Home › Guides › Loop Checks

Loop Checks Explained: I/O Verification in Commissioning

Every control system is only as good as its inputs and outputs. A distributed control system or PLC can run the most elegant control logic ever written, but if the pressure transmitter it reads is wired to the wrong input, or the valve it commands is wired backwards, the plant will not behave — and the first sign of trouble may be a process upset rather than an alarm. Loop checks exist to prevent that. A loop check proves, one loop at a time, that the complete signal path from a field device to the control system and back again is correct: the right device, the right wires, the right input or output channel, the right engineering units, the right alarms, and the right final element moving in the right direction.

This guide explains loop checks in plain language: what a loop is, the difference between cold and hot checks, how analogue and digital loops are verified, what a loop folder contains, and how punch lists turn failures into a managed path to handover. As with everything on this site, always work to your project specification, the manufacturer's instructions and your site safety rules; the general guidance here never overrides those documents. Loop checking is done by competent persons under permit, with outputs to operating equipment made safe or isolated where required, so that a test signal cannot accidentally start a motor or open a valve on a live process. If you are new to instrumentation as a whole, start with our transmitters guide and our control valves guide, and for the wider project picture see our substation commissioning guide.

What a loop is, and what a loop check proves

In commissioning language, a loop is the complete chain for one measurement or one control function. For an analogue input loop, the chain runs from the field transmitter, through its cable and marshalling cabinet, through any barriers or isolators, to the I/O card in the control system, and into the database point that the operator sees on screen. For an analogue output loop, the chain runs the other way: from the control system's output, through the I/O card and marshalling, to the final element — typically a control valve positioner or a variable speed drive — and the valve or drive's actual position is fed back and displayed. A digital loop is simpler but no less important: a switch or pushbutton input, or an on-off output to a solenoid or motor starter, proved from field contact to screen indication and back.

A loop check proves that chain end to end, and that phrase matters. Checking the transmitter in isolation proves the transmitter. Checking the I/O card in isolation proves the card. Neither proves that the transmitter is connected to that card, through those wires, with that scaling, displaying on that screen graphic, and alarming at those limits. Most commissioning errors live in the connections between pieces of equipment — a wire landed on the wrong terminal, two loops swapped in a marshalling cabinet, a range entered wrongly in the database — and only an end-to-end check finds them. That is why loop checking is treated as a distinct commissioning activity with its own documentation, rather than as a by-product of equipment testing.

Cold loop checks: wiring proved before power

Loop checking usually begins cold, with the system unpowered. A cold loop check verifies the physical installation: cable continuity from field device to marshalling cabinet to I/O card, correct terminal numbers at each end, correct cable identification, insulation resistance of the cables, and correct earthing and shield termination. It is done with a multimeter and with the loop drawings in hand, point by point, and it catches the errors that are cheapest to fix — before panels are energised and before a wrong connection can damage an input card or drive a valve the wrong way.

Cold checks also verify the field device installation itself: the transmitter is the correct model and range for the loop, it is mounted in the correct location and orientation, its impulse lines or sensors are connected correctly, and its nameplate matches the loop number on the drawing. A transmitter with the wrong range will pass a continuity check perfectly and still report the wrong process value forever, which is why cold checks include device verification, not just wire verification. The cold check is signed off loop by loop, and only loops that have passed cold checks proceed to hot checks — a discipline that keeps the hot-check programme from drowning in wiring faults.

Hot loop checks: signals proved with power on

A hot loop check is done with the system powered, and it proves that signals flow correctly through the whole chain. For an analogue input, the field transmitter's output is simulated or driven across its range — typically at zero, twenty-five, fifty, seventy-five and one hundred percent — using a signal calibrator connected at the transmitter, and the control system display is checked at each point for the correct engineering value, the correct units, and the correct response. The five-point check is not arbitrary: zero and span prove the range, the middle points prove linearity, and the return points prove there is no significant hysteresis. Our transmitter calibration guide covers the instrument side of this in detail.

Alarms and trips belonging to the loop are proved during the hot check as well. Each alarm setpoint is driven through, and the alarm is confirmed on the operator screen, on the alarm list, and on any annunciator or horn where applicable. Where a loop has a trip or interlock function, the trip is proved to operate the correct output — with the final element made safe or isolated, so that the test does not trip a live process — and the reset behaviour is confirmed. A loop whose alarms were never driven through during commissioning is a loop whose first real alarm may be the first time anyone discovers the alarm was configured wrongly.

For an analogue output loop, the direction reverses: the control system output is driven from the operator station across its range, and the final element's response is observed in the field — a valve strokes from closed to open, a drive runs from minimum to maximum speed — with position feedback confirmed on screen. Fail-safe behaviour is proved explicitly: on loss of signal or loss of air, the valve must move to its designed fail position, and that movement must be observed, not assumed from a datasheet. For digital loops, each input is operated in the field and confirmed on screen, and each output is commanded from the screen and confirmed in the field, including the correct indication of the operated state. A motor starter that shows "running" when it is stopped, or "stopped" when it is running, is a loop failure, not a cosmetic issue.

Loop folders and punch lists

A substation or plant may have hundreds or thousands of loops, and loop checking at that scale is a document-control exercise as much as a technical one. Each loop has a loop folder: the loop drawing, the instrument datasheets, the cable schedule entries, the I/O list entry, the alarm and trip schedule, the calibration certificates, and the loop check sheet itself, with space for the measured values at each test point, the tester, the witness and the sign-off. The folder is the loop's evidence file, and it is what the client reviews at handover. A loop without a complete folder is not a commissioned loop, however well it appears to work.

Failures are managed through a punch list. Every defect found during loop checking — a swapped wire, a wrong range, a failed alarm, a valve stroking backwards — is recorded as a punch item with a description, a category, an owner and a due date. Categories matter: a category A punch blocks handover or energisation of the affected system, while lower categories may be accepted for completion after handover with an agreed plan. The critical discipline is re-checking: when a punch item is fixed, the affected loop is re-checked end to end, not just at the point that was fixed, because a repair in a marshalling cabinet can disturb a neighbouring loop, and a second fault in the same loop is common. Loops are only signed off when their folder is complete and their punch list is clear, and that sign-off — by the commissioning engineer and the client's witness — is what authorises the loop to be put into service.

Common loop-check failures and how to prevent them

The failures found during loop checking fall into a small number of repeating patterns. Swapped loops in marshalling cabinets are the classic: two cables landed on each other's terminals, so that two measurements read each other's values. Wrong scaling in the control system database is nearly as common: a transmitter ranged 0 to 10 bar entered as 0 to 16 bar, so that every reading is wrong by a consistent factor that looks plausible on screen. Reversed valve action — a valve that opens when the controller says close — is usually a wiring or configuration error at the positioner, and it is exactly the kind of fault that causes a process upset if it reaches operation. Shield and earthing errors cause noisy, drifting readings that are hard to diagnose after handover and easy to prevent during installation.

Prevention is mostly about sequence and discipline. Check loops in the order the drawings say, not in the order that is convenient. Never sign off a loop from the control room alone — someone must be in the field, watching the device, for every check. Never accept "it worked yesterday" as a substitute for a recorded check. And never let schedule pressure convert a failed loop into a signed-off loop with a verbal promise to fix it later: the punch list exists precisely so that known defects are visible, owned and tracked, rather than hidden in someone's memory. The projects that energise smoothly are rarely the ones with the fewest loop faults; they are the ones whose loop faults were found, recorded and fixed before energisation, one loop at a time.

Frequently asked questions

Related guides: power transformer testing and commissioning guide, winding resistance test, protection relay testing, substation commissioning guide.