Print

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

[Weaver]

@Chrysalis - Don't know the answer to that, but in any event it's A & A's bad bug if it does what it appears to do.

[Weaver]

Have reported this to A & A for investigation.

[Weaver]

Dopey question: what effects am I likely to see?

[Weaver]

Do I have to power-cycle the modem for this to take effect, I wonder?

[Weaver]

Have confirmed that A & A don't touch the tick box, it comes unticked by default and a non-A & A supplied owner said it's unticked on his device too. So it's nothing to do with A & A, but they are not "fixing it" either by ticking it.

This seems to be a complete red herring. Perhaps it's set to auto on anyway and the UI does nothing, a UI bug. Perhaps DLink just thought it was a useless idea but accidentally left the web UI in for it. Lots of speculation.

Notice that the pic earlier has it ticked (for wrong device model).

[Weaver]

I tried to find a way to report this issue to DLink, but I found their www UI bewildering. If anyone knows of a more civilised way to get in touch with real humans at DLink, then I would be grateful.

[Chrysalis]

A resync would be needed but a reboot of modem should not be needed.

If you able to monitor the bitloading of each tone, and the bitloading shifts about then bitswapping is working.

[Weaver]

@chrysalis - don't yet have the technology, three modems hidden behind a router. But maybe in the foreseeable future. See http://forum.kitz.co.uk/index.php/topic,16700.0.html

[Weaver]

I had better luck when I reported it straight to RevK via a tweet. Thanks Adrian. I suspect he will look into it.

[NewtronStar]

Here is my bitswap per tone graph it's only been running for 10 minutes and you will see each tone that is being effected by noise on this xDSL circuit.

[Weaver]

That's a very slick tool there.

[NewtronStar]

That's a very slick tool there.

It is slick tool by Roseway one just needs an unlocked HG612 or compatible modem/router but the graphs won't persuade your OR engineer visit to go deeper when looking for a fault unfortunately they won't understand what they are seeing in those graphs they only look at their JDSU no amount of printed graphs will help, it's only for personal use.

[Weaver]

I wrote
> This seems to be a complete red herring. Perhaps it's set to auto on anyway and the UI does nothing, a UI bug. Perhaps DLink just thought it was a useless idea but accidentally left the web UI in for it. Lots of speculation.

I can't remember what evidence I based the idea on that this was a red herring. Has DLink been sending these out with bitswap turned off by default according to the UI at least? Think so. Because we can factory reset one. And if the UI actually does anything, then AA has not been 'fixing' it, which is a real shame. I failed to persuade AA to change their config.

How would I actually test it? What kind of proof do I need to get as an outcome from ticking it,  doing the right thing, to show it actually gives an improvement of reliability or speed or both? (Given that it could be that the UI does nothing, because it has a bug in it, and it just does the right thing anyway.)

[ejs]

I think you'd need evidence that the modem has done some bitswaps, which would imply recording the bits per tone data at different times, and seeing if it has changed. Of course if no bitswaps have happened, that might be because none were needed, but over several days, there are going to be changes in conditions from day to night, so it would be very unlikely for no bitswaps to ever be needed, especially on a long line.

[Weaver]

I don't recall offhand whether user g.dmt worked out how much in the way of stats can be extracted from these modems. I would have to isolate one temporarily because my router acts as a barrier between the modems and the main Lan, and then pick its brains using telnet.