Enclosure internal temperature sensor without losing NEMA 4X

Measure the box from inside and nothing has to pierce it: a two-wire Pt100 transmitter and its probe both live on the 35 mm DIN rail, and one M20 cable gland carries the 4-20 mA loop out through the wall. One penetration, one rated part, and the enclosure keeps its rating. Set the PLC alarm at 40 C (104 F), the ambient rating most drives and PLCs carry, and the same signal doubles as proof that a cooler or a derate is doing its job. The steps below pick the parts, place the probe, run the loop arithmetic, list the checks, and name the mistakes that quietly cost the NEMA 4X.

How do you monitor the air inside without drilling the box?

Put the whole measurement chain inside: a Pt100 probe and a two-wire DIN rail transmitter clipped to the 35 mm rail, with only the 4-20 mA loop leaving through one M20 gland, so the wall keeps a single rated penetration and the box holds NEMA 4X / IP66.

You are measuring the air the electronics actually breathe, and that is the number that decides component life. A monitor on the outside wall reads the room or the boiler house, never the box. A probe poking through reads wherever its tip happens to land. Inside, the RTD sees the mixed air temperature and the only thing crossing the wall is a cable your PLC already expects.

Rail transmitters come in two shapes. A fixed-sensor unit reports the temperature at its own terminals, which sit beside your power supply on the rail. A remote-probe unit takes a 1-2 m lead and measures wherever you clip the tip, so take that kind. Many universal models can also run the span reversed, 20-4 mA, if your alarm logic prefers a falling signal as the box heats.

When the point you care about sits outside the box — a process line, incoming air, a chiller — the probe itself takes the gland route instead; the questions at the foot of this page cover that variant. For the temperature of the box itself, the all-inside layout wins on every count: fewer holes, one gland, and a sensor you can reposition with a zip tie.

What do you need for the job?

Four parts do the whole job: a 35 mm rail two-wire transmitter with a Pt100 input, a 6 mm [1/4 in] stainless-sheathed probe on a 1-2 m three-wire lead, one M20 gland its maker rates IP66 or better, and 24 V DC from the panel's existing supply.

• Transmitter: two-wire, DIN rail mount, Pt100 input, 4-20 mA output; universal models accept thermocouples too.
• Probe: Pt100 per IEC 60751 — 100.0 Ω at 0 C, rising about 0.385 Ω per K — in a 6 mm stainless sheath, three-wire so lead resistance cancels.
• Gland: M20, clamping range matched to the loop cable's roughly 6 mm [0.24 in] jacket; nickel-plated brass or stainless on a 316L washdown box so the gland cannot rust-streak the wall.
• Loop cable: shielded two-core instrument pair, 0.5 mm² (20 AWG) or heavier.
• Receiver: a 250 Ω resistor if the PLC analog input reads 1-5 V.
• Checkout gear: a resistance decade box or loop calibrator, plus a handheld thermometer.
• The enclosure: a stainless wall-mount box rated NEMA 4X / IP66 by the manufacturer; not UL listed. The size finder for stainless wall-mount enclosures filters by grade and outside size, and the gland cutout can be ordered with it.

Where should the sensor tip sit?

Clip the tip two-thirds of the way up the box, in free air, at least 2 in [50 mm] clear of every heatsink, power supply and wall, because a probe inside a sink's airflow reads that component's skin, which at load can sit 20 K above the air you are judging.

Warm air stratifies against the top of a sealed box, so the upper third reads the worst air the top-shelf electronics see — which is the number that matters. Work the placement with these rules:

• Put the tip in the upper third, 2 in [50 mm] off every sink, supply and wall.
• Keep it out of a vortex cooler's outlet stream; that jet reads cold while a far corner cooks.
• Allow for a door-mounted display: its electronics, and any sun on the door, warm the front zone, so favour the rear half of the backplate.
• Leave the lead in a loose coil, tied off, so opening the door never tugs the tip.
• Pick a spot your handheld meter can reach for the comparison check later.

A 6 mm sheath in still air follows enclosure temperature swings well within a minute, so position, not response time, decides what the reading means.

How do you install it step by step?

Eight steps take about an hour once the box is on the bench: rail the transmitter, place the probe, order the M20 cutout, gland the loop, land the wiring, set a 0-80 C span, simulate both endpoints, and put the alarm where the electronics' rating sits.

1. Rail the transmitter. Clip it to the 35 mm rail; mounting DIN rail on a backplate shows the spacing and the end brackets.
2. Place the probe. Zip-tie the tip two-thirds up, 2 in [50 mm] clear of sinks, lead in a loose coil.
3. Order the cutout. One M20 hole in the bottom wall keeps drips off the gland; the cutout designer checks the spacing and sends the layout with the quote, and the finished box is delivered in 10 business days anywhere in the world, counted from drawing approval on custom sizes. Side entry works when the layout forces it — plan a drip loop outside. Hole size per material is in M20 gland hole sizes.
4. Fit the gland. Installing a cable gland without losing the rating gives the full sequence: right hole, seal seated flat, cap torqued, jacket gripped round.
5. Land the loop. 24 V positive into the transmitter, loop out through the gland to the 250 Ω receiver, shield landed at the receiving end only.
6. Set the span. 0-80 C scaled to 4-20 mA puts 40 C mid-scale at 12.0 mA and keeps the reading off zero when the box is healthy — the live zero that proves the loop is intact.
7. Simulate the endpoints. A decade box at 100.0 Ω should read 4.0 mA; at 115.5 Ω (Pt100 at 40 C) it should read 12.0 mA.
8. Set the alarms. High alarm at 40 C (12.0 mA); treat 3.6 mA and below as a broken probe, the fault band NAMUR NE43 defines.

How does the 4-20 mA loop leave a sealed box?

Through a single gland whose maker publishes IP66 or better, seated on the cable jacket: the gland becomes the lowest-rated part in the wall and the assembly rating rides on it, which is why hole size, seal seat and cap torque decide the result.

The digit 6 in IP66 is the powerful-jet test of IEC 60529: a 12.5 mm nozzle throwing 100 L/min at the wall. A gland whose datasheet carries second-digit 6 or higher, with the jet test shown, stands up to the same blast the box passed, and with the seal seated and the cap torqued the assembly holds together — the NEMA 4X vs IP66 guide maps which test proves which digit. The lowest-rated part in the wall sets the number for everything inside it, so the gland, not the transmitter, is the part that matters here. For US drawings, spell the NEMA type out beside the IP code, because NEC 110.28 treats the two as different answers and the inspector reads the type.

Size the loop before you promise a long run. Assume a 24 V supply, a 250 Ω receiver and 100 m of 24 AWG shielded pair at 84 Ω/km per core: loop resistance is 250 + 2 × 100 m × 0.084 Ω/m = 266.8 Ω, the 20 mA drop is 266.8 Ω × 0.02 A = 5.3 V, and the transmitter sees 24 V − 5.3 V = 18.7 V. That clears an assumed 10 V floor with 8.7 V to spare; the datasheet's own floor figure replaces the assumption. The same run on a 12 V loop leaves 6.7 V and starves, which is why two-wire transmitters get 24 V.

What reading means the box is in trouble?

Alarm at 40 C (104 F): that is the ambient rating most drives, PLCs and power supplies carry unless the datasheet allows more, so a sustained reading above it is spending component life even while everything still runs.

Worked example — what a 60 W load does to a sealed box. Take a 16 x 12 x 8 in [406 x 305 x 203 mm] stainless wall-mount box. Total surface is 2 × (16×12 + 16×8 + 12×8) = 832 in²; the face flat against the wall moves little heat, leaving 640 in² = 0.41 m² to do the work. Stainless passes roughly 3.7 W/m²K through the wall, the value behind our sizing charts, so the steady rise is 60 W ÷ (3.7 W/m²K × 0.41 m²) ≈ 39 K. At 35 C ambient the probe settles near 74 C and the PLC sees 18.8 mA on the 0-80 C span. The 12.0 mA alarm tripped hours before that, which is the point: the trend warns while there is still time to act.

Read the same arithmetic backwards for the limit: passive walls hold 40 C only while ambient stays below about 1 C. Halve the load to 30 W and the rise halves to about 20 K, which is still 55 C at 35 C ambient. The watts set the number, so cut the load or add a cooler, and let this transmitter prove which one worked.

How do you prove the reading and the seal are good?

Four checks close the job: decade-box points of 100.0 Ω → 4.0 mA and 115.5 Ω → 12.0 mA, a handheld meter within 2 K of the tip, a gland cap that passes a firm tug, and a 24 h trend matching the predicted rise.

• Endpoints: 100.0 Ω → 4.0 mA and 115.5 Ω → 12.0 mA, each within 0.05 mA.
• Tip against a handheld meter: agree within 2 K after a minute at the tip; a bigger gap means the probe sits in a hotspot or a draught.
• Fault injection: lift one probe lead and watch the reading fall to the 3.6 mA fault band.
• Gland: cap at the maker's torque, jacket visibly gripped, and a firm hand-tug moves nothing; drip loop present on a side entry.
• Gasket and door: gasket seated all round, latch pressure even top and bottom.
• 24 h trend: the logged curve lands within a few kelvin of the rise the load arithmetic predicts; if it does not, move the tip 2 in [50 mm] and re-run the day.

Six checks, about ten minutes, and you have both a calibrated channel and a seal you can point to.

What mistakes cost the NEMA 4X rating?

Three errors undo the rating, and all three show up before the door closes: a probe hole closed with epoxy putty instead of a rated gland, a gland gripping a connector moulding instead of the jacket, and an old cutout plugged with sealant alone.

• Epoxy pass-through. Putty has no test behind it, so the hole counts as a field cut. A hole already in the wrong spot can be recovered — sealing a wrongly drilled hole in an enclosure explains what brings it back.
• Connector through the gland. The clamp seals on a round jacket; a moulded plug cannot seat. Use a jacketed tail and make the joint inside — getting a connector through a cable gland covers the options.
• Sealant-only plugs. A spare cutout takes a rated metal plug sized to the hole.
• A louver for cooling. The louver's own rating becomes the box's rating, so keep the wall sealed and cool externally; the sealed-box vs vent question outdoors shows why the sealed wall also stays drier.
• Cutting without a plan. Holes decided at order time keep the rating story simple; what a panel keeps after you cut it and mount devices maps which devices and holes survive which rating.

Ordered this way — sensor inside, loop out through one rated gland, alarm at 40 C — the monitor adds a measurement and never touches the rating.

Catalog dimensions to compare

Compare these catalog dimensions with your component layout. Open each enclosure for its rating, material and usable space, then check those specifications against your application.

Questions