This is a graph of the attainable downstream speed as reported by "adsl status" on the Vigor 120v1. As you can see, it's all over the speed range and it's this characteristic that concerns me. Is this "normal"?

The sync speed is much more stable but may, I suspect, be only what was negotiated at the last connect whereas the attainable speed is what actually affects your downstream speed.



Advice anyone?

Kelvin

This strikes me as pretty flat with only a 4% variation from the mean. If the data was displayed with a true rather than false origin (ie the vertical axis starting at zero) it would look reassuringly flat.
I note that the cycles are daily; isn't this what we are all plagued with? I know that my own daily variations, both in London and on the south coast, are rather more than 4%!

Regards, LA

Just to keep the thread up to date.

I changed the ADSL firmware from the "UK - 332201" to the "STD - 321311" version. As a result, from the 28th onwards, the speeds have increased markedly. That's just above the "4" in "Week 04" on the graph.

However the cyclic variance still occurs. I'd love to know why...

"However the cyclic variance still occurs. I'd love to know why..."

Why is my connection slower in the evenings?
ADSL broadband is a shared service. Therefore the more people who are using it, the slower it will be. The amount of people using a broadband connection in an area affects the “contention ratio” which is the ratio of users to the area’s exchange and the coping capacity the exchange has for those connections. Most home users use their computers in the evening, so connections tend to be a lot slower at these peak hours. The best times to use broadband may be between 12am and 8am!
Cable Broadband speed does not change during the day, making it a more consistent alternative.

Regards, LA

I agree that there is contention on the service, however, this is not the area that I am interested in.

The cyclic quality issue is at the ADSL ATM layer and related to physical signal quality rather than the contention related throughput.

As far as I am aware, the physical link between my ADSL modem and the exchange DSLAM is a private 1:1 connection and not subject to contention.

It is the characteristics of this part of the connection that SNR and Attenuation values represent.

Kelvin wrote: Once you had changed to the more stable modem code, how did the reported downstream speed compare with what had previously been reported?

In other words, what is the penalty for stability?

Thanks.

Sorry for the late reply! I don't think I set up my account here properly since I didn't get an email saying there was a reply to my post.

There was no penalty at all for switching, as you found out. I was getting 3500-3900 down and 800-890k up... until recently that is. Something's changed on their end (everything's exactly the same on mine) and now we barely get 3000 down and 700 up, according to the modem, and the SNR margin has increased to 9db where we were previously stable at 6db. The OpenReach van has been parked in front of the phone box on the corner a fair bit of late... maybe I'll ask the BT man what the problem is.