dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1394
share rss forum feed

funkadelic

join:2000-08-02
Fremont, CA

slower speeds at NID than in the house?

hi,

i have single line Fusion (since late 2010) and was getting stable & consistent 14-15Mbps download speeds until maybe the end of November 2011 when it dropped down to 12 Mbps, so I thought that maybe something changed inside my home which was creating new interference, so I decided to test at the NID to confirm my suspicions.

i have a home run (shielded cat5e) from the NID/MPOE to the phone jack inside my house and then another 15' shielded cable to the actual modem. I have the Corning xDSL splitter (little bit over 1 year old): »www.hometech.com/hts/products/wi···dsl.html

using the speedtest.sonic.net page, i get 12 Mbps down inside the house, and when I connected directly to the NID test jack, i can only get 9-10 Mbps down.

Shouldn't I get better speeds at the NID test jack than inside the house?

using DMT-tool, here's a graph of the DSL tones inside the house



and here is the result at the NID test jack:



could the test jack be faulty, but the rest of the splitter still ok? trying to understand why i'm getting faster speeds inside than outside....

for reference, here's a graph from the NID back in september 2010:



DaneJasper
Sonic.Net
Premium,VIP
join:2001-08-20
Santa Rosa, CA
kudos:9
That's really weird. How are you connecting the modem to the test jack?

-Dane


leibold
Premium,MVM
join:2002-07-09
Sunnyvale, CA
kudos:10
Reviews:
·SONIC.NET
reply to funkadelic
The one immediately obvious difference is that the first and third chart (taken inside) are stable connections while the second chart (taken at the NID) has a large spread in SNR. This despite the fact that in all three cases the modem was monitored only for a very short time.

Guesses for possible reasons:
- poor quality (not twisted pair) cable from test jack to modem
- poor contact (cable not fully plugged in on either end)
- cable and/or modem being moved around while the line is retraining (with loops or sharp bends in the cable effecting noise levels)
--
Got some spare cpu cycles ? Join Team Helix or Team Starfire!

klui

join:2001-11-08
Castro Valley, CA
Reviews:
·SONIC.NET
said by leibold:

- cable and/or modem being moved around while the line is retraining (with loops or sharp bends in the cable effecting noise levels)

Which leads me to ask: how important are the wires not exposed out of their sheathing when they're connected to the NID? I've always heard that when terminating CAT5/CAT6 cables, you need to preserve the twists and be out at most 0.5 inches from the outer encasing or else error rates would unnecessarily increase. But often times I see the telco guys just strip the heck our of their wire (CAT3?) and lay at least several inches of bear wire with twists removed then into the line interface module.

»en.wikipedia.org/wiki/File:Apl_germany.jpeg

Is it because DSL speeds (less than 50 Mbps) are not as important?


leibold
Premium,MVM
join:2002-07-09
Sunnyvale, CA
kudos:10
Reviews:
·SONIC.NET
You are right about "telco guys" not to be overly concerned with the strict rules for Cat5+ wiring. This is most certainly due to the fact that for an analog voice connection you only need about 5kHz (that is all it takes for human speech) and at such low frequencies twisted pair cable is very forgiving.

DSL is up to 1.1 MHz for ADSL1/2 and 2.2 MHz for ADSL2+ and therefore much more sensitive the analog voice. Removing the outer insulation isn't that big of a deal (it does help to maintain a minimum distance to other nearby wires/cables but it is not providing any form of shielding). For the best results the exposed wires should retain the original twist rate. Unnecessary untwisting is just as bad as the opposite: giving the pair a lot of extra twists (because someone heard that untwisting is bad and thinks that extra twists must be good). The main reason for the Cat5/Cat6 rule about not removing too much of the cable insulation is that it is very hard to maintain the correct number of twists once the insulation is removed. However those rules are made for 1G and 10G Ethernet signaling at frequencies 100 times higher then DSL frequencies.
--
Got some spare cpu cycles ? Join Team Helix or Team Starfire!

funkadelic

join:2000-08-02
Fremont, CA
reply to DaneJasper
said by DaneJasper:

That's really weird. How are you connecting the modem to the test jack?

i've tried 2 different plain jane (non-twisted pair) phone cords and had the same results

this is what my "test" setup looked like



funkadelic

join:2000-08-02
Fremont, CA
reply to leibold
said by leibold:

The one immediately obvious difference is that the first and third chart (taken inside) are stable connections while the second chart (taken at the NID) has a large spread in SNR. This despite the fact that in all three cases the modem was monitored only for a very short time.

The 3rd chart was actually outside at the NID taken 9/2010, not inside. The 2nd chart is at the NID as of earlier this week.

said by leibold:

Guesses for possible reasons:
- poor quality (not twisted pair) cable from test jack to modem
- poor contact (cable not fully plugged in on either end)
- cable and/or modem being moved around while the line is retraining (with loops or sharp bends in the cable effecting noise levels)

i tried 2 different non-twisted pair phone cords, one of which was brand new. I guess I should try with my shielded twisted pair cord as well, just to eliminate that as a factor.

the cable was fully connected on both ends when i was doing the testing, and nothing was moved during testing (the pic in the previous post was how it was situated)

any other thoughts on what's giving me bad results at the NID? i'll run the tests again with the STP cable and see how that goes.


leibold
Premium,MVM
join:2002-07-09
Sunnyvale, CA
kudos:10
Reviews:
·SONIC.NET
reply to funkadelic
said by funkadelic:

i've tried 2 different plain jane (non-twisted pair) phone cords and had the same results

That test setup looks fine, but if you still have the UTP (unshielded twisted pair) phone cord that came with the modem (most DSL modems include one) I would suggest to try that one.

Shielded twisted pair (STP) can help to keep noise out, but also weakens the signal strength at higher frequencies.
--
Got some spare cpu cycles ? Join Team Helix or Team Starfire!