Twisted pair (TP) capacitance, shielded vs unshielded.

Does having a grounded shield around TP increase the inter lead capacitance? By like, 1/2 the capacitance of each wire to the shield.

Or put another way, if I ground one side (at one point) of the TP does the inter lead capacitance stay the same?

George H.

Reply to
George Herold
Loading thread data ...

It would certainly increase each wire's capacitance to ground, and if you were driving a balanced signal on it that would appear as an increased inter-lead capacitance.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

Thanks Tim, That's how it looks to me, but I was feeling uncertain.

Geroge H.

Reply to
George Herold

gnd | | C1 | | =================================================== | C2 | =================================================== | | C3 | | gnd

Without the shield, C1 and C3 are the capacitances from the wires to the grounded universe.

With the shield, C1 and C3 are the capacitances to the grouded shield.

When you add the shield, C1 and C3 go way up, and I'm guessing that C2 goes down a little bit.

And I don't have numbers!

Interestingly, most CAT5/6 cables have different twists on the four different pairs.

formatting link

formatting link

(I know more about CAT5/6 than I ever wanted to know.)

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

Semantic problem; one lead of TP tied to shield; capacitance of other lead to first cannot be isolated so semantically, the inter lead capacitance has no meaning. Obviously capacitance is larger (to both shield and lead as it is a combo).

Reply to
Robert Baer

The three capacitances are distinct and can be separately measured.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

** Not when one of the TP conductors is linked to the shield.

... Phil

Reply to
Phil Allison

The twist helps isolate the differential signal from the environment, but if all the pairs had the same twist you get a lot of crosstalk from transformer coupling. The four pairs should all have a different twist, with the common factors calculated so each 1 metre length has minimum cross-talk.

Clifford Heath.

Reply to
Clifford Heath

** Depends if the TP are driven with anti-phase signals or not.

When such is the case, signal current flow into the shield is very low as the two electric fields from the TP tend to cancel.

** As you have just created a co-axial cable, it is going to be higher.

Typical STP cables specs give twice the static capacitance per metre from cores to shield compared to core to core.

.... Phil

Reply to
Phil Allison

Tricky question. Do you actually mean "ground the shield vs. leave the shield open" or "ground the shield vs. have no shield at all"?

In the first case, grounding the shield (as opposed to letting it just sit there) will decrease the effective inter-lead capacitance considerably (think of the shield as the series connection between 2 capacitors to the leads). It won't reduce the load on the signal source (basically the inter-lead capacitance becomes converted to lead-to-ground capacitance, but the source will still have to drive it), but if the pair is driven asymmetrically, the effective impedance between the leads will change.

In the second case, the tricky part is which end of the shield you ground, how long it is, and what timescales you are interested in.

Grounding the shield (as opposed to having no shield) will always increase the inter-lead capacitance a little, due to the finite impedance of the shield and the ground connection, and it will also introduce a significant additional load on the signal source.

If your signal is perfectly balanced with respect to ground, there would be no difference except for the loading on the source, but any imbalance present will drive a net AC current into ground via the shield, so the ground impedance will become the significant factor since a part of the common mode signal will be then converted to differential mode through the finite impedances.

If you ground the near end (where the driving source of the signal is), then the effect will be relatively benign (as long as you have a properly set up transmission line with matching impedances and termination and the propagation speeds of common mode vs. differential mode mode signals in your cable are not too different).

If you ground the far end, then, at short timescales (and therefore high frequencies) the shield would "look" to the signal source as if it was not grounded at all but floating. In this case you'd have to add to the inter-lead capacitance half the shield-to-lead capacitance as well. As the timescales get lower the shield will gradually start to look more and more like being grounded at the near end (in the corner case at DC, there would be no more difference but then capacitance would have no meaning either any more).

Regards Dimitrij

Reply to
Dimitrij Klingbeil

That's the perfect picture... and yeah I see the shield as bringing infinity up close to the wires. But now I've got a (voltage) signal on the twisted pair, with some significant source impedance. Does the shield change the bandwidth? I guess if I have to I can do the experiment.

George H.

Reply to
George Herold

Yes signal on TP is voltage source with significant impedance. With a differential amp at the end of the TP to measrue the signal.

Hmm yeah... I see that too. Do you know of any references? I figure I can read about someone else doing the measurement, or do it myself.

OK I made a math mistake last night. Taking JL's picture, and assuming C1 and C3 are the same.. then I drew the inter lead capacitance as increasing by the series combo of C1 and C3 ~1/2C1. If I ground the bottom cable then C3 goes away.. but I now have C1 + C2... without the 1/2.

George H.

Reply to
George Herold

OK sorry for the confusion. There are two questions, the first was how I was trying to think about the problem. Does the interlead capacitance of TP increase when a grounded shield is added?

My real problem is that there is a (balanced) signal on TP, that is running inside a grounded tube. The signal is isolated (at DC) from ground, with a differential amp at the end. (Note to self: I better check the input capacitance of the diff amp.) Now if I ground one side of the TP (by connecting it to the shield.) I get a decrease in the electrostatic interference/ pickup. (Mostly 60 Hz and it's harmonics.) And I'm uncertain by how much this changes the capacitance/ bandwidth.

Ahh, Source impedance is 10 Meg., It's a photoconductor that's down the bottom of a 4 foot (1.3 meter) stainless steel low temperature probe. (I'm having fun with physics grad students every Friday evening*.) The twisted pair is taped to a 1.5" diameter SS tube with another ~2" SS tube around it. We measured a 3dB corner frequency of ~1k Hz last night. That was with the voltage source grounded at the bottom, (and the outer can/ tube not in place.)

Dimitrij, Thanks so much for your detailed reply. It's the loading of the signal source that is my biggest concern.

George H.

Reply to
George Herold

Oh I forgot the asterisk.

  • As the old physics joke goes; I'm telling my wife that I'm with my mistress. And my mistress that I'm with my wife. :^)

George H.

Reply to
George Herold

What's strange is that most U/FTP or S/STP cables, which have individually shielded pairs, also have different twists. That makes delay skew, which Ethernet doesn't care about but other apps might.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

Is the far end terminated? Is the cable long enough to have prop delay matter? If it's relatively short, you can model the whole thing as capacitance.

Probably. My educated guess (about 54% educated) is that the shield increases high frequency attenuation. Which won't matter for a short cable with a high impedance drive. The shield certainly increases capacitance, which is probably the "attenuation" that you care about.

There are circuits that help ignore the cable capacitance. Like charge-amp sorts of things, or bootstrapped shields.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

It helps with the magnetic coupling. The math is the same as coupled modes in fibres--if the beat length (2 pi /(k1 - k2)) is short, you can't couple much power from one line to the other before it starts coupling back out again.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

That degrades to the equivalent of a single coax, even easier to measure. If one wire is connected to the shield, we don't care about its capacitance.

You can always disconnect the short when characterizing the cable.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

That's what you will measure if you use a 2-lead c-meter, but that hides what's really going on inside the cable.

A 3-wire c-meter can measure the actual C2 value. It could be much less than C1 or C3. That's the beauty of a 3-wire meter: it can measure a fraction of a pF at the end of a couple of long coaxial measurement leads, and ignore the cable capacitances to ground.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

Save a lot of thinking and ground all shields on both ends.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.