How a Permissive Chain Is Supposed to Work

Why this matters

A string of contacts in series ahead of a coil is the most common control arrangement in the trade and the most commonly mis-measured. The method everyone knows is to work down the chain until the voltage disappears. The part that gets skipped is why the voltage behaves that way, and without it a tech will take three readings that are each individually correct, assemble them into a wrong conclusion, and condemn a good contact.

The chain also hides things. It reports exactly one open element no matter how many are open, which means "found it, fixed it" is a sentence that sends people back to site.

Before the meter goes in

The control circuit may be 24 V. The enclosure it lives in is usually not, and that is where the hazard is: line-voltage terminals, contactor line sides and incoming lugs sit inches from the terminal strip you are probing, and a slipped probe or a dropped lead reaches them.

Energized measurement is permitted under 29 CFR 1910.333(a)(1) where de-energizing would introduce additional or increased hazards or is infeasible given the equipment design, and tracing a live permissive chain is a standard case of the second, since a de-energized chain has no state to read. Use a test instrument whose measurement category rating suits the highest voltage present in that enclosure, not the 24 V you intend to read, inspect the leads first, and use the shortest exposed probe tips available. When the work moves from reading to landing wires, that is 29 CFR 1910.333(b)(2) for de-energizing, with live-dead-live proving per NFPA 70E-2021, 120.5.

What a chain is, and why its order is not arbitrary

A permissive chain is a set of conditions that must all be true, implemented as contacts in series between a source and a load. Logically it is an AND. Physically it is a single path, and that single path is what makes it diagnosable.

The order is a design decision worth reading. Protective and safety-related elements are conventionally placed nearest the source, so that when one opens it removes potential from the entire remainder of the string rather than from one branch of it. Sequence permissives sit after them, roughly in the order the sequence needs them. The last element before the coil is very often the controller's own output, which is the one that changes state constantly in normal service.

So a device's position tells you how the designer ranked it, and a device you expected near the source that you find near the coil is worth a second look before you trust the drawing.

There is a second split that the single-chain picture hides, and it changes what a symptom means. Some machines have one chain that gates starting only, with a separate and shorter chain that can drop a machine already running. A condition wired into the start chain alone will prevent a start and will not stop a running machine, which is exactly right for a prerequisite like proof of an open damper and exactly wrong for a protection. The field symptom sorts them for you before you open anything: a machine that will not start at all is in the start chain, and a machine that starts and then drops out a few seconds later is being taken down by something in the run chain, most often a proof that had time to make and then fell back. Ask which behaviour the customer actually saw, because "it will not run" gets used for both.

The electrical fact the method rests on

In a series circuit feeding a single load, essentially all the source potential appears across the open element. Closed contacts drop close to nothing, because they carry the same current with near-zero resistance.

Clip the common lead to the circuit's own common and walk the hot lead node by node from the source toward the coil. You read source voltage at every node up to and including the input side of the open contact, and near zero at every node after it. There is exactly one transition from source voltage to zero, and the open element is the one straddling it.

That statement holds under specific conditions, and they are the conditions under which it fails in the field: a single series path with no parallel feed, a low-impedance load actually connected, and readings referenced to the same common the load returns to. Change any of those and the readings stop meaning what they appear to mean.

The worksheet, filled in

A packaged unit, contactor coil not pulling in on a call. Six elements in the chain. Meter common clipped to the transformer common leg, coil connected, readings taken in one pass.

One transition, between N3 and N4. Element 4, the airflow proving switch, is open. Measuring directly across it, N3 to N4, reads 24.2 V, which is the same fact stated the other way and a useful confirmation that you are not chasing a bad common lead.

Then the discipline that separates a diagnosis from a parts swap: 24.2 V across an open contact establishes that the contact is open. It establishes nothing about whether it should be. A proving switch that opens because there is no airflow is a switch doing its job, and replacing it lands in the same place as jumpering it, one invoice later. Measure the quantity it senses at the location it senses: differential across the proving taps read 0.15 in w.c. against a switch scale set at 0.40 in w.c. The switch was right. The airflow was not, and the sibling article on establishing why a safety device opened is the procedure for the rest of that thread.

The second open the first one was hiding

Airflow restored, the chain was re-walked before anyone pressed start.

The transition has moved. Element 6, the controller output, is open now, and it was open the whole time. It could not be seen on the first pass because everything downstream of the airflow switch was at zero, and a second open contact sitting in a dead section of the string looks exactly like a closed one.

This is a structural property of the method rather than a mistake anybody made: a series chain reports the first open element and no others, forever, one at a time. Re-walk the chain after every correction, before the start attempt, not after the customer calls back.

Three readings that are not what they look like

A partial reading with the load connected. Source 24.3 V at R, 24.1 V at the node ahead of a suspect contact, 15.2 V at the node after it. That is a drop of 24.1 - 15.2 = about 8.9 V across a contact that should drop near zero, and the coil is being offered roughly 15 V. This is a real finding and a good one: a contact with resistance divides the voltage only while current is flowing, which is exactly why the reading is taken with the coil connected. Pitted contacts, a corroded terminal, a strand-damaged conductor. The equipment will often work when cold and fail when the connection warms.

A partial reading with the load disconnected. Same-looking number, completely different cause. On a high-impedance digital meter, an unloaded conductor running alongside energized wiring picks up capacitively coupled voltage. Readings of a few volts, sometimes more, appear at nodes past the open contact and look like a failing contact. The check is to repeat with the meter in a low-impedance mode, or with a solenoid-type tester, or with the load reconnected. A coupled reading collapses to near zero; a real drop does not.

Source voltage everywhere and no pull-in. The chain reads 24.3 V all the way to the coil terminal and the contactor sits still. The chain is not the problem. Either the coil is open or the return path to common is broken, and the way to tell them apart is to read across the coil rather than from the coil to common: full source across the coil with no pull-in means an open coil, while near zero across it means the potential is not getting through the return leg.

Where this method stops being reliable

Parallel paths. A hand-off-auto selector, a bypass contact, a second feed, an auxiliary contact wired around part of the string, or a jumper somebody left in all break the single-path assumption. The symptom is a chain that reads sensibly and a machine that behaves impossibly, and the reading that gives it away is source voltage at a node downstream of a confirmed open contact. Do not explain that away as a coupled reading before you have looked for a second feed.

A shared common that is not shared. If the meter's common lead is clipped to a chassis ground or a different transformer's common rather than to the common the coil returns to, every reading is referenced to the wrong thing. Confirm the reference by reading source to your chosen common first and getting nominal voltage before you walk anything.

Software permissives. Where several of the conditions live in a controller rather than as physical contacts, there is only one terminal to measure, at the controller output, and the chain inside it is not accessible with a meter. Read those from the controller's own status display, and treat the controller output as a single element in your physical chain rather than pretending you can probe it.

References

• 29 CFR 1910.333(a)(1) for the conditions permitting energized troubleshooting, and 1910.333(b)(2) for de-energizing electrical circuits and equipment
• NFPA 70E-2021, 120.5 for establishing and verifying an electrically safe work condition
• NFPA 79 for electrical standards for industrial machinery, including control circuit practice
• See related: Interlocks and Why They Are Not Control; How to Establish Why a Safety Device Opened; How Control Voltage Differs From Line Voltage