extraordinary cable

On Feb 26, 11:03*am, Ian Jackson
wrote:
In message , Bill Wright
writes

I put a length of dielectric from this cable and one from a known good cable
in the microwave to see if they behaved differently. Neither of them became
warm at all.

Bill

Worth a try though. Interesting though.
--
Ian

I have experienced a very similar problem in the past with a screened
data cable made by a very reputable American company.

This turned out to be a problem with the foil screen. The overlapping
edges of the foil only made contact with each other at irregular
intervals, due to the way in which the insulation layer had been
deposited. This resulted in frequency selective attenuation, which
sounds exactly like the problem you have described.

In all other respects the cable seemed perfectly OK. It took some time
for the manufacturer to agree that there was a problem (we had to
provide frequency sweeps, despite the fact that the company states
that it 100% certifies all cable as part of the manufacturing process)
and identify the cause.

UKM

wrote in message
...
On Feb 26, 11:03 am, Ian Jackson
wrote:
In message , Bill Wright
writes

I put a length of dielectric from this cable and one from a known good
cable
in the microwave to see if they behaved differently. Neither of them
became
warm at all.

Bill

Worth a try though. Interesting though.
--
Ian

I have experienced a very similar problem in the past with a screened
data cable made by a very reputable American company.

This turned out to be a problem with the foil screen. The overlapping
edges of the foil only made contact with each other at irregular
intervals, due to the way in which the insulation layer had been
deposited. This resulted in frequency selective attenuation, which
sounds exactly like the problem you have described.


Like all cheap downlead cables with 'silver paper' foil, the foil is
actually plastic with something conductive stuck to one side. Hence it does
not contact itself where it overlaps.

Bill

Bill Wright wrote:
"Terry Casey" wrote in message
news
I think you'll find that there is another factor to be born in mind - the
dialectric constant. This is 1.0 for air but differs for other materials.

I seem to have lost the details of a couple of on-line calculators I've
looked at before but here is a useful one, assuming the maths behind it is
better than some of the spelling!

http://www.mogami.com/e/cad/electrical.html

Using 1.0mm for the centre conductor diameter and 4.6mm for the outer
diameter of the dialectric (actually the inside diameter of the outer
conductor) gives a 75 ohm impedance for foamed poly-ethylene dialectric.

Keeping the physical measurements the same and changing the dialectric
material gives some interesting results:

Dialectric Z(ohms)
Air 92
Foamed PE 75
PTFE 63
Solid PE 60
PVC 51


Very interesting. The diameter of the dielectric is 4.5mm. The diameter of
the dielectric of Cavel QF100 is 4.7mm. The foam of the oddball cable
appears to be slightly more dense. The diameter of the Cavel inner is
exactly 1mm, whereas that of the oddball cable is 1.05mm.

When I was looking at this recently, comparing RG6 with CT100 (WF100,
actually), I thought I'd got some of my figures muddled because I
expected, for the same style of construction, that the ratio for both
cables would be the same - but they're not!

At the time, I ran both sets of figures through the on-line calculator I
mentioned earlier. I've now added more data and repeated the exercise.
Data id taken from the manufacturers' data sheets except for the mystery
cable, of course. Here's the result:

Type inner outer D/d Zo
RG6 1.02 4.57 4.48 73
WF100 1.00 4.8 4.8 77
QF100 1.00 4.75 4.75 76
??? 1.05 4.5 4.29 71

Manufacturers typically quote impedance as 75? ±3 so the results are
reasonable - the mystery cable is only just outside the range and, of
course, slight variations in dialectric composition could narrow the field.


The response is remarkably flat. It just looks like you've added an
attenuator! Most of the crap cables become really very crap indeed above
1GHz, but this one behaves consistently right to the top of the satellite IF
band. It's just as if you're measuring the loss on a cable times as long.


So ... no further progress, I'm afraid ...


I'm baffled.

Bill


.... so am I ...


Terry

"Jim Lesurf" wrote in message
...
In article , Ian Jackson
Is the outer braid also 'steel'? I can't recall what was said about that
earlier...

No, it appears to be copper. The braid cover is only about 10%. The foil is
the usual plastic with a conductive substance on one side.

Bill

On Feb 26, 2:33*pm, "Bill Wright"
wrote:
"Jim Lesurf" wrote in message

...

In article , Ian Jackson
Is the outer braid also 'steel'? I can't recall what was said about that
earlier...

No, it appears to be copper. The braid cover is only about 10%. The foil is
the usual plastic with a conductive substance on one side.

Bill

My money is still on incorrect manufacturing of the foil screen.

The lack of a conducting overlap suggests that there is capacitive
coupling beteen the edges of the foil. This will result in the foil
not acting as a screen but as a long inductor with parallel
capacitance which will be resonant somewhere in the GHz region, seems
like a recipe for trouble to me.

Dielectric loss or resistive loss due to a steel centre core would
produce a log attenuation curve across the whole frequency range -
this is frequency selective.

UKM

wrote in message
...
On Feb 26, 2:33 pm, "Bill Wright"
wrote:
"Jim Lesurf" wrote in message

...

In article , Ian Jackson
Is the outer braid also 'steel'? I can't recall what was said about
that
earlier...

No, it appears to be copper. The braid cover is only about 10%. The foil
is
the usual plastic with a conductive substance on one side.

Bill

My money is still on incorrect manufacturing of the foil screen.

The lack of a conducting overlap suggests that there is capacitive
coupling beteen the edges of the foil. This will result in the foil
not acting as a screen but as a long inductor with parallel
capacitance which will be resonant somewhere in the GHz region, seems
like a recipe for trouble to me.

Dielectric loss or resistive loss due to a steel centre core would
produce a log attenuation curve across the whole frequency range -
this is frequency selective.

UKM

This type of foil is used on all the RG6-type cables. They all have the
problem that the overlap doesn't conduct. Remember though that the foil is
in contact with the braid.

The losses seems to be in line with frequency all the way up; it's just that
they are about ten times worse than they should be.

Bill

Hil says, "You seem to be fascinated by that bit of wire. Why do you keep
picking it up and peering at it like that?"

Bill

Bill Wright wrote:


wrote in message
...

My money is still on incorrect manufacturing of the foil screen.

The lack of a conducting overlap suggests that there is capacitive
coupling beteen the edges of the foil. This will result in the foil
not acting as a screen but as a long inductor with parallel
capacitance which will be resonant somewhere in the GHz region, seems
like a recipe for trouble to me.

Dielectric loss or resistive loss due to a steel centre core would
produce a log attenuation curve across the whole frequency range -
this is frequency selective.

UKM

This type of foil is used on all the RG6-type cables. They all have the
problem that the overlap doesn't conduct. Remember though that the foil is
in contact with the braid.

I'm clutching at straws here with no technical expertise whatsoever...

Is the foil on the right way round? Is the conductive side in contact
with the braid?
--
Phil Cook looking north over the park to the "Westminster Gasworks"

Funny stuff, cable.

You never know when it's going to cause you grief ...

When DTV was being rolled out on what was then the ntl cable network, a
problem occurred which caused more than a little head scratching.

The same analogue/digital mix is fed to every subscriber - except for
the MCNS (Multimedia Cable Network System) signal which comes from a
broadband router. Today, Broadband Internet signal ares carried by the
MCNS although, originally, much more modest use was made of it for
interactive tv, etc.

The point is that every block of subscribers has its own MCNS and, as
more use is made of it and traffic increases, the blocks are becoming
increasingly smaller.

Provision had to be made for combining the broadcast signal with the
individual MCNS signal, immediately before it was fed to the laser
transmitter. This was done in one of two ways, dependant on the original
build strategy. This description is of the method used at almost all the
sites I was concerned with. A directional coupler was inserted into the
broadcast feed and the RF drive levels re-adjusted to make up for the
insertion loss. The MCNS was then added via the DC and the level
adjusted to match the DTV carriers. This was the last stage before
acceptance testing at the roadside cabinets began and was pretty
straightforward. In fact, it had been done many, many times without a
problem - until we got to Bournemouth.

The BER and constellation in one area was rubbish. This was not unusual
- cleaning fibres usually cleared up the problem. If not, it was
normally an optical transmitter that rebelled at the new demand made of
it - occasionally a receiver.

It wasn't a dirty fibre. In this case, two receivers were fed by the
same laser, so that seemed to rule that out. That night, the laser was
replaced. Nothing changed! Everything was checked and double checked -
to no avail. Fibres were quickly swapped between two working
transmitters - clean as a whistle! Two faulty transmitters? Put things
back as they were and swap the RF feeds - clean as a whistle!

Replace the directional coupler - no change. The splitter on the
incoming RF feed was replaced - no change. Replace the connecting cables
- no change. Pull hair out - no change! The problem was proven to be at
RF and everything had been eliminated - everything was reasonably
compact so there were no long cable runs, so no hidden damage (or so we
thought!) but, try as we might, the cause eluded us.

Bournemouth is a long way from home. We'd done everything else and the
test results were perfect - except for this one problem. We left, went
home, and the next week we started somewhere else. We thought, we
talked, we came up with the most bizarre theories but couldn't think of
anything constructive to resolve the problem we'd left behind ...

.... a few days later, my mobile rang: "Found it! And you'd never guess
in a thousand years ...!"

Somebody had decided that, as nothing that anybody else had tried had
worked, he would systematically go through the entire Bournemouth hub
site and check every single F connector - no matter what it was being
used for! He unscrewed every plug and visually checked it and its mating
socket, then put them back again.

He was refitting one plug when he stopped - "It's the wrong colour", he
thought. "No, that's it! It ISN'T coloured!"

The cable we were using for the upgrades was Commscope RG59 Headend
Cable. It has very high levels of screening and a solid poly-ethylene
dialectric. The inner screen is bonded aluminium foil with a blue
adhesive on the inside. When you look at the cut cable or a properly
terminated F connector, the clear dialectric takes on a bluish tint.

In this case, though, it was clear - no trace of colour! As it was a
crimped connector, there was no way of dismantling it but we assume
that, when the wireman was making up the cables, when this one was
stripped, the foil had somehow been stripped of as well. One new F plug
and everything worked fine! The strange thing is that this cable and the
directional coupler it was connected to was nothing to do with the
affected laser at all.

Shortly after that, the bean counters changed the cable type for a
cheaper one. It was still Commscope RG59 Headend Cable - but this time
with opaque foamed PE dialectric ...

Strange old world isn't it?

Terry

In article , Bill Wright
scribeth thus

"Jim Lesurf" wrote in message
...
In article , Ian Jackson
Is the outer braid also 'steel'? I can't recall what was said about that
earlier...

No, it appears to be copper. The braid cover is only about 10%. The foil is
the usual plastic with a conductive substance on one side.

Bill



Be interesting to use another braid covering to see what difference
that'll make if you've got the time/inclination etc;!..
--
Tony Sayer