cancel
Showing results for 
Search instead for 
Did you mean: 

ADSL2+ Tones & Bitloading

GrahamC
Grafter
Posts: 257
Registered: 19-07-2009

ADSL2+ Tones & Bitloading

Hi all
I'm hoping that someone can help me understand the status of my connection. I don't have a problem as such, but I would like to understand what is going on.
There are several gaps in my tone and bitloading map for ADSL2 which have been there since day one. I don't beleive this is AM radio interference as the noise is too focused on one particular tone and seems to be very stable.
I've scoured the internet trying to find info that may explain the configuration that my particular connection has (See pics below). There are 7 gaps in total as follows:
Gap1 474kHz-479kHz (Tone110)
Gap2 548kHz-552kHz (Tone127)
Gap3 811kHz-815kHz (Tone18Cool
Gap4 824kHz-828kHz (Tone191)
Gap5 1048kHz-1052kHz (Tone243)
Gap6 1255kHz-1259kHz (Tone291)
Gap7 1501kHz-1505kHz (Tone34Cool
these tones are noted in nearly all the topics I have found on this subject, so I'm now wondering if this is some sort of MSAN/Chipset issue.
I'd appreciate any feedback especially concerning the same tones and the Vendor ID of the MSAN you are connected to.
cheers
Graham
57 REPLIES
Community Veteran
Posts: 2,321
Thanks: 230
Fixes: 2
Registered: 04-08-2009

Re: ADSL2+ Tones & Bitloading

I reckon it's almost certainly radio pickup from somewhere, probably broadcast services. There are lots of broadcasters in Europe that use these frequencies. What happens is that the phone line acts as an aerial and feeds interference into the chipset, and that does the trick. Get hold of a ferrite toroid from eBay and wind as many turns of the phone line round it as you can. Do the same for the mains cable and ethernet cable. Type 43 ferrite is what you are looking for, such as...
http://www.ebay.co.uk/itm/6x-FT114-43-Fair-rite-Ferrite-Ring-1-14-Diam-Ham-Radio-EFI-Balun-PAs-Ham-R...
x47c
Grafter
Posts: 878
Registered: 14-08-2009

Re: ADSL2+ Tones & Bitloading


Personally I'd agree with the OP in it's not AM radio interference - for the same reasons  - too focused on one frequency.
I'd be expecting to see a curve in and curve out centred around the relevant channel for the interference.  I get this from Radio 5 live, Talk Sport and Virgin AM channels though I am on the top of a hill in central UK.
As well as interference from the above I also have the same gaps
at: 110, 127, 188,191,243, 291 and 348
Like you I've no idea why
If you don't get any more suggestions try asking on the Kitz Forum - there are some there who know a lot about a lot.
(sorry my router will not supply vendor ID info - the applicable command is removed from the cut down CLI version installed by Dlink)
Community Veteran
Posts: 2,321
Thanks: 230
Fixes: 2
Registered: 04-08-2009

Re: ADSL2+ Tones & Bitloading

It's focussed on specific frequencies precisely because broadcasters use exact frequencies to broadcast on, not a band. Seeing several gaps missing together indicates broad frequency interference which is not characteristic of a broadcaster. He has specific single gaps and that's typical of radio broadcast pickup.
maximod
Grafter
Posts: 139
Registered: 15-08-2010

Re: ADSL2+ Tones & Bitloading

If the radio frequency interference is strong enough, it can cause several tones to be reduced or lost due to the bandwidth of the am modulation.
Uk uses 6khz audio modulation so will cause rf energy +/- 6khz of carrier frequency.
I suffer from reduced bitloading of tone 46 caused by radio 4 lw, and also tone 58 caused by Rte ireland.
Rte ireland seems to have wider bandwidth and so causes reduced bitloading over several adjacent tones.
Also note that if radio transmissions are between tone frequencies, it will affect several tones as well.
Also if the lost or reduced bitloading is harmonically related, this is probably due to switch mode power supply, for instance if the supply switches at 100khz, you may see harmonics of this going up the bitloading graph.
Community Veteran
Posts: 19,100
Thanks: 437
Fixes: 21
Registered: 31-08-2007

Re: ADSL2+ Tones & Bitloading

No point in me repeating what maximod has said, but that's pretty much what I would have said.
Gaps 2` to 6 and probably 7 are most certainly due to MW broadcast stations both in the UK and Germany & Holland, considering your location.
Gap 1, could be a number of things from Switch Mode PSU's to a possible aeronautical/maritime beacon or a modem - msan/dslam oddity.
A number of x47c's gaps could be Switch Mode PSU and harmonics.
GrahamC
Grafter
Posts: 257
Registered: 19-07-2009

Re: ADSL2+ Tones & Bitloading

Sorry for the late reply guys...some interesting thoughts although I'm still not convinced on the "noise" being attributable to AM frequencies, or indeed if it is noise induced.
I have included a DMT screenshot for direct comparison with that of x47c's, and a snapshot of both Tone/SNR graphs for a clear side by side view.
One reason I'm sceptical of the AM interference stance, is that I've just switched from 20CN/ADSLMax to 21CN/ADSL2+ and have access to archived bitloading graphs from the same router over the same line for both the DSLAM and MSAN (one is included below for ref.).
The 20CN/ADSLMax graphs show absolutely NO radio interference and NO gaps in the spectrum. As 4 out of 7 of the tones mentioned above appear in both the 20CN/21CN frequency spectrum, I cannot believe that several new AM stations are broadcasting since I happen to have upgraded to 21CN/ADSL2+ or that my line has degraded to such a point it is now susceptable to radio interference.
@x47c, can you let me know your general geographical area, I'm in East Yorks, quite literally just East of York  Smiley
Unless x47c is a neighbour of mine I couldn't accept local noise or aero/maritime frequencies as a possible source.
So, for me the question still remains.
x47c
Grafter
Posts: 878
Registered: 14-08-2009

Re: ADSL2+ Tones & Bitloading

I'm in Oxfordshire
Edit:
I've also recently transferred to ADSL2+/21CN
I've now just dug out my DMT stored images from the days of 20CN - see graph attached
Like you I also see no gaps on 20CN (apart from the pilot tone).
x47c
Grafter
Posts: 878
Registered: 14-08-2009

Re: ADSL2+ Tones & Bitloading

Sometime it's also helpful to look at the Quiet Line Noise (QLN) - if your router will supply it out to you via its telnet access.
This is the noise on the line found by the router in its initialization process and before any ADSL signal is applied.
Obviously it is only collected/stored at the moment of switch on/re-sync.
Attached is mine - showing the spikes corresponding the the dips in SNR and bitloading displayed in the DMT program..
GrahamC
Grafter
Posts: 257
Registered: 19-07-2009

Re: ADSL2+ Tones & Bitloading

Interesting...
not only that you've recently moved to ADSL2+, but the data supplied from your time on Standard DSL reinforces my thoughts on this being equipment related.
After a quick refresh on graphs in Excel  Embarrassed I've attached my QLN graph, which also seems to support the hypothesis.
Your data clearly shows AM interference over several tones, denoted by the curved spikes, as you have previously stated. Mine shows no such behaviour.
Community Veteran
Posts: 19,100
Thanks: 437
Fixes: 21
Registered: 31-08-2007

Re: ADSL2+ Tones & Bitloading

Well, apologies to you x47c, that will teach me not to get up too early in future. I looked at your DMT plot and missed the words "the same gaps" in your post  Embarrassed  Now Graham has posted his alongside yours it stands out a mile. It's also clear the 21CN MSAN and 20CN DSLAM are behaving differently. My earlier comment was based on Graham's list rather than a graph, and glancing at x47c's graph I ignored the curved dips/depressions which are typical of broadcast interference and just looked at the pattern of missing tones.
The reason you don't see the curved dips Graham is because you are much much closer to the exchange, a short line, probably underground, whereas x47c's is much longer, showing the same dips as 20CN and I'd guess he may have some overhead.
Why the MSANs are producing the results they are is curious to say the least. Thinking cap on.
A couple of questions to save me re-reading and digging through threads that may not have the answer, roughly what time of day did each of you sync at?
Are you both using the same or different modem/routers on 20CN as 21CN and what were/are they and do you know what chipsets they use?
x47c
Grafter
Posts: 878
Registered: 14-08-2009

Re: ADSL2+ Tones & Bitloading

A quick reply from me -I've got to go..................
No need to apologise - I don't mind
Yes same Dlink router used on 20CN and 21CN - broadcom chipped
No idea when I re-sync'd.  Under 20CN the unit stayed in sync for multiple weeks at a time.
I think the 21CN DMT image was taken shortly after I switched on/re-booted after a time away so probably mid morning.
My line is rural (2.4Km) but underground till joint box on D side.  Then up pole, then around 50 yards total of drop wire to house, along eaves and down to NTE5 in hall. Being on the ridge of a hill I guess my dropwire is more suseptible to RFI.
GrahamC
Grafter
Posts: 257
Registered: 19-07-2009

Re: ADSL2+ Tones & Bitloading

All data so far posted (20CN and 21CN) is from the same router, a Zoom X7n, Broadcom chipset.
Currently the router has been sync'd just over 12 days, and last resync was around Noon on 30th May.
I'm on what can only be described as a small "market town" exchange (~3000 lines). My particular line is all underground, approx 1km from the exchange.
Exchange equipment is identified as IFTN - Infineon, which I beleive is a derivative of the Texas Instruments equipment, as Infineon bought the TI DSL CPE business back in '07.
GrahamC
Grafter
Posts: 257
Registered: 19-07-2009

Re: ADSL2+ Tones & Bitloading

Quote from: Anotherone
Now Graham has posted his alongside yours it stands out a mile.

Exactly, hence my initial enquiry. I find it strange that something like this does not have a "detailed explanation" or "contrived conspiracy" somewhere on the interweb.
Quote from: Anotherone
The reason you don't see the curved dips Graham is because you are much much closer to the exchange, a short line, probably underground, whereas x47c's is much longer, showing the same dips as 20CN and I'd guess he may have some overhead.

Looking at it another way, it is that my line does not suffer from RF interference at all, and should not have any tones marked as unusable.
Quote from: Anotherone
Why the MSANs are producing the results they are is curious to say the least. Thinking cap on.

Thinking cap on, Indeed  Wink
Bald_Eagle1
Grafter
Posts: 313
Registered: 24-06-2011

Re: ADSL2+ Tones & Bitloading

If it is of any interest whatsoever & if you are Windows users & if you have access to a Netgear DG834GT, 834DGN and/or possibly other Netgear routers and/or a Huawei HG612 modem, you may find the scripts for various flavours of ADSL connections, available from this link of some use:-
http://forum.kitz.co.uk/index.php/topic,11216.0.html
They can provide a snapshot of your connection at any given time, or provide ongoing statistics graphs over a minute by minute, hour by hour, day by day or month by month period.
I have attached a couple of typical graph set examples.

Note that in the "snapshot" graphs QLN & Hlog data is only updated during the training period of a resync.
However, Bit-loading & SNR are constantly updated throughout the life of a connection.