Print

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

[jaydub]

Just noticed that my bit swap graph on MDSW (un=jaydub) has just got a shed load noisier (see attached graph).

Any ideas what might have made this happen, as I can't relate this to any changes I have made?

Thanks.

[Weaver]

New neighbour or a neighbour getting a generational protocol upgrade perhaps?

[kitz]

There's a few lines which run quite normally with anything up to 50 bitswap p/min.   It has slightly increased the amount of CRC's but your Err/Secs seem ok and still well within limits.      Whatever it is, it is so slight that your chances of finding what's causing it is slim.

You could try rebooting your modem to get a new bit allocation table to see if that helps.

[jaydub]

Thanks, Kitz.

Having had a bit more time to diagnose it, the increase in bit swapping corresponded with me changing a crappy modem cable to a Tandy high performance cable.

I don't think the problem is the cable itself rather than the fact it is being asked to bend at 90 degrees almost as soon as it comes out of the master socket.

The cheaper cables are actually much more flexible and therefore more tolerant of being bent like this.

Having changed the cable back (and rebooted the router at the same time), the bit swapping has reduced significantly.  Cable or Bit Allocation Table still remains to be proven, I guess.

The error count is pretty low, so I recognised it wasn't anything to particularly worry about.

[Weaver]

The Tandy ADSL Nation cable, which I use, might be a bad fit given that it has an RJ11 plug rather than R45 at the wall socket end and the latter might be what you need. I have RJ11 to RJ45 cables too. If my understanding is correct, then the gold-plated contacts will be at a slightly different angle with an RJ11 plus RJ45 socket.

Apologies: Is it possible to put a big 270 degree loop into the cable say, or something like that? -  ugly and in the way probably.

[jaydub]

I've reverted my ZyXel VMG8924-B10A to modem/router mode rather than using a separate HG612, which has reduced the bit swapping to negligible levels.

The issue with where the master socket is that there is a drawer unit that sits within 30mm of the front face of my Mk3 master socket.

I might try again tomorrow and see if the noise increases if I put the Tandy cable back in place.  This might just be a scenario where a cheap nasty modem cable works better than a better quality one, as the the terminator on the cable is that much smaller and the cable itself much more flexible.

[tiffy]

My DS bit swapping appears to follow a very repeatable trend, tapers up to around 50 BS/min as daylight fades and then tapers back down to a much lower value again as daylight returns, certainly appears to be reflecting a night/day pattern.

For the record, I'am also using the shortest possible Tandy modem cable, didn't record any difference in any modem parameter when I changed this out from the original modem (flat) DSL cable.
My line is on a Huawei cab with DS G.Inp enabled, same ID on MDWS.

[kitz]

I've reverted my ZyXel VMG8924-B10A to modem/router mode rather than using a separate HG612, which has reduced the bit swapping to negligible levels.

That looks quite a bit better. 

My VMG8324 has a much lower error rate on my line than the HG612.   Possible benefit of the internal filter in the -B10A models.

[Weaver]

The 30 mm thing is a bit of a nuisance, I understand the problem. I imagine shifting the NTE5 might be a total nightmare too. But if you wanted to BT would do it in no time flat. Alternative cables, as you said, might be an easier solution.