Further test carried out, shows noteable difference, normal performance acceptable but once the 81 test is carried out it's all downhill as the graph shows.

However even look at the consistency of the normal test, far better than previous, should this not be the way a test should look always, or is it normal to have drop downs to zero speed and so much fluctuations.

http://img75.imageshack.us/my.php?image=plusnetport81exampleoj7.jpg

I think it's fairly safe to ignore the results for port 81 and 8080 as it doesn't seem to be indicative of the experience that you're actually getting.

Just rely on port 80 like normal speed tests will give you. The tests you're running seem to be making you think you're getting a poorer service than you actually are.

No the vat majority of my tests are not on other ports, I merely done these to show there is a difference and as mentioned traffic management can go wrong?

However that one stable test I showed you is one of very very few, I could have hosted a stack that have up and down / dropout.

An expanation as to whether such behaviour is normal or not would be good.

I would test it elsewhere to compare but will not be able to for around 4-5 weeks perhaps.

So guess this means TTB meter is a waste of space and not accurate?

We'll take a bit more of a look at what's happening with those test results from our side and come back to you when we've got an answer.

Regards,

Ian

Are you using a Speedtouch 585v7 router by any chance?
I ask because I know Plusnet now supply those and I've recently had poor throughput result from a newly supplied one although on a different ISP.
The solution was simply to reset the router to factory default.
Note: One symptom of the problem with the router was a incorrect and very low report of the Rwin setting from the following site.
http://www.speedguide.net/sg_tools.php > tcp/ip analyzer.

Nigel

http://www.farina1.com/compare.jpg

First block 20MB port 80 file, second port 81 and third port 8080 done at 19:44 today. Machine itself was busy running other stuff but mainly CPU bound rather than internet connection.

Links to the files for download without using tbbmeter
http://download/thinkbroadband.com/20MB.zip
http://download/thinkbroadband.com:81/20MB.zip
http://download/thinkbroadband.com:8080/20MB.zip

I was using a Netgear DG834Gv2, then got a v4, Plusnet supplied a BT Voyager 2110 to eliminate any problems with my own equipment

Hi there,

You could always wireshark capture the speedtest traffic to see if it's getting prioritised correctly by the system?

Interesting - looking at packet headers to count TCP/UDP etc is on the list and should be possible to build something like this into the app