Crosstalk in Multi-Pair Cables: When Overall Screening Is Not Enough

TL;DR

An overall screen protects the pairs inside a cable from the world outside. An overall screen does not provide the same protection between pairs as individual pair screens. When a switching signal shares a cable with analog loops, the coupling happens inside the screen, where no external shielding, earthing or re-calibration can reach it. That is what individually screened pairs are for.

A level reading blips, cleanly and repeatably, every time a solenoid valve fires. Two flow transmitters on opposite sides of the plant drift loosely in step. A temperature indication takes a small step whenever a motor starts, then settles as though nothing happened.

The technician recalibrates the transmitter. The DCS engineer swaps the input card. The loop is meggered and passes perfectly.

The fault stays exactly where it was.

Every instrumentation engineer meets this fault eventually, and it never introduces itself honestly. Cross talk is one possible cause, especially when the affected signals share a cable.

The Fault That Passes Every Test

Instrumentation cable crosstalk survives commissioning because the three standard checks are all measuring something else. Each one returns a clean result on a cable that is actively corrupting readings.

The megger says healthy: Insulation between pairs is perfect, and it should be. Crosstalk is electromagnetic coupling across healthy insulation. Hence, an IR test cannot see it.
The loop check says healthy: Calibrate a static loop and it reads accurately. The fault only exists while the aggressor signal is switching. A static loop check may miss interference that occurs when a solenoid switches.
The drawings say healthy: Pair allocation lists record tag numbers. They rarely record signal type per pair, so the drawing cannot tell you that a discrete output is sitting two pairs away from a thermocouple.

What does catch it is correlation.

Random noise looks random. Crosstalk looks like one signal wearing another signal’s clothes: the disturbance on loop A is time-locked to events on loop B.

Line the trend up against the event log. If the blips land on the solenoid commands every time, the cable is the suspect and no further calibration will help.

Why an Overall Screen Cannot Stop Crosstalk

A multi-pair instrumentation cable contains several twisted pairs. Depending on its construction, it may have one overall screen, individual pair screens, or both. An aluminium-mylar screen wraps the whole bundle, a drain wire terminates it, armour and sheath sit outside.

That wall does one job well: it keeps external interference out. Fields from nearby power cables, from drives, from radio sources.

Every pair inside that wall still lives in the same room. Pairs lying beside each other for hundreds of metres couple two ways:

Capacitively. Adjacent conductors form small capacitors. A fast voltage change on one pair injects current into its neighbour.
Inductively. Current changes on one pair induce a voltage in the loop area of the next.

Twisting is the defence, and it is the reason instrumentation pairs are twisted at all. Each twist reverses the orientation of the loop, so induced pickup cancels itself twist by twist. For a great many installations, a twisted pair cable inside an overall screen is genuinely enough.

It stays enough until somebody puts the wrong two signals in the same cable.

The Spare Pair Nobody Documented

It is almost never a design decision. It is a construction decision made under time pressure, and it looks reasonable at the moment it is made.

The 12-pair cable to a junction box has three spare pairs. Late in the build, someone needs one more signal back from the field:

A solenoid command
A pulse output from a flow meter
A 24 V discrete status

Pulling a new cable means trays, transits, glands, documentation, and a week nobody has. The spare pairs are right there.

So the switching signal goes into a spare pair.

From that day, every time it fires, its sharp edges couple into the analog loops sharing the twisted pair cable beside it. This can create a problem that routine commissioning check may miss and that becomes harder to trace during operation.

If You Have This Fault Today

Separating the signals into different cables can help, but first confirm where the interference is coming from.

Move the aggressor: Moving the switching signals into a separate, suitably routed cable can reduce interference with nearby measurement signals.

Check the screen terminations first: It costs an hour. A screen earthed at both ends, or an individual screen chopped at a junction box, may be compounding the coupling, and the fix is free compared to pulling cable.

Re-allocating pairs is not a fix: Where no spare cable exists, moving the aggressor to the pair furthest from the most vulnerable loop reduces coupling. Pair positions inside a laid-up bundle are not always what the numbering suggests, so treat any improvement as a reprieve while the proper cable is arranged.

Filtering at the input card is a mask. Damping the signal in the DCS hides the symptom and slows the loop response you were paying for. It is a defensible holding measure before a shutdown, and it should be recorded as one.

Whichever route you take, write the signal type against the pair in the allocation list before you close the job. The next person to find a spare pair will make the same decision the last one did unless the drawing stops them.

What Can Share a Cable, and What Cannot

The discipline that prevents all of the above, in descending order of importance.

Never mix signal categories in one cable

Analog measurement signals (4-20 mA, RTD, thermocouple) in one cable. Discrete and switching signals in another. Pulse and communication signals in their own. The cheapest crosstalk fix in history is an extra cable pulled on day one.

Respect the vulnerability ladder

The lower the signal energy, the greater the harm. Thermocouples and RTDs work differently, but both can be affected by electrical interference, and coupling that a robust 4-20 mA loop shrugs off will walk straight into a temperature reading. Low-level signals earn individual screens even inside a same-category cable.

Specify IS + OS wherever the first two rules are at risk

An individually screened pair is a walled room inside the walled compound: each pair wrapped in its own aluminium-mylar screen with its own drain wire, then the overall screen outside. Individual pair screens help reduce interference between pairs. On a spec sheet it reads “12P x 0.5 sq mm, IS+OS”, and those five characters are the whole difference.

Remember that distance multiplies everything

Coupling accumulates with every metre that pairs lie together. A 20-metre panel tail forgives. A 400-metre trunk to a far-flung junction box does not.

Terminate the screens properly, or they are decoration

Individual screens must run continuous through junction boxes, screen to screen through the terminals rather than chopped and forgotten. Single end earthing is common, but the connection should follow the equipment instructions and project requirements. A screen earthed at both ends invites circulating current. A screen earthed nowhere is a floating antenna.

Do Your Signals Need Individual Screens?

The choice between overall-screen only (OS) and individual plus overall (IS+OS) is not about cable quality. Both constructions are legitimate.

The question is what the pairs inside will be asked to carry.

Property	Overall screen only (OS)	Individual + overall (IS+OS)
Blocks external interference	Yes	Yes
Blocks pair-to-pair coupling	No	Yes
Mixed signal categories	Not acceptable	Tolerable, though segregation is still better
Millivolt signals (RTD, thermocouple)	Risky alongside other pairs	Recommended
Long parallel runs (300 m and beyond)	Coupling accumulates	Handles it
Spare-pair reuse for discrete signals	Creates the classic fault	Contained
Relative cost	Base	Premium
Spec sheet reads	12P x 0.5 sq mm, OS	12P x 0.5 sq mm, IS+OS

When Overall-Only Screening Is the Right Call

Honesty cuts both ways, and over-specification wastes project money that could have bought a second cable.

An overall-screen construction is a legitimate and economical choice when all three of these hold:

Every pair carries the same category of robust signal, typically 4-20 mA loops at similar levels
Discrete, pulse and communication signals are routed in their own screened cables
Run lengths are ordinary rather than long parallel trunks

That is a clean, disciplined design and there is nothing to apologise for in it.

The failure mode was never the overall-screen construction. It is mixing signals inside it.

Three Things the Label Cannot Show You

Two shielded twisted pair cables carrying the same printed specification can behave very differently in service. Noise immunity lives in construction details that no label records.

Screen coverage. A cheaply made foil screen with gaps or narrow overlap leaks. Coverage should be complete, with the drain wire in continuous contact along the run.
Uniform twisting. Consistent lay length is what makes twist cancellation work. Sloppy, variable twisting quietly gives away the cable’s main defence, and nothing on the label will tell you.
Drain wire continuity. End to end, without breaks, so the screen can actually be terminated as designed.

These are manufacturing-discipline properties. They are invisible at the counter and decisive in service. Ask your vendor how each one is controlled and tested in production, and expect a specific answer rather than a reassurance. Serious shielded cable manufacturers will have one ready, and any instrumentation cables manufacturer worth specifying can show you the in-process controls rather than describe them.

From BMI’s plant: screen application and pair lay are controlled in-process at Bhiwadi, and every batch is tested in-house on instruments calibrated by NABL-accredited agencies before dispatch, with batch test documentation available for project QA records.

Where BMI Fits

BMI Cables manufactures multi-pair instrumentation cables at Bhiwadi, Rajasthan:

Twisted pairs in overall-screen and individual-plus-overall-screen constructions
Armoured and unarmoured, with FRLS and zero-halogen sheathing options
Alongside the power, control and building-wire ranges from the same plant

A BIS-licensed, ISO 9001:2015 certified instrumentation cables manufacturer in India, BMI is an approved vendor to leading PSUs and an approved make on major metro rail projects. See the complete record on our Certifications and Approvals page.

Specifying instrumentation cable for a project?

Send us your instrument index or cable schedule. Our technical team will review the pair allocations, flag where individual screening is and is not warranted, and respond with a recommendation and quotation within 24 hours. Email sales1@bmicables.com or WhatsApp +91 92570 35793.

Frequently Asked Questions

1. What is crosstalk in an instrumentation cable?

Unwanted coupling of a signal from one pair into another pair of the same cable, through the capacitance and mutual inductance between adjacent conductors. It shows up as disturbances on one loop that are time-locked to activity on another.

2. Does the overall screen prevent crosstalk?

No. The overall screen protects the pairs from external interference. Crosstalk happens between pairs inside the screen, where the overall screen has no effect at all. Only pair segregation and individual pair screens act on it.

3. Can I use spare pairs in an analog cable for a solenoid or discrete signal?

Do not. Switching signals are the classic crosstalk aggressor, and putting one inside an analog cable creates a fault that passes megger and loop checks but corrupts readings in service. Route discrete and switching signals in their own cable.

4. What does "IS + OS" mean on a cable specification?

Individually screened plus overall screened. Each pair carries its own screen and drain wire, and the whole assembly has a common screen as well. The construction offers extra protection for sensitive signals. Whether different signals can share it depends on the equipment and project requirements.

5. Why does crosstalk not show up in a megger test?

Because insulation resistance measures whether conductors are isolated, and they are. Crosstalk is electromagnetic coupling across healthy insulation, not leakage through failed insulation. It only appears while the aggressor signal is active.

6. Are screened cables and shielded cables the same thing?

Yes. Screened is the common term in Indian and British practice, shielded in American practice, and both describe the same conductive layer around a conductor, pair or bundle. A specification asking for screened cables and one asking for shielded cables are asking for the same construction, so read the screen type (foil, braid, individual, overall) rather than the word.

7. When is an overall-screen-only cable acceptable?

When every pair carries the same category of robust signal, typically 4-20 mA loops, with switching, pulse and communication signals routed separately, over ordinary run lengths. Same category, separate cables per category: that discipline is what makes the economical construction safe.