Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[mywifeshusband]

[waltergmw]

Hello MWH,

Here are just a few thoughts:-

Why do the various speed tests return such different results?

All speed tests use some form, often different, of statistical sampling as well as routing to different far-end-points all subject to contentions etc. on the way.
BT's tests obviously do a great deal more including determining whether you are still on a fixed speed service, an ordinary ADSL1 service, a 21CN ADSL2+ service or a VDSL FTTC service.
BT make you agree every time that you will obey their rules but I observe no difference on a test of one line whether I do the test at the house or remotely using LogMeIn which obviously increases the traffic on the line being tested.
I've even tried launching a BT test remotely and then disconnecting LogMeIn and logging back on 5 minutes later and comparing that result an hour later by remaining logged in.
BT's results are, as usual, provided with very astute commercial wording (and different scaled upstream and downstream histogram displays) as well.

TBB speed tests seem to produce more consistent results.

Speedtest.net actually show you the sampling values they then amalgamate them to produce a final result.
The download speed graph is also quite helpful.
They are subject to the problems described with Flash recently here.

Why did I get such a vast difference between Wired and Wifi tests ... etc.

There are differences in the firmware (and their associated "features" and bugs) behind the different types as equipment you use.
The actual wireless devices will vary with different types and sizes of antennae, the positioning of the devices, other WiFI traffic on different networks etc. etc. I.e a very large bag of worms !
The most reliable type of WiFi is not to use it at all ! That said, there are many who would be dead in the water without it !

I eagerly await more erudite answers from the cognoscenti !


Kind regards,
Walter

[burakkucat]

I think the important thing is not to start Bald_Eagle1 going off on another saga about FTTC services. 

The questions asked all relate to the end-user experience with different hardware. What would have been the most interesting, to me, would be the actual sync speed of the VDSL at the same time that the throughput was measured. Obviously that can only be observed with the unlocked HG612 modem firmware in place. 

[Bald_Eagle1]

Hi mywifeshusband,

Thank you b*cat

I think the important thing is not to start Bald_Eagle1 going off on another saga about FTTC services. 

No comment 

Assuming the configured speeds you mention are the results from BT's speed/performance checker, these are the same as your IP Profile/BRAS Rate.

I suspect the increase to 37.565Mbps in December was due to the FTTC/VDSL2 profile change from 8c to 17a (higher frequency band available for BT's soon to be rolled out 80Mb/20Mb service).

As IP Profiles for FTTC usually work out at 96.79% of Sync speed, that would suggest an actual Sync speed of 38.811Mb.

Your initial reduction in speed may well have been the telephone noise issue that was "cured".
However, it appeared to have left your connection in a poorer/slower condition.
This was apparently not resolved depite assurances from India.

However, it may by coincidence have stayed lower, at that time, due to other causes such as increased crosstalk from more subscribers becoming connected, or other external sources of noise interference.

Without access to an unlocked modem, there is no way to monitor/record any changes in connection conditions such as increased attenuation, SNR, SNRM, bit loading etc.

So, it is quite feasible that if your connection had remained in the same condition as it was back in May, with the same external factors affecting it (or not), being switched to the 17a profile would have given you the full 40Mb Sync speed.
Most other users with similar IP Profiles to yours are now indeed seeing the full 40Mb following their switch, with IP Profiles of 38.717Mbps.

As you are not quite achieving the full 40Mb at the moment, your "Attainable Rate" will not be much higher than 40Mb, so it is unlikely that you would see much (if any) advantage from the 80Mb service as & when it is introduced.

A lot of this is "guesstimation" as only an unlocked HG612, or a different make "accessible" modem could confirm your connection's current condition etc. without an engineer's visit.

As far as differences in different speed testers goes, I think the jury is still out.

Some users have reported that they see large discrepancies in speed tests when using the HH3 as supplied for BT's own Infinity service, particularly when connected wirelessly. I'm not sure how much speeds are affected in normal use though.

Walter has mentioned the TBB speed test being the most reliable.
It may well be for certain routers & users' PC settings.
I use a Netgear WNR1000 v 3 router as my FTTC is supplied via Plusnet.

I use Windows 7 on my ethernet connected desktop PC & Vista on Mrs. Eagle's wireless connected laptop.
All my speed tests are conducted from the desktop PC as there is a significant speed reduction when using the laptop.

The speed reduction to the laptop is in part due to it having an old (low speed) wireless adaptor & it is located a couple of 2' thick walls away from the router.
With FTTC it is perfectly adequate for Mrs. Eagle's occasional use.

Throughput on the laptop is around 16Mb at best & can now range between 24Mb & 27.5Mb on my desktop, directly related to my Sync speeds which do vary due to frequent uninitiated "on the fly" re-syncs.

Again, without access to an unlocked modem, you would be blissfully unaware of any of these "on the fly" re-syncs as they do not always result in a dropped PPP session, your dynamically allocated IP address does not change, & most (if not all) ISPs do not detect them in their own logs.

I find the TBB speed tester & TBB's test download files give me the most unreliable results.

My own preference, for my particular setup, is speedtest.net as it seems very consistent, showing a slight slow down at peak periods (presumably contention) & at quiet times the results are usually around 97% of my IP Profile.
mybroadband.co.uk & bbmax speed testers also deliver very similar results to speedtest.net.

I suppose it is best to experiment with different speed testers & settle on one that most suits your own setup & then stick with it to ensure at least consistency in your results, even if they are ALWAYS optimistic or pessimistic.
That way you would be able to see any significant changes in your throughput.

Also, unless attempting to prove that contention is at an unacceptable level i.e. that the service is potentially oversubscribed, it would be advisable to ALWAYS conduct speed tests at quiet times e.g. between midnight & 08:00 a.m.

Different antivirus programs can also apparently affect throughput on speed tests.
Kapersky, for one, appears to do so.
I use AVG - Free edition. That appears to have no effect whatsoever on my usual speed tests.

Finally, I believe Windows 7 auto detects the router's MTU settings & adjusts the PC's sttings accordingly to ensure the highest compatibility between the different hardware in use.
XP doesn't do this & a manual adjustment of MTU settings (& possibly Rwin) may be required.


Paul.

[mywifeshusband]

Oh dear I do seem to have opened up a can of worms when I  intended my questions to be mainly rhetorical. Why though it is so hard to get a reliable FTTC speed test though is quite beyond me. I'd have thought that a file of size X downloaded in a time T should by simple arithmetic give a more or less uniform answer with only issues that qualify factor T effecting the result: and I would have thought the limiting factor would still be the FTTC download speed of a 'mere' 35Mbps rather than any home computer or local networking speed restricting factors: - unless still using a Sinclair Spectrum of course!.   (Now there a thought to ponder; FTTC Broadband on a Spectrum, such a thread could run for months. )
Be that as it may, my main point in posting was to let it be known that BT is not a complete waste of space if you keep on at them and having kept good statistical records make sure that they take note by elevating complaints up to the top of the tree. Mind you I concede it is an up hill struggle.
Thanks anyway chaps for the trouble you have taken to respond.