Kitz Forum
Broadband Related => ADSL Issues => Topic started by: konrado5 on June 03, 2018, 10:19:44 PM
-
Sometimes I have small noise on tone 453. It is noticeable on attached QLN. I've had about week and a half very stable ADSL connection on 1.1 dB snr margin. However, this small noise caused reconnection. What can be source of this noise on this one tone?
-
Sub-carrier number 453 is centred upon a frequency of 1.95 MHz. In what band plan does that frequency fall? Ah, yes. We've been here before, it's the Amateur Radio "top band" (1.81 - 2.00 MHz). So my suspicion is that it is a licensed transmission or beacon in that frequency band.
How have you proved that signal was the reason for your circuit's re-train? Surely it is just a mere co-incidence? To prove that it causes your circuit to re-train, you need to know the source of that signal and, with the co-operation of the person responsible, arrange to have it turned off and then back on again. If, after a repeated sequences of the above, you see that every time the signal appears your circuit re-trains then you can claim proof.
To avoid circuit re-trains you will need to increase the target SNRM. Try with it set to 2 dB.
-
How have you proved that signal was the reason for your circuit's re-train? Surely it is just a mere co-incidence?
I was observing SNR per tones when I saw increasing CRC errors number. Always II saw notch on SNR graph on this tone.
To avoid circuit re-trains you will need to increase the target SNRM. Try with it set to 2 dB.
It is curious that I have significantly less errored seconds on SNR margin 1.8 dB than on SNR margin 6 dB. On SNR margin 6 dB I have about 100 errored seconds per hour. On SNR margin 1.8 dB I have about 30-50 errored seconds. On SNR margin 1.8 dB I have slightly lower power output (18.4 dBm instead 18.6 dBm).
-
Over the years, we have all become accustomed to the fact that you have a rather unique circuit in the way that it operates. In theory, an xDSL circuit operating in Poland should be broadly identical to an xDSL circuit operating in, say, Spain or the UK.
-
I start getting errors when I get down to that kind of SNR margin. My SNR did, until changing modems recently, sometimes go rather lower than that, down to 0.8dB or 0.6dB occasionally. This was starting from a supposed target downstream SNR margin of 3dB, but with the older modem model the observed SNRm could be quite a bit lower than 3 dB soon after a resynch.
-
Sub-carrier number 453 is centred upon a frequency of 1.95 MHz. In what band plan does that frequency fall? Ah, yes. We've been here before, it's the Amateur Radio "top band" (1.81 - 2.00 MHz). So my suspicion is that it is a licensed transmission or beacon in that frequency band.To avoid circuit re-trains you will need to increase the target SNRM. Try with it set to 2 dB.
This is a "SHARED" frequency band, and above 1850 kHz amateurs are restricted in power, i.e. there are other almost certainly more powerful transmitters which will stick to a particular frequency. Suggest your suspicion is only one possibility.
-
To avoid circuit re-trains you will need to increase the target SNRM. Try with it set to 2 dB.
Unfortunately xDSL doesn't allow to set different target SNR margin on different tones. It would be good idea to set target SNR margin 6 dB at HAM band and 1 dB on other tones.
-
Now I have wide noise in HAM band.
-
Is that not rather too wide for the HAM band around 136kHz ? That band is really narrow according to Wikipedia.