Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Author Topic: SNR vs frequency peak  (Read 2286 times)

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
SNR vs frequency peak
« on: March 25, 2019, 06:49:26 PM »

Here is an extract from the SNRM vs freqency data obtained from my modem. DSL tone numbers vs SNRM. See the peak at tone 125:

Code: [Select]
   113 22.9375
   114 22.8125
   115 22.3750
   116 21.8125
   117 21.1250
   118 21.1250
   119 20.3750
   120 19.8750
   121 18.3750
   122 17.7500
   123 13.0625
   124 10.5625
   125 8.8125
   126 11.3125
   127 13.7500
   128 15.1875
   129 15.1250
   130 16.2500
   131 16.1875
   132 15.8750
   133 15.1875
   134 15.3750
   135 15.1875

What could that be? That would equate to what -539 kHz?

Thing is, I can’t see a corresponding thing in the QLN.

That doesn’t prove that it is noise as it could be reduced received signal level too - frequency-dependent increased attenuation, a loss of power or cancellation - is that even possible?

What am I looking at?
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: SNR vs frequency peak
« Reply #1 on: March 25, 2019, 09:12:59 PM »

The extract you have shown is not of SNRM but SNR per sub-carrier.

Let's check the calculation . . . 125 x 4.3125 kHz per sub-carrier = 539.0625 kHz

Either the signal strength has been reduced or the "noise" strength has been increased. You will be looking at a dip (not a peak) in the SNR plot. What is shown in the bit loading plot for that frequency?
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: SNR vs frequency peak
« Reply #2 on: March 25, 2019, 11:36:09 PM »

Sorry, brain error regarding SNRM, I wasn’t thinking straight but I suppose that I could have realised that that didn’t make sense if I had bothered to engage brain first. I’m glad you pointed that out.  ???

What is odd is that I can’t see any corresponding ‘feature’ in the bit loading. I’ve just rechecked and it’s just featureless, gently declining monotonically, as we expect.
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: SNR vs frequency peak
« Reply #3 on: March 25, 2019, 11:53:24 PM »

I have also checked for usage of that frequency by broadcasting stations and could only find a Belgian station. I think that can be ruled out!

As there is no corresponding perturbation in the bit-loading and QLN plots, I suspect it is going to remain a mystery.
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: SNR vs frequency peak
« Reply #4 on: March 26, 2019, 03:21:05 AM »

I found the Belgian station too.

What is you opinion of the width of the anomaly? It does not seem to be quite a sharp peak as we usually see because there is some region around it which is more cone-like, roughly 3dB per tone spacing over several tones in either side.

I am quite baffled.
Logged

aesmith

  • Kitizen
  • ****
  • Posts: 1216
Re: SNR vs frequency peak
« Reply #5 on: March 26, 2019, 07:58:55 AM »

Do you see the same on all your lines?
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: SNR vs frequency peak
« Reply #6 on: March 26, 2019, 05:05:45 PM »

What is you opinion of the width of the anomaly? It does not seem to be quite a sharp peak as we usually see because there is some region around it which is more cone-like, roughly 3dB per tone spacing over several tones in either side.

Not having sight of the plot, it is rather difficult for me to comment.  :)  However the preceding question (above) is relevant. Is the anomaly "present and correct" on all four lines?
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: SNR vs frequency peak
« Reply #7 on: March 27, 2019, 01:03:46 AM »

On line 1 there is a tiny 3dB deep depression centred on tone 102, again cone-shaped/symmetrical-triangular.

On the other lines just some minor irregularities.

So the original pattern has vanished altogether. It was just an unidentified fleeting artefact.  :-[ ???

It felt very fishy given that I could not match it against other stats in the same modem.
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: SNR vs frequency peak
« Reply #8 on: March 27, 2019, 09:31:12 PM »

b*cat nods, knowingly. There are so many temporary, minor, perturbations in the average xDSL circuit's statistics that it is best to view the data with an "it may not be explainable" mind-set.

I hardly ever look at my own . . . as long as the service works, I just make use of it.  :)
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Re: SNR vs frequency peak
« Reply #9 on: March 27, 2019, 11:31:12 PM »

Quite so. It was the depth of it that led me astray though in this case. Weird.
Logged

DiggerOfHoles

  • Member
  • **
  • Posts: 64
Re: SNR vs frequency peak
« Reply #10 on: May 06, 2019, 05:03:13 PM »


Back of a fag packet calculation for your 539Kz Quafor Poobel Belgie. Belgian haircut of BINs


You have Belgian squatters in BIN 125 with their belongings scattered either side.

If you evict them, make line longer or shorter, they will only move BIN along with there tat.

Unless the router or house wiring etc are picking up 539KHz or it's harmonics you can't fix it.

Rough harmonic calculations.
Integers are bad news.

Line length 7000M

539 = 12.98
417 = 16.78
348 = 20.11
270 = 25.92
139 = 50.35

Line length 7007M

539 = 13.00
417 = 16.80
348 = 20.13
270 = 25.95
139 = 50.41

Line length 7100M

539 = 13.17
417 = 17.02
348 = 20.40
270 = 26.29
139 = 51.07

I suspect you are 7.007Km from the exchange.
Logged
 

anything