Short answer: a purge low-pressure alarm that survives a known-good bypass connector and a round of board swaps is almost never an electronics fault. It lives in the air path or the sense path, so work it in this order: the supply regulator and filter, then the leak paths — one missing blanking plug or a cut gasket — then the blocked or split sense tube. The margin you are chasing is small. NFPA 496 and IEC 60079-2 set the minimum enclosure overpressure at about 0.1 in w.c. above ambient, and one open 13 mm knockout vents it instantly.
Why does the alarm show only when you request motion?
Because the safety chain evaluates the purge-OK input at the moment servo power is requested — the teach-mode enabling switch or the cycle-start command — so a pressure drifting just under the minimum trips exactly there and reads normal at idle.
Purge and pressurization holds the enclosure slightly above the atmosphere around it, so flammable gas outside cannot flow in toward ignition-capable parts. NFPA 496 gives the method for US plants and IEC 60079-2 the international version. Type X covers a Division 1 area and interlocks power off until pressure is proven; type Y drops Division 1 to Division 2 inside the box; type Z drops Division 2 to unclassified. For the type Y and Z levels on most motion-controller cabinets, the minimum operating overpressure is about 0.1 in w.c. [25 Pa] above ambient — the value NFPA 496 and the pzc level of IEC 60079-2 work to.
Once the timed purge has completed, the rule is blunt: lose the pressure, lose the power. That contact sits in the servo-enable chain, which is why the fault bites the instant you press the enabling switch or start the cycle. Purge is also not the only protection out there — a flameproof enclosure contains an internal explosion rather than excluding gas, and when its pressure test has to be repeated is a separate question.
Pull the full error history before anything else. A companion code in the same event — a purge-timeout or a sensor code — names the stage that failed, and the safety monitor screen shows which contact is open at that instant.
What does a fault that survives board swaps and a bypass connector tell you?
It tells you the failing item is common to everything you replaced: the pressure itself, the sensing element, the tube or cable between them, or the one input the bypass does not cover.
A bypass connector stands in for the purge unit's output contact, or for its sensor input — the manual says which — and it does not stand in for every input in the chain. Prove it with four checks, power off for the first two:
- Fully seated and locked on the socket the manual names.
- Continuity across the jumper pins inside the connector.
- Pin condition at the socket: bent, recessed or corroded contacts.
- The monitored input must change state the moment you fit it.
If the monitored input holds its state with the bypass fitted, the bypass is not reaching it. Ring out every conductor between the sensor, the purge unit and the controller input, both ends off, each wire to its neighbours and to ground. A wire shorted to a neighbour passes an end-to-end check and still fails in service. Work with the area cleared of gas and the supply locked out, and keep the bypass for service only — it comes out before the machine runs again.
If the input does change state and the alarm stays, the pressure or its sensing is genuinely at fault. Move to the air path.
What does each symptom point to?
Read the symptom before you open the cabinet: what the pressure gauge does at the instant the alarm fires points at one of six faults, from a half-closed supply valve to a sense tube full of condensate.
| Symptom when it alarms | What it points to | First check |
|---|---|---|
| Only at servo-on or cycle start | Margin drifting under the minimum: small leak or supply sag | Read the port gauge at the request |
| During the purge timer, with a companion code | Sequence or sensor stage, not pressure | Full error history for that event |
| Gauge reads good, controller still alarms | Sense line or controller input | Measure at the sensor terminals |
| Pressure builds, then decays over minutes | Slow leak: gasket, latch, relief setting | Isolate the supply and time the decay |
| Pressure never builds | Supply: valve, filter, kink — or a wide-open entry | Regulator output and filter drop |
| After washdown or a cold night | Water in the sense line, stiff wet gasket | Drain the tube; inspect the seal face |
Each row hands you one measurement to take before you order any part. Two of the six settle the direction in a single reading: the good-gauge case sends you to the sense path, and the never-builds case sends you to the supply.
In what order do you check the causes?
Check the supply first, the open entries and gasket second, and the sense line third; those three carry most low-pressure alarms, and the sensor element and the controller input come last, after the air path measures good.
- Supply sag — isolation valve partly shut, plugged filter element, kinked tube, regulator set below its label value. Giveaway: pressure never builds.
- Open entry — a missing blanking plug, an unloaded gland or a spare knockout, usually fresh wiring work. Giveaway: the first start-up after that work.
- Door gasket — cut, flattened, hardened, grit on the seal face, latches short of home. Giveaway: builds, then decays over minutes.
- Sense line — blocked, crushed or split tube, or water standing in it. Giveaway: port gauge good, controller low.
- Sensor element or its cable — the swap you make only after causes 1 to 4 measure good.
- Controller input or interlock logic — the one thing the bypass connector actually tests.
How do you check the supply and the leak paths?
Read the enclosure pressure at the sense port with the door shut and latched, and compare it with the 0.1 in w.c. minimum before you touch anything else. That single reading splits the fault in half: low or falling sends you to the supply and the leaks, at or above the minimum sends you to the sense line.
Feed the unit dry instrument air — a plant header typically runs 80 to 120 psi [5.5 to 8.5 bar] — and set the unit's regulator to the value on its label. Check downstream of the filter while the unit is flowing, because a loading element shows its pressure drop right there. Trace the supply tube for a kink, especially where it crosses a hinged door.
Then walk every entry on the box. The classic find after wiring work is a spare knockout or conduit entry with no plug in it; UL 50's Table 71-1A sets the knockout diameters and fitting dimensions, so a listed blanking plug exists for every hole — the cutouts and glands guide carries the sizes. For the gasket, close the door on a strip of paper at six points around the seal; where the strip pulls free, clamp is low. It is the same hunt as finding water inside a sealed enclosure: water and pressure leave by the same roads.
Worked example. A 36 x 24 x 12 in [915 x 610 x 305 mm] wall-mount cabinet on type Z purge alarms at every cycle start. The port gauge reads 0.04 in w.c. [10 Pa] latched — below the minimum. One 0.5 in [13 mm] conduit knockout stands open from a recent add: π/4 × 13² ≈ 133 mm² [0.2 in²] of leak area. For scale, at the minimum overpressure the 36 x 24 in door face (0.56 m²) carries 25 × 0.56 ≈ 14 N, about 3 lbf — that is the whole pressure budget, spread across the door, and the open hole dumps it. A listed blank goes in, the re-read is 0.11 in w.c. [27 Pa], and the alarm clears at the next start.
How do you check the sense line, the sensor and the controller input?
Clear the tube before you condemn the sensor: a blocked or split sense line reports low no matter how healthy the enclosure pressure is, and proving it takes ten minutes with a regulated air source and a gauge.
Disconnect the tube at the sensor and pass a few psi of regulated air through it, listening at the enclosure end. Drain any water you find — a low spot near a hinge collects condensate, and the reading then swings with temperature. Reconnect and compare the port gauge with the controller's display. When the port holds the minimum and the controller still says low, replace the sensor element; the boards you already swapped were never the problem. Only when sensor and cable both measure good does the fault sit at the controller input, and that is the one case the safety monitor view settles.
How do you stop the alarm coming back?
Age out the parts that fail on schedule: keep a spare gasket and a set of listed blanking plugs on the shelf, service the supply filter with the plant's routine, drain-loop the sense tube, and pull the bypass connector after every service visit.
- Fit the pressure-relief device the system maker offers. Most purge installations run without one, and a regulator that drifts high then sends its air somewhere — usually past the gasket.
- Where the hazard is combustible dust, the protection is pressurization that keeps dust out, and the standard has you clean the enclosure before you seal and pressurize it: pressure holds dust out, it does not remove what is already inside.
- Feed the system dry air. You are exchanging the enclosure volume with header-dew-point air, and the enclosure relative humidity calculation shows where that leaves the inside air.
- Write the healthy reading — say 0.12 in w.c. [30 Pa] — on the panel schedule. An alarm at 0.05 in w.c. [12 Pa] after a year at 0.12 is a leak; a reading that never builds is a supply.
- After any wiring change, walk the spare entries before you latch the door.
When should you re-specify the enclosure?
Replace the body when the box itself can no longer hold 0.1 in w.c.: a gasket channel gone corroded, a door face sprung, or a wall of unused holes from twenty years of changes costs more in stopped cycles than a new body.
Order the next body with the entries you actually use, gathered on a gland plate — when to specify a gland plate — so the next change happens on the plate rather than the body. The stainless enclosure size finder filters 304 and 316L wall-mount bodies by inside size, and a body built to your cutout drawing is delivered in 10 business days from your approval of the drawing.
Our enclosures are rated NEMA 4X / IP66 by the manufacturer; not UL listed — what a UL listing covers explains the difference when a project specification asks for a listing. The protection itself stays specialist work: we build general-purpose stainless enclosures, and we do not build hazardous-location enclosures or purge systems. The purge unit, its listing and the area classification belong to the purge system's maker and your authority having jurisdiction.
Enclosures for this job
Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 36 × 24 × 12 in | 304 stainless | 36 x 24 x 12 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 12 × 10 × 6 in | 304 stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 16 × 14 × 8 in | 304 stainless | 16 x 14 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 304 stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
How much pressure must a purged enclosure hold to stay out of alarm?
About 0.1 in w.c. above ambient. NFPA 496 and IEC 60079-2 set that minimum for the type Z / pzc level, and one purge-system vendor's technical note computes only about 9.4 lb [4.2 kg] of door force at that pressure. The margin really is that small, which is why a single open knockout can hold the alarm in permanently.
Why does the gauge read good while the controller still alarms?
Because the controller reads the sensor, and the sensor reads its tube. A blocked, crushed or water-filled sense line reports low no matter what the enclosure holds. Pull the tube at the sensor, prove it clear with a few psi of regulated air, and compare the port reading with the controller display; if the port holds its 0.1 in w.c. minimum and the alarm stays, swap the sensor element.
Can the purge bypass connector stay in for production?
No. The bypass is a service tool for an area the site safety authority has cleared of gas, and it satisfies the input without proving pressure. Fitted full time, it lets ignition-capable equipment run in a classified atmosphere with the protection defeated. Use it only as the manual directs during fault-finding, then remove it and re-run the complete purge cycle before the machine returns to service.
What is the fastest leak to find after new wiring goes into a purged cabinet?
An unplugged spare entry. A single open 0.5 in [13 mm] knockout is about 133 mm² [0.2 in²] of leak area, and it holds the enclosure below the minimum no matter where the regulator is set. Fit a listed blanking plug sized to the knockout, re-latch the door and re-measure; the reading should come back above 0.1 in w.c.
