5 , Plusnet have been in contact with BT several times but still await a fix from BT

AND
https://community.bt.com/t5/ADSL-Copper-broadband/Is...

Re: Is there a better, more accurate, speed test than the BT Wholesale code?
the btspeedtester has been giving incorrect results as others have posted but it appears to happen when firefox is used so can you try chrome or edge/IE and see if that helps get more accurate results, it is only test accepted by wholesale

HELP !.

So there is something wrong with the connection "my BB does not seem as quick at 80 than it did at 38" our test is designed to not mask this, whereas speedtest.net will often mask a slightly poor connection, i.e. it has better bragging rights but less diagnostic use.

I did say our test was showing things were not right with the connection, only thing now is to figure out where.

Using Chrome
Your standard BB test today gave 27.4/11.6 and burst 44.8/12.4
The special test version gave 41.28/12.9 and burst 45.6/14.
Using the download file test on port 80 I achieved 512Mb 69 sec, 200Mb 27 sec indicating a 60 Mb speed.

Using Edge, which the test takes much longer to load and complete
Your standard BB test gave 26.0/4.7 and burst 27.4/5
I could not complete a special version test with a can't reach page error, tried a few times.
Using the download file test I achieved 512Mb 96 sec, 200Mb 39 sec

Modem sync still 74.98/20.

Yesterday I made several changes to my setups, swopping out both modem and router.
Ran the tests on my PC and an old laptop (XP) and saw no noticeable speed difference with either setup.
I cannot see the problem being PC or laptop related, wiring, version of Windows, browser choice or connection method.
But what ?

http://tbb.st/1533984438772496555

Latency while idle
[41,46,39,43,39,46,49,70,49,40]
Latency while downloading
[4242,6455,5815,4719,3193,2781,7699,1768]
Latency when uploading
[42,42,53,45,50,46,49,42]

While we expect the latency to increase when the line is busy, the more normal would be a rise from 40ms to around 100 to 150ms.

With the latency so high what will be happening is that the packets saying we got that bit of data send the next will be seriously delayed and thus reducing your actual speeds.

This test http://tbb.st/1533982594604626055 gave a better set of speeds and the latency when downloading was [250,202,172,258,192,178,174,419] so more normal.

How sure are you that nothing is downloading other than the speed test?

Best test seems to have been http://tbb.st/1533897201501377455

The connection can deliver it seems, but results seem variable, cause unknown

Thanks for the insight into my issues.

I have today, starting around 11am, run 9 speed tests around 10 minutes apart and all using a 100Mb wired connection direct to the modem/router ( with one port so no AP etc ) guarantying nothing else is connected to my network.

First three from usual win 10 PC

Second three from laptop ( XP )

Third another Laptop ( XP )

The modem sync speed is as before (74.98/20) and my expectation at 170 meters from the cabinet was a throughput around 70Mb.( 96%)

Even your file download test never got anywhere near it.

I am not comfortable switching off anti virus software, the laptops are running AVG the PC Avast.
Using Chrome in all cases best I can tell there were no upgrades running on any of them during the tests.

Re packets being delayed, where does that message come from, is that the receiving computer task or some other network device ?.

At times all three computers have been connected to the network so could all be infected by the same virus ?.

A couple of weeks back I ran a speed test using a wifi Android device and the speeds were a little slower but much the same.

The DLM seems happy and the connections have been up for weeks, no reports of disconnection in router logs, sort of rules out bad wiring.

Do you think that the incorrect BT record re available speed at https://www.btwholesale.com/includes/adsl/main.html might be influencing a network device somewhere ???.

I'm told an official engineers login contains details for two cabinets.

Thanks

Re packets being delayed, where does that message come from, is that the receiving computer task or some other network device ?.

Without testing your computers on a known working well connection hard to say i.e. so can rule them out.

The question is this, when someone else is web browsing is a second person web browsing seeing pages sometimes load slowly?

If this was a virus effect, it would be showing as constant traffic and easy to spit by using Windows Task Manager and looking at the traffic over the network interface.

DLM is only worrying about the DSL elements, i.e. that 100 to 200m of copper, what happens on your PC and from cabinet back to the Internet is where your problem is likely to be.

Do you think that the incorrect BT record re available speed at https://www.btwholesale.com/includes/adsl/main.html might be influencing a network device somewhere ???.
NO

http://tbb.st/1534073014850086155
http://tbb.st/1534072458325314555
http://tbb.st/1534071879148691055
http://tbb.st/1534071261563727355
http://tbb.st/1534070602316870455
http://tbb.st/1534070008672272355
http://tbb.st/1534069347409989255
http://tbb.st/1534068508656304455
http://tbb.st/1534068007750044755

When I said posting the speed test links, this is what I meant, as makes it easy for people to look at the data, rather than rely on longer verbal descriptions.

I'm told an official engineers login contains details for two cabinets.

Can you expand on what this means?

I apologise for the long verbal descriptions, I was just trying to answer questions I thought people would eventually ask.

I don't quite follow your comment ...........

"Without testing your computers on a known working well connection hard to say i.e. so can rule them out."

Are you saying you are ruling out the PC's ?

"The question is this, when someone else is web browsing is a second person web browsing seeing pages sometimes load slowly?"

I thought perhaps my descriptions as to my method would have ruled out the possibility that slowness was due to two people browsing is it happens with any one of computers connected to the network connected one at a time, and is very consistent which is why I asked which device was responsible for the delay message.

My thought also was that as the fault occurs on all three computers this would rule out it being a particular pc issue, clearly the hardware is very different so I thought unlikely to be down to a network card driver..

Unless of course it's due to some rogue software that got installed on all three or it's outside of my connection to the cabinet.

Sorry to repeat this again but my browsing is OK but it seems no faster at a sync of 75Mb than it was at 38Mb which is why I started running the speed tests and discovered the major differences between the BT and your tests.

During the move between cabinets I was connected, in error, at 50Mb, it was a noticeable difference, when it got put back to 38Mb is was a noticeable difference, when it went to 75Mb it was not.

Connecting an Android pad to my 2.4 Mhz wifi router capable of up to 145Mb and standing next to it, using Chrome has just given a TBB speed test of 24.0/2.1. perhaps not the best test but the numbers are so similar.

Thanks

"The question is this, when someone else is web browsing is a second person web browsing seeing pages sometimes load slowly?"

The idea is to show that the PC is good, and you know then that the AV/CPU or other elements are not affected.

Your description of web browsing feeling no better than the older slower connection is bourne out by the speed test results

IGNORE the BT test results as others have said it is a bad speed test to use

JOhn83

I should have added that https://www.btwholesale.com/includes/adsl/main.html also give the date it was last tested, in my case 30/7/18, so it must bear some influence somewhere or why would they keep testing it ?.

If that test was from 8 it is clearly very wrong, but if they think it's conneted to 7 it would be correct.

My ISP could not process an order to upgrade to Max because of this info and it took weeks to find some somebody there who could override the system.

It's a long shot guess but, what if the engineers when moving my line from 7 to 8 did not take out the jumper at 7 ??????.

The BT portal does seem to think I'm connected to both !.

Just saying is all..............

The only problem in ignoring the BT tests are that when reporting faults to my ISP they tell me they are the only test results BT will accept and will not investigate or send an engineer to site if the test speeds are above what the https://www.btwholesale.com/includes/adsl/main.html site says are your acceptable speeds for your line.

So 15 to 24 does it for me !.


.

If the line was physically connected to two cabinets you would not have the high sync speed that you have

Is there a way that I can see the latency figures from new tests as those of mine that you posted earlier please ?.

thanks

The idle latency for small TCP packets is shown in the image. The full detail that you can see when viewing the live result is not available.

I would be grateful to know the general condition of the latency of some recent tests

https://www.thinkbroadband.com/speedtest/15360691926...

https://www.thinkbroadband.com/speedtest/15360693722...

An OR engineer carried out a couple of tests earler today using his laptop which "eventually" showed a throughput in the 80's Mb.

Another OR engineer, also using his laptop on the 28th Aug, who like me only saw 38/40 Mb, as I still do !.

Thanks

You can see what I am about to post by clicking the analysis button at the end of the test...

The difference is it does not take 10 to 15 minutes of digging that I have just done.

Small TCP package Latency while idle 46,49,42,47,45,44,49,40,138,49
Latency while downloading 3125,7406,6148,5141,3539,2142,4485,1234
Latency while uploading 44,54,40,54,54,42,47,43

What exactly showed a throughput in the 80's on a laptop? Since this is VDSL2 throughput above 76 Mbps at the TCP level should not be possible.

Onto your latency, the latency while downloading should increase usually but to nothing like what you are seeing.

That the latency was stable during idle and upload suggests you were not downloading or uploading interfering with the results.

The problem may be something at the ISP end that is buffering your traffic.

Tips: Try a different modem/router or switch ISP

Same conclusion on the other test

Thank you Andrew for your time again with this.

I know you see a large number of post everyday so I don�t expect you would have remembered after you had provided latency data earlier in the post that I had asked you a little while ago if I had access to see those figures

"Is there a way that I can see the latency figures from new tests as those of mine that you posted earlier please ?."

and you replied...........

"The idle latency for small TCP packets is shown in the image. The full detail that you can see when viewing the live result is not available. "

So I was under the impression I could not see any of the detailed data I wanted to see.

I have now looked as you suggest and it does give some insight but nothing like the figures you have just provided for which I sincerely thank you.

Without any scale there is no real way to know the difference in values between the small red on upload and large green on download.

Still the odd thing here is, now and only now, the OR engineer tests today showed using TBB test a smidgen over 80, say 80.2 ( last week 38 ) and I have run a couple of other sites tests and also see 80.22 and 74.39 a very long way from my last TBB test for which even the burst speed is only 43.2.....

.https://www.thinkbroadband.com/speedtest/1536076663444122555

I understand you have what is very likely a truer speed test method but sitting here it's not easy to understand who's over fast (BT wholesale - have seen 112 and others) or over slow and some in the middle.

"The problem may be something at the ISP end that is buffering your traffic.", it now looks as though they are buffering your tests and as I can't see other testers latency data I can't know for sure.

Looking back over this post it can be seen a whole range of computers, OS's, browsers modems and routers have been tried.

Still a work in progress

Regards

The analysis bar chart will show the value for each bar if you hover the mouse over the bar, the download test latency should only be a couple of times higher than the idle latency.

On the other testers, without access to your network card monitoring (Windows 10 does this in Task Manager) impossible to say if they are reflecting reality or not, or the way they handle traffic is getting around this problem.

BT Wholesale test - please just ignore it.

The key thing is how is the connection behaving in day to day use?

Also how do the devices behave when using other connections that are otherwise known to behave well.