"shielded" twisted pair dsl modem cables
Denis McMahon
not as if it's grounded anywhere, and this bear of little brain wonders
if such shielding would (a) have any effect, and (b) if the effect could
even be detrimental.
The question is prompted, in part, by the fact that I get over 1 Mbit/sec
more using a "flat" rj11-rj11 than I do using a round shielded twisted
pair.
Could be that there's something fundamentally wrong in the twisted pair
cable too I guess, but I'm wondering if we're being sold a crock of shit
for the last 2 metres.
After all, it's unshielded twisted pair over most of the path from the
exchange. Surely all that STP on the last 2m is going to do is provide
more capacitive coupling between the two wires in the pair than
equivalent utp would, for no significant gain in noise reduction, tp
being fairly resilient to noise anyway.
Rgds
Denis McMahon
Peter Able
"Denis McMahon" <denis.m....@gmail.com> wrote in message
news:4dea4ca0$0$357$bed6...@gradwell.net...
Don't quote me Denis, but I could imagine that when the cable runs parallel
with cabling that might be carrying signals/noise there could be common-mode
induction that might be outside the modem's working range (I've seen it)
and, possibly some differential induction due to relative proximity -
although the latter is what twisting is meant to minimise! There are STP
cables where the connection of the shield continues through one or both
cable end connectors - which makes sense - but, as you say, many STP cables
just float the shield.
PA
Graham.
"Denis McMahon" <denis.m....@gmail.com> wrote in message news:4dea4ca0$0$357$bed6...@gradwell.net...
Your logic appears perfectly sound to me, and I suggest you trust your own
instincts rather than other peoples sales hype.
I've seen plenty of twisted pair RJ11 cables, there is one bundled
with every Cisco router I install in the course of my work. I don't recall
seeing a twisted one coming with a domestic grade router, they tend to
be parallel.
What I have never seen myself, is a shielded twisted pair RJ11 cable,
and as you say, where are you supposed to connect the shield?
--
Graham.
%Profound_observation%
Denis McMahon
> Your logic appears perfectly sound to me, and I suggest you trust your
> own instincts rather than other peoples sales hype.
>
> I've seen plenty of twisted pair RJ11 cables, there is one bundled with
> every Cisco router I install in the course of my work. I don't recall
> seeing a twisted one coming with a domestic grade router, they tend to
> be parallel.
>
> What I have never seen myself, is a shielded twisted pair RJ11 cable,
> and as you say, where are you supposed to connect the shield?
The shield is floated, in fact it's not brought out at the ends of the
cable at all.
The cable is a "belkin high speed adsl modem cable" which I purchased
because I wanted twisted pair all the way - but as I've said, I actually
get lower performance, 5500 kbits / sec vs 7000 kbits / sec, when I use
this cable compared to a "flat" (d section) rj11-rj11 cable for the last
2 metres.
So either the belkin cable is duff (and I don't have test equipment to
test beyond basic continuity / resistance) or stp with a floating shield
is actually detrimental .... and I remember enough of my electronics
engineering to realise that the capacitive coupling between a twisted
pair inside a floating screen will be much more complex than that between
a normal twisted pair.
Rgds
Denis McMahon
Graham.
"Denis McMahon" <denis.m....@gmail.com> wrote in message news:4debf766$0$13337$bed6...@gradwell.net...
between a satellite Rx and an 8 bit computer, I think the data was
9600bps with no handshaking. We has some special cable to wire
the link, one pair of wires with a screen. The pair was not twisted,
and the screen was left floating at both ends. I never could see the
point, and I am sure telephone cable would have done, It would have been
one less roll to carry in the car.
--
Graham.
%Profound_observation%
Denis McMahon
> years ago I was involved with a contract that required a data link
> between a satellite Rx and an 8 bit computer, I think the data was
> 9600bps with no handshaking. We has some special cable to wire the link,
> one pair of wires with a screen. The pair was not twisted, and the
> screen was left floating at both ends. I never could see the point, and
> I am sure telephone cable would have done, It would have been one less
> roll to carry in the car.
Yeah, and like I said before, my recollection of electrical engineering
from my ONC and HNC is that if you put a twisted pair inside a floating
screen, then not only do you have the mutual capacitance of the wires to
each other, but you'd also have capacitance between each wire and the
screen, so your total capacitance between the wires per unit length would
be c1 + c2/2 where c1 is the direct capacitance and c2 is the capacitance
between each conductor and the screen.
Or am I mis-remembering?
Rgds
Denis McMahon
The Natural Philosopher
not the shield is grounded.
The grounding of the shield helps remove external interference, thats all.
> Rgds
>
> Denis McMahon
Denis McMahon
Yes, I know all of that.
I'm speculating that stp with any shield will have *more* inter-connector
capacitance than utp.
I'm also speculating that an the ungrounded shield on the wall socket -
modem connection will, regardless of effect on inter-connector
capacitance, have no additional noise reducing benefit over that which is
achieved by using utp.
Finally, I'm observing that a belkin stp (floating screen) "high speed
adsl modem cable" drops my downstream rate from approx 7000 Kbits / sec
to approx 5300 Kbits / sec compared to a flat unscreened rj11 - rj11 lead.
Now, I might have a duff belkin "high speed adsl modem cable", as the
only tests apart from observing its actual performance that I can do are
basic continuity, but the observed performance of the cable combined with
my own understanding of the electrical engineering involved suggests to
me that the shield on the belkin "high speed adsl modem cable" is
essentially snake oil.
Rgds
Denis McMahon
The Natural Philosopher
If it is, its just a bit more from each to ground and a bit less between.
> I'm also speculating that an the ungrounded shield on the wall socket -
> modem connection will, regardless of effect on inter-connector
> capacitance, have no additional noise reducing benefit over that which is
> achieved by using utp.
>
Ithink it will actually.
> Finally, I'm observing that a belkin stp (floating screen) "high speed
> adsl modem cable" drops my downstream rate from approx 7000 Kbits / sec
> to approx 5300 Kbits / sec compared to a flat unscreened rj11 - rj11 lead.
>
Hmm.
> Now, I might have a duff belkin "high speed adsl modem cable", as the
> only tests apart from observing its actual performance that I can do are
> basic continuity, but the observed performance of the cable combined with
> my own understanding of the electrical engineering involved suggests to
> me that the shield on the belkin "high speed adsl modem cable" is
> essentially snake oil.
>
yes if its just a short length.
The majority of interference will come from the unshielded link to the
exchange.
> Rgds
>
> Denis McMahon
Charles Lindsey
>The grounding of the shield helps remove external interference, thats all.
I think even an ungrounded shield will remove most of the interference.
Essentially, the interference can still only get inside it at the ends.
Or, which comes to the same thing, only very low frquencies would get
through (wavelength comparable to the length of cable I should imagine).
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Web: http://www.cs.man.ac.uk/~chl
Email: c...@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5
Sam
> I'm kind of wondering, what's the point of the shielding on these? It's
> not as if it's grounded anywhere, and this bear of little brain wonders
> if such shielding would (a) have any effect, and (b) if the effect could
> even be detrimental.
Not sure of the technical terms but I recently replaced my faulty RJ11
cable and my download speed almost doubled.
Not sure of the difference between the two cables.
Denis McMahon
Well, I recently installed an adslnation splitter at the NTE, and have a
separate cat 5 run from the adslnation "unfiltered" punchdowns to near my
adsl modem / router.
The original phone wiring is attached to the filtered side punchdowns on
the adslnation faceplate.
I've tried the following:
2 different rj 11 modules
2 different rj11 - rj11 flat cables
an rj11-rj11 stp (floating screen)
an rj11-rj11 utp
I also tried using a BT secondary jack and cutting one rj11 connector off
of one of the rj11-rj11 flat cables and attaching a BT431A in its place.
So far, none of the results have approached the original rates I was
getting with microfilters, despite the fact that there's a long parallel
connection from the NTE to 3 sockets, and the computer etc is located
near the middle one of the three.
At the moment, I'm starting to think that the adslnation faceplate isn't
all that it's cracked up to be. I may look for a passive filter faceplate
with filtered and unfiltered punchdowns and try that instead.
Rgds
Denis McMahon
The Natural Philosopher
> Rgds
>
> Denis McMahon
Denis McMahon
>> I've tried the following:
>>
>> 2 different rj 11 modules
>> 2 different rj11 - rj11 flat cables
>> an rj11-rj11 stp (floating screen)
>> an rj11-rj11 utp
>>
>> I also tried using a BT secondary jack and cutting one rj11 connector
>> off of one of the rj11-rj11 flat cables and attaching a BT431A in its
>> place.
>>
>> So far, none of the results have approached the original rates I was
>> getting with microfilters, despite the fact that there's a long
>> parallel connection from the NTE to 3 sockets, and the computer etc is
>> located near the middle one of the three.
>>
>> At the moment, I'm starting to think that the adslnation faceplate
>> isn't all that it's cracked up to be. I may look for a passive filter
>> faceplate with filtered and unfiltered punchdowns and try that instead.
>>
> you are damed right. Its nothing special. But its neat.
I'm now considering the Pressac and Austin Taylor faceplates as stocked
by Solwise. Anyone here got opinions on these two?
Downstream attenuation is 42 dB on 2+, which I believe indicates 3km ish
of cable between me and the dslam (which looks about right from street
maps, exch is about 2.5 km by what I would consider the most logical
route for cables).
Rgds
Denis McMahon
Windmill
Not sure if anyone mentioned this already, but there used to be
security issues with unshielded UTP in the long gone days of 10Mb/sec.
Apparently it was possible to decode the signals from some distance away.
Whether that is still true with 100Mb/s and 1000Mb/s networks I don't
know.
--
Windmill, Use t m i l l
Til...@Nonetel.com @ O n e t e l
. c o m
Denis McMahon
Yes, any radiated signal can be received, and any received signal that
contains some sort of encoded data can be decoded if the encoding is
known.
google "tempest security"
Rgds
Denis McMahon