Print

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

[Weaver]

As many if you will know, I have three lines which are over "63.5" dB d/s attenuation. I think my line is 4.6 miles long (7400 m) on the assumption that it follows the road which seems to be the only practicable route. Given that 63.5 dB is the reporting limit in some protocol within the DSL system, in that case, any ideas about what the true attenuation figure might be?

Apologies if I've posted on this topic before, I forget.

I've tried various tools, but most of them simply say that 4.6 miles is off the scale or won't handle values higher than 63.5 dB.

[roseway]

I seem to remember that the old 2-Wire 2700HGV used to report the actual attenuation even if over 63.5 dB.

[aesmith]

 I've seen people reporting higher figures from Netgear kit for example.   For what it's worth we show 57dB for 5.3km so if you scale that up to your line length you'd get something like 76dB.   However it won't be linear as the higher frequencies will drop out quicker, so you presumably won't be using them.

[Weaver]

@roseway - thank you I had forgotten that.

@AESmith - I was wondering about a tool that could do extrapolation to longer lengths correctly beyond 63.5 dB, taking into account the effects you mentioned.

Why do tools and graphs published simply give up, believe that service over such line lengths is impossible? Is it perhaps the case that service would normally be impossible without a very, very good metallic path, one that's unusually good? Some of my neighbours have much more difficult conditions, on lengths 1 km longer, getting sub 1 Mb/s service.

[Weaver]

@everyone - If you have a line with such an attenuation or similar length or worse, would you post something telling me/us all about it?

[aesmith]

By the way, I should have asked him where it originated but the OR guy who did the SFI visit has a graph on screen with speeds vs distance, and he was coming up with around 4meg at 5 km, pretty similar to what we actually get.   I'm sure 5km wasn't the end of his range, but I can't say I noticed whether it went up to your 7+ km.

[Black Sheep]

It's not the distance so much, as the actual attenuation, as I'm sure we're all aware. Unless the cable size (poundage) is known, then it really is pure guess-work ??

I tend to use 13dB attenuation per Km, as an approximate guide for 0.5 Cu cable. I have no converters or tables to tell me what that same Km's attenuation would be, if it was run in 0.9 Cu, or conversely 0.4 Al ?? As there is a damn site more 0.5 Cu than Al in the D-side network, then I would just use the 13dB guide. 

[Weaver]

@BlackSheep  - woah :-  7.4 × 13 = 96.2 dB ! Sounds like I'm in trouble if I'm on 96dB d/s attn.

It shows you how inadequate a lot of the tables graphs and tools are if they pack up somewhere around the 63.5 dB - 75 dB point.

[Black Sheep]

Ha ha ...... believe me, you won' have an attenuation of 96dB and be receiving DSL over fixed-line cables.  .

This is why it is paramount to know the exact distances involved combined with the exact make-up of the cable. Long D-side line lengths have often been made up of a hybrid of higher poundage cables, such as 0.63 Cu and 0.9 Cu, which would bring your attenuation right down to a, 'within range' integer.

Ezzer (an ex-BT engineer with a wealth of knowledge), along with other educated members of the forums had a mini-discussion here ... http://forum.kitz.co.uk/index.php?topic=1448.0 .... about something similar. You will notice he had a line in-synch at 9.6Km (I think he said ?), but this was fed over mainly 0.63mm Cu cable.

My own personal experience has seen me attaining synch (ADSL) on an 86dB cable-run, albeit only at around the 0.3meg mark. Spent days wih another engineer, clearing umpteen battery contact faults in wet buried joints up a very big hill, in order to get a 'Line Test OK'. Knowing full well, that even when we had achieved this LTOK result, the circuit would struggle like hell to synch. But them's the rules, the ISP's require a LTOK and quite rightly so.

[Weaver]

Well the 4.6 mi figure is accurate presuming that the cable follows the road, the only other option is for it to go over the moor. The road route is just due north then due west, two roads and you're at the exchange. You can see the cable by the side of the road on the northbound section.

So that leaves the 13 dB figure?

[aesmith]

Does your ISP have a figure for your line length?   In among the figures that Pxxxxxx paste into my fault ticket, there's often a couple of length figures which my case are ..
  Line Length (DN)   5300
  Line Length (DP)   5300

[burakkucat]

4.6 miles = (4.6 x 8) / 5 km = 7.36 km

Are you able to harvest the Hlog data for one of your lines? If so, I would appreciate a copy of the raw data (that is the list of the individual attenuations (in dBm) per 4.3125 kHz wide tone).

[Weaver]

> Are you able to harvest the Hlog data for one of your lines?

Beyond my physical abilities, I'm afraid. (As explained earlier, you'll recall.)

It's not one of the parameters that A & A's "clueless" server harvests when it extracts info from modems remotely (somehow) to prepare me a report. It pulls out sync rates error counts hec errors, errored seconds - the usual stuff.

(I'm assuming that it would require telnet access into the modem ?)

[Weaver]

A perfect job for a raspberry pi - excuse for buying one needed - except that it would need more Ethernet ports on it.

[Weaver]

I used https://www.freemaptools.com/measure-distance.htm

And 1 mi = (1760 × 3 × 12 × 0.0254) m  :: exactly
              = 1609.344 m