Kitz Forum
Broadband Related => ADSL Issues => Topic started by: konrado5 on June 27, 2018, 07:32:22 PM
-
I've noitced that intermittent noise is more often in HAM band (tones 400-483) than in other bands. I know this is band for amateur radio however as I know nobody in Poland (all the more in my town) use this band for amateur radio.
Furthermore, at one time (at this time I had disabled HAM band) I heared buzzing on AM radio and I had reconnection. I've noticed that most increase on QLN graph is in the tones 400-483. It seems that my circuit is the most susceptible to any wideband noise on these frequiences. I would expect that circuit is the most susceptible to noise on the end ones (483-511) because as I know the higher frequency the worse longitudinal balance.
I suspect it is becasue I am only user on DSLAM with these frequiences unblocked and it results with no crosstalk on these frequiences. On the other frequiences crosstalk attenuates other type of noise. Is it possible?
Best regards
konrado5
-
I don't understand your use of the term ‘attenuate’. ;D ‘Attenuation’ means multiplying by a positive number that is less than 1, so the value is reduced. It is not the same as obscuring or covering up something. As of course you know, crosstalk is always additive, that is it is a special type of noise that is strongly correlated with your own data signal or some other data signal.
-
I don't understand your use of the term ‘attenuate’. ;D ‘Attenuation’ means multiplying by a positive number that is less than 1, so the value is reduced. It is not the same as obscuring or covering up something. As of course you know, crosstalk is always additive, that is it is a special type of noise that is strongly correlated with your own data signal or some other data signal.
I mean crosstalk reduce other noise.
-
Perhaps a problem, with English Konrado, my friend. :-) But crosstalk cannot reduce other noise. I would need you to explain - help me understand a bit. Your English is superb by the way. My Polish is a source of shame, given that it has now taken second place in ranking amongst the most commonly spoken languages of Scotland.
-
`i would need you to explain - help me understand a bit.
I mean that if my router receive crosstallk then it receives fainter other noise than without crosstalk.
-
I would say that the level of other, non-crosstalk, noise received should be the same, be independent of whether you are getting crosstalk or not. What do you think?
I cannot think of a mechanism by which the two could be linked. Others may be able to think of something? Do you have something particular in mind?
-
I would say that the level of other, non-crosstalk, noise received should be the same, be independent of whether you are getting crosstalk or not. What do you think?
I also think that level of received non-crosstalk noise should be the same when router receive crosstalk. However I have observation indicating it is dependent on received crosstalk. I have described it in the first post of this thread.
-
@burakkucat What do you think?
-
There is the wanted signal and the unwanted signal(s). For simplicity we refer to the latter as noise.
Cross-talk is, by definition, an unwanted signal. A lawful broadcast transmission is an unwanted signal. The radio-frequency signal emitted by a poorly constructed switching PSU is an unwanted signal. The high-frequency "mush" emitted by an old television set -- one that is faulty -- is an unwanted signal. The cumulative sum of all those unwanted signals is the noise.
I cannot see how any one unwanted signal from one source can be responsible for the decrease in another unwanted signal for another totally unrelated source.
I would say that the level of other, non-crosstalk, noise received should be the same, be independent of whether you are getting crosstalk or not.
I also think that level of received non-crosstalk noise should be the same when router receive crosstalk.
I believe, therefore, that I am agreeing with you both. :)
-
I cannot see how any one unwanted signal from one source can be responsible for the decrease in another unwanted signal for another totally unrelated source.
But my observations indicates this is a case. Look at my two QLN graphs. On first graph (with noise causing buzzin on AM radio) there is noise on end tones. However difference between two graphs is the largest on the 400-483 tones (with very little crosstalk).
-
I will not dispute your observations. However it appears that your interpretation of the observations is lacking.
-
However it appears that your interpretation of the observations is lacking.
I don't understand.
-
You have made certain observations and have shown the evidence.
However your attempted interpretation of the observations cannot be correct.
-
How about QLN being more accurate on tones wihout crosstalk? Perhaps my router see better other type noise on tones with very small crosstalk.
-
As long as the other user(s) are connected, then cross-talk is a constant. EMI/RFI noise is variable.
You wouldn't normally spot crosstalk on QLN. Your line is an exception because you are in a truly unique situation whereby you are the only user on your MSAN who is able to use those notched tones.
As others have already said, one type of noise (crosstalk) does not decrease or mask noise from another source (RFI).
If you were not in the unique situation of being the only user who doesn't have the notching applied, then in theory I would anticipate that the QLN would show a peak at those tones during the time of RFI from the buzzing on the radio.
-
As long as the other user(s) are connected, then cross-talk is a constant. EMI/RFI noise is variable.
I know.
You wouldn't normally spot crosstalk on QLN. Your line is an exception because you are in a truly unique situation whereby you are the only user on your MSAN who is able to use those notched tones
I know.
As others have already said, one type of noise (crosstalk) does not decrease or mask noise from another source (RFI).
I suspected it is possible because of my observation.
If you were not in the unique situation of being the only user who doesn't have the notching applied, then in theory I would anticipate that the QLN would show a peak at those tones during the time of RFI from the buzzing on the radio.
Why peak at these tones? As I know higher frequency the worse longitudinal balance (sensitivity to noise). I would suspect the peak at end tones (483-511) and not 415-483.
-
If
Graph [1] is no Crosstalk and no RFI
Graph [2] is no Crosstalk but with RFI
Then
Graph [3] with Crosstalk and RFI in theory should look something like the blue line.
Please note the graph is NOT to scale, nor have I analysed individual tones, but just based on tones 432+ are affected more than tone 414. Its a very quick and crude projection.
-
I don't understand why these tones (432-483) are most affected by impulse noise. This is not HAM band noise. This is wideband noise (heared on AM radio on differerent frequiences).
-
The MSAN manufacturers are aware those particular tones can be problematic. That's why they decided to mask them in the first place by applying a notch.
What the cause is I've no idea, the MSAN manufacturer will know more.
-
A I know only reason why this tones are masked is that they are used by amateur radio (HAM band). But impulse noise heared on AM radio is not amateur radio.
It is evident on my graphs that QLN caused by impulse noise is not larger than about -120 dBm/Hz. It is not visible on tones with crosstalk significantly larger than -120 dBm/Hz. Perhaps buzzing on AM radio causes intermittent noise (not whole second) and crosstalk causes stable noise and thereby these noises are not summed. It is evident that crosstalk mitigates impulse noise effect on ADSL.