Print

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

[sevenlayermuddle]

Since installation of my FTTC, the downstream speed has consistently tested at 44 to 45-ish.  Ten minutes ago, the same result - 45.51 on the BTW speed test.  So I dug out a fairly expensive purpose made Cat 5 modem modem lead I bought a while ago, branded 'Belkin Pro', too see what difference it would make.  And...

Downstream dropped by over 3Mbps.  Repeated three times, always the same, a bit over 42.   

I wondered if DLM might have kicked in as a result of my resyncs, so reconnected again with the cheap-looking flat cable provided by BT, DS tested at 46.57.   

Upstream was much the same, 7.5-ish with either cable.

So, consistently, the cheap flat cable is getting significantly faster downloads than the posh 'Belkin Pro' Cat 5 .  I would not of course advise anybody else to necessarily expect the same effects, the Cat 5 ought to do better.  But it does go to show, IMHO, we can theorise all we like about what is scientifically optimal, but in the real world random factors do often conspire to surprise us as to what actually works best.

I have no means of getting line stats at the moment, so can't investigate in more detail.   One possible clue might be that the Belkin cable, although terminated with RJ11s is fully loaded Cat 5 with 4 pairs, thus has redundant wires perhaps adding to the interference pickup?  The Cat 5 is also slightly longer at 4 feet vs  1 metre, but it's hard to believe that would make any odds.

[roseway]

I'm no cabling expert, but I would say that Cat5 cable is wholly unsuitable for a modem lead for purely physical reasons (thickness of cable and size of connector). Apart from that, is it a screened cable perhaps? I could certainly believe that this might degrade the performance.

[JGO]

So, consistently, the cheap flat cable is getting significantly faster downloads than the posh 'Belkin Pro' Cat 5 .

But it does go to show, IMHO, we can theorise all we like about what is scientifically optimal, but in the real world random factors do often conspire to surprise us as to what actually works best.

Where is the screening connected ?   I suspect it isn't  - as few modems have a screened box.  If so it would act something like a (capacitivly coupled) Ring wire !!

[burakkucat]

I assume the lead in question is that which connects the SSFP (RJ11 or RJ45 port) to the VDSL2 (RJ11) port of the modem?

Is that cable wired such that a twisted pair is used between 3 & 4 at either end?

The cable is labelled with the Belkin name . . . which is a good reason for me to be suspicious of its quality!

[sevenlayermuddle]

Here's a link.

As you will see, it is supposedly purpose designed as a modem cable.   I don't think it is screened. The legend written on the cable reads 'High speed internet modem cable'.

http://www.amazon.co.uk/gp/product/B004SMNT7W/ref=oh_aui_detailpage_o08_s00?ie=UTF8&psc=1

Note the generally favourable Amazon reviews... 

Within the bounds of common sense, I would generally expect a genuine Cat 5 cable to outperform any flat cable, purely because of the twists.   Belkin may be a naff name, but as it does say on the cable, 'cat  5' I suspect it does meet cat 5 twist requirements.   ISTR there was a small but demonstrable improvement in ADSL1.

[roseway]

Hmm.. extract from the first of those customer reviews:

"As this cable is cat5e cable the attenuation has been lowered dramatically and SNR vastly improved, Reporting no errors on the Draytek since changing over."

I think we all know what to make of that.

[burakkucat]

Hmm.. extract from the first of those customer reviews:

"As this cable is cat5e cable the attenuation has been lowered dramatically and SNR vastly improved, Reporting no errors on the Draytek since changing over."

I think we all know what to make of that.

 

Now having looked at the link which 7LM has provided and progressed to this page, we can see that cable appears to have two (twisted) pairs connected -- 2 & 5 and 3 & 4. And that may be sufficient to account for the observed performance failing. 

[sevenlayermuddle]

we can see that cable appears to have two (twisted) pairs connected -- 2 & 5 and 3 & 4. And that may be sufficient to account for the observed performance failing. 

Yes, the cable has all 4 pins of the RJ11 connected at each end.   

But I would not expect the fact that they are connected to make any difference (beyond the presence of the redundant pairs, as already mentioned) unless the modem and/or SSFP RJ11 sockets also had anything connected to the redundant pins?  I can confirm that for the ECI modem, as the modem's socket has only two pins physically present.  I can't confirm whether it applies to the SSFP as there is a big cupboard in the way that stops me getting a close look, but I'd be very surprised if more than two connectors were seen.   

edit, wrong smiley!

[les-70]

  I can't see that the extra pair will make any difference at all.  I would expect the best xsdl cable to 1. have same impedance as the drop wire and cable to the master socket and 2. ideally to be twisted.  Impedance is hard to judge by looking at a cable.  The best may be proper BT cable with carefully done do it yourself RJ11's added to the right length of cable.

[NewtronStar]

I can't see that the extra pair will make any difference at all.  I would expect the best xsdl cable

Those extra unused pairs will act like an ariel and pickup interference and disturb (transfer the interference) onto the used pair.

[tommy45]

I can't see that the extra pair will make any difference at all.  I would expect the best xsdl cable

Those extra unused pairs will act like an ariel and pickup interference and disturb (transfer the interference) onto the used pair.

If that was the case then by the same token you would expect the same thing from the BT wiring to the master socket, as most of that is 2 or 3 pair cable, or am i missing something?
Maybe the belkin cable is just poor quality or has a defect of some kind,
I changed the modem cable (supplied flat cable) to a twisted pair 0.5mtr cable from Tandy,(adsl nation)
And personally cannot tell any difference in snr attainable rate, or error rates

[sevenlayermuddle]

Les's suggestion of impedance transition is probably the most technically convincing for me.

But personally I'll stick with my original theory that it is, to some extent, a black art.

One thing that might be coming into play is the slightly different cable lengths.   Acting as antennas, the different length, and thus resonance, could be tipping the balance towards reducing a problematic interference frequency, or amplifying a weak tone.  Such effects are of course impossible to predict, but can sometimes produce unexpected benefits or detriments.

Another possibility is that the extra stiffness of the Cat 5 held it further off the wall, pushing it into a 'hot spot' of higher interference.   And again, random and unpredictable.

For those of us old enough to remember 405 line TV on VHF, it is a bit like the best way to get to see BBC1 was sometimes to balance the antenna on the mantlepiece and then persuade Gran to stand poised in the doorway, acting as part of the complex system of reflections.   Whereas, for ITV, or if it was raining outside, she had to stand in the far corner of the room rather than the doorway.

[NewtronStar]

If that was the case then by the same token you would expect the same thing from the BT wiring to the master socket, as most of that is 2 or 3 pair cable, or am i missing something?

The difference between the BT wiring is the unused pairs are not terminated to anything were as the Belkin unused pair looks like both ends of the RJ11 plugs are terminated to pins 2 & 5 most DSL cables have only a single pair and use pins 3 & 4.

It would be my preference just to use standard single pair DSL cable from modem to Master Socket.

[sevenlayermuddle]

The difference between the BT wiring is the unused pairs are not terminated to anything were as the Belkin unused pair looks like both ends of the RJ11 plugs are terminated to pins 2 & 5 most DSL cables have only a single pair and use pins 3 & 4.

That was mentioned earlier in this thread.   As said then, it would only be relevant if the modem and/or the SSFP also has anything connected to the redundant pins.  The modem ( ECI) certainly does not have anything connected, as the socket itself only has two connectors present.   Access problems prevent me from confirming same for SSFP, but I would be very surprised to find that the redundant pins were connected therein. 

I'm actually getting more interested in this as I have experimented more today and get the same results from a variety of cheap-n-chearful freebie untwisted cables - some are better than others but results are surprisingly repeatable, and all the flat untwisted cheapos surpass the Cat 5 for downstream speed tests.   The Cat 5 has the edge for upstream, by about 0.3 to 0.5 Mbps.

The  most convincing scientific theory I have seen is that of impedance mismatch.   There is also the possibility that the Cat 5 cable is screened which, I agree would be bad, as it is not grounded.  I may be tempted to carve a slit in the insulation to see inside so as to confirm or deny that possibility.   

If we are to conclude that Cat 5 really is simply inferior to flat (untwisted) extension cabling for extending the modem connection, I suspect the consensus of advice in a number of past threads might merit reconsideration...? 

[NewtronStar]

I suspect the consensus of advice in a number of past threads might merit reconsideration...? 

Lets just say there is a lot of misinformation out there on the WWW, the good thing about the KITZ forum we do are own tests and report back into our findings, rather than taking an unkown Web Sites consenus at face value 

edit: reason 1 > typo