Print

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

[roseway]

I'm actually using Routerstats Lite v2.3, but I'll checkout the full product and put a howto in the Linux section of the forum shortly.

[graevine1]

Thank you Roseway, Im sure that will be appreciated by many with difficulties.
Im sure that the owner of Routerstats a fellow Radio Amateur is however extremly competent in Linux from the looking into it that Ive done, so maybe somewhere it is already a chore already done but where maybe is the question.

Have a great day, and I shall await with interest.

[roseway]

The howto is there now.

[roseway]

I'm certainly learning something about the problem by running routerstats. DLM has increased my target noise margin to 12 dB so my attempt to avoid this by getting in first with a DMT tweak didn't work. Now, most of the time my noise margin is between 10 and 12 dB, but occasionally there is a horrible pulse of interference from somewhere, as you can see below. I'm suspicious of my central heating thermostat, but haven't yet been able to nail it firmly to that source.

By the way, I think this is quite separate from the general problem which started 3 days ago - that seems to be settling down, with only a small variation in noise margin where it was hopping up and down by 6 dB every few seconds a couple of days ago.


[attachment deleted by admin]

[roseway]

If anyone is still interested in this little saga I got a mains interference suppressor and used it in the supply to the router. Since then my connection has been up for nearly 24 hours and there haven't been any of those nasty interference spikes. That may be coincidence of course, but it lends some support to the theory that mains interference was causing the spikes.

When I pick up the phone I can sometimes hear a bit of crackling, but I think it's too intermittent to be worth reporting as a fault. I'll leave things as they are for another day or two and see how it goes. My IP profile has already recovered slightly to 4000 (it went down to 3000 at the worst point).

[graevine1]

Another beauty of the Netgear DG834G is that they run of 12 volts DC so I run them on a battery 7amp sealed lead acid aka intruder alarm and my own I power from a filtered intruder alarm 12volt 8 amp battery floted supply to include hubs switches VOIP units etc etc each 12 voilt filtered and obviously get no supply glitches.

BUT

to eliminate mains bourne the battery supply not connected to mains is a great test setup.
Have a great day

[kitz]

>> Since then my connection has been up for nearly 24 hours and there haven't been any of those nasty interference spikes. That may be coincidence of course, but it lends some support to the theory that mains interference was causing the spikes.

Very possibly.  lets hope it stays that way for you eric.

[mr_chris]

to eliminate mains bourne the battery supply not connected to mains is a great test setup.

Indeed - I never even thought of trying that

Eric, any recommendations for mains suppressors - what kind did you get?

[roseway]

I can't say it's a recommendation - I just went to PC World and bought the only one they had, which was a Belkin SurgeCube. It claims 43 dB reduction in the range 150 kHz - 100 MHz.

[graevine1]

Some years ago I built all my Hi Fi systems if I remember correctly they were from designes published in "Wireless World" Then came the odd unwanted clicks, I got on well at that stage with Hewlett Packard and borrowed their latest fast Scops. Wow got it in one gee those (mains bourn) transients are fast and sharp, so I fitted some three wire mains filters inside the Ali HiFi boxes. ----Job done ____ but gee were they expensive

So I really like Netgear because they are 12Volt DC input and you can prove a case and drink a nice bottle of Red whilst you watch your neighbour smile (and of course pay for a nights great wine, my choice obviosly) whilst taking many hours thinking about watching routerstats.

[roseway]

After several days of drying out (or whatever it's being doing) my line is now clear again, and Routerstats shows the noise margin as virtually constant. The unfortunate legacy of the episode is that my target noise margin is now 15 dB. The final 3 dB rise was triggered by a single disconnection after 40 hours uptime, when I perhaps foolishly swapped to a different router.

The reason I wanted a different router is that I wanted one which is both tweakable and supported by Routerstats, and from my little collection the DG834GT fitted the bill. I tweaked the target noise margin as low as it would go (about 10 dB) and it connected at a speed which will get my IP profile back to what it was before the episode. It's been running for 20-odd hours with no more than 1 dB change in noise margin.

I'm going to leave it as it is now and see how it goes. A bit later I might ask for my connection to be retrained if it remains as stable as it seems now.

[kitz]

>> my target noise margin is now 15 dB. The final 3 dB rise was triggered by a single disconnection after 40 hours uptime, when I perhaps foolishly swapped to a different router.




>> It's been running for 20-odd hours with no more than 1 dB change in noise margin.

In view of the fact that a couple of the disconnections over the past few days were swapping to the Netgear and installing the new firmware... and the fact the higher Target SNR was caused by you changing the router and things are stable again..
 
Perhaps one of the nice guys from Zen could see if they can do anything about that

[PhilT]

hmm trading off the upstream -v- downstream?

This isn't a feature of G.dmt ADSL, is it ? the up and down frequency bands are fixed.

Annex M ADSL2+ can optionally use some higher frequencies for upstream rather than downstream, but vanilla ADSL doesn't do that.

Imagine the crosstalk if you started using the lower downstream tones for your upstream and next door was using them for downstream - you would drown out the exchange.

Phil

[mr_chris]

Phil, don't you remember 512k ADSL could drop the upstream down to 64k in order to get 512k downstream - from memory, it was even called RADSL in the BT docs?

So it *is* a feature of ADSL1 and has been used by BT for years. Echo cancellation techniques are used in order to allow upstream bins to be used for downstream to minimise problems caused by doing this.

Whether it's active on Max lines is another question and is certainly a theory to explain some of the oddness Eric's been seeing. Another possibility is that the interference / water / whatever was causing the issue was causing problems around the lower frequency bins which was causing upstream issues?

[PhilT]

I remember very well that 512k was and is rate adaptive on the upstream.

What it did not do was use the upstream frequencies to allow the downstream to sync, it simply adapted the upstream to achieve what it could - in the same way that Max adapts the downstream.

It was an Urban Myth that one was exchanged for the other but it didn't happen - you aren't going to hear the exchange downstream if I'm on the same channel upstream in an adjacent pair of wires at the DP.

A check of the modem settings or bin plots will show a fixed regime around the allocation upstream / downstream.