Touch safe control panel validation: IP2X test and label scope

Short answer: validate each panel with the IEC 60529 jointed-finger test, done by a qualified person, and label the exact condition you tested. The probe is IEC 61032 probe B, an articulated finger 12 mm in diameter and 80 mm long, pushed against every opening with 10 N of force while every cover and barrier is fitted. When it cannot touch a live part anywhere, you document the pass and label the panel IPXXB for that configuration. Claiming IP2X adds one test on top: a 12.5 mm rigid sphere must fail to pass through any opening. The pass buys shock protection only. An arc flash study showing under 1.2 cal/cm² at the working distance sets no arc-rated PPE for work at that panel, and that decision comes from the study, not from the label.

What IPXXB and IP2X actually test

IEC 60529 grades two things in one code: the digits grade ingress, and an additional letter grades access to hazardous live parts. IPXXB uses the letter B alone, so it claims one thing — the jointed test finger cannot touch a live part — and says nothing about water or dust. IP2X uses first digit 2, which runs the same finger test and adds an object test: a 12.5 mm rigid sphere must stay outside every opening. The full digit meanings sit in the IP rating chart, and what makes an enclosure finger safe covers the concept in the round.

The articulation and the stop plate are the point of the calibrated probe. It bends around a terminal shroud the way a hand does, and the 50 mm plate stops it after 80 mm of penetration, so the depth of live parts behind an opening is part of the result. Push with 10 N — that is firm hand pressure, and a barrier that deflects aside under 10 N fails. IEC 60204-1, the electrical standard for industrial machinery, asks for IPXXB at minimum for live parts on a machine, so machinery auditors reach for this same probe.

Frame the paperwork the way the enclosure business does. IP and NEMA ratings are manufacturer ratings — our stainless enclosures are rated NEMA 4X / IP66 by the manufacturer and not UL listed — and an in-house IPXXB validation is the same kind of documented self-assessment. When a customer or inspector wants third-party backing, that is the enclosure listing route for a control panel. NEC 110.28 keeps the two systems separate in US installations: an IP code never replaces an enclosure type rating for the location, so the box still carries its type while IPXXB speaks only to finger access.

What you need on the bench

• A jointed test finger built to IEC 61032 probe B: 12 mm diameter, 80 mm of finger, a 50 mm stop plate, and a handle force gauge that reads 10 N.
• A 12.5 mm rigid sphere when the label claims IP2X.
• The panel drawing with every opening numbered, plus the arc flash study page showing incident energy and its working distance.
• A record sheet with one row per opening, a camera, and a mirror or borescope for the back of a deep cabinet.
• Label stock that survives the panel's environment, and training time for the qualified staff who will read it.

Validate the panel in eight steps

1. Scope the panel. Record the panel ID, the highest voltage inside, the criterion you chose — IPXXB or IP2X — and the barrier condition you will test: door closed, door open with all covers on, or both. The label grants only the condition you test.
2. Pull the study numbers. Copy the incident energy and its working distance from the arc flash study onto the sheet, for example 0.8 cal/cm² at 455 mm [18 in], the usual working distance for a low-voltage panel. That number decides arc-rated PPE; the finger test has no vote on it.
3. Open it safely. Lock out and verify absence of voltage where production allows, then test de-energized — the finger result is identical. When a check must happen energized, the tester works as a qualified person under your NFPA 70E rules, in the PPE those rules set.
4. Number every access path. With all covers fitted, walk the panel and list every place a finger could enter: the door seam, the latch area, the gland plate, louvres, the gap around a display, the clear covers over terminal blocks. Expect 10 to 20 openings on a typical machine panel.
5. Apply the finger. At each opening push probe B with 10 N straight in, then angled, then articulated around whatever sits behind it. One touch of a live part fails the whole panel.
6. Add the sphere for IP2X. Offer the 12.5 mm sphere at every opening. Anywhere it passes through, the ingress claim fails even though the finger test passed.
7. Correct, then retest the whole list. Fit the missing cover, blank the oversize hole, then re-run all openings — 14 of 14, not only the two you repaired.
8. Sign, label, train. The qualified person signs the record; the label states the code, the barrier condition, the date and the next retest; and every qualified person on the floor learns what it grants.

Validating a 480 V PLC panel: 14 openings, two failures

The inputs: a machine panel fed at 480 V three-phase with 24 V control; incident energy 0.8 cal/cm² at 455 mm [18 in] from the facility's arc flash study; 14 access paths counted on the door drawing; every barrier fitted at the walk-down.

The rule: probe B — 12 mm diameter, 80 mm of reach before the stop plate — pushed at 10 N at every opening, with the 12.5 mm sphere added because the plant wants an IP2X label. Any opening where the finger touches a live part, or the sphere passes through, fails the panel.

The walk-down: 12 of 14 openings pass. Opening 13 is a terminal strip whose clear cover went missing after a repair, so the finger lands on 480 V terminals: fail. Opening 14 is a 15 mm hole left by a removed knockout, with a live busbar 70 mm behind it. The finger fits, because 15 mm clears its 12 mm body, and it reaches, because 80 mm of finger beats 70 mm of depth: contact. The 12.5 mm sphere also slips through the 15 mm hole, so the ingress claim fails too. Twelve passing openings out of fourteen is a failed panel — the share of finger-safe components never enters the test.

The arithmetic of the fix: blank the knockout with a plug carrying an 8 mm hole. At 8 mm the opening sits under the finger's 12 mm and under the sphere's 12.5 mm, so both probes stay outside and the depth question disappears. Refit the clear cover on opening 13, re-run all 14 openings, and the panel passes 14 of 14.

The decision: 0.8 cal/cm² sits under NFPA 70E's 1.2 cal/cm² (5 J/cm²) threshold, so the study sets no arc-rated PPE at the working distance, and the validated barriers close the shock path, so insulated gloves and leather protectors stay on the shelf for covers-on work. The label reads IP2X validated, all covers installed, door closed or open, with date and signer. Pull one cover and the panel is unvalidated again — the label says so in those words.

What the label grants, and what it leaves out

A pass answers one question: can a finger reach a live part in the tested condition. It lets a qualified person work at the panel without insulated gloves while the tested barriers stay in place, because the shock path is closed. Approach boundaries and safe access still govern how you come at the enclosure, and a metal box that tingles needs the shock from a metal enclosure checks before anyone trusts a label.

Arc flash stays a separate decision with its own inputs. NFPA 70E draws the arc flash boundary where incident energy reaches 1.2 cal/cm² (5 J/cm²), and inside that boundary the study sets the PPE for the task. A terminal cover that stops a 10 N finger stops no arc; plasma crosses a plastic barrier about as easily as air. Panels holding several circuits fed from different sources carry their own incident-energy math, covered in arc flash incident energy for multiple circuits in one box. Who counts as a qualified person at these voltages is spelled out in NFPA 70E qualified-person training for arc flash compliance — put that training on the record next to the signature.

The label also changes nothing physical around the panel. NEC 110.26(A) still requires clear working space in front of equipment likely to be serviced energized, with room for the door to swing at least 90 degrees.

Checks when the label goes on

• The full opening list passes end to end after every correction — 14 of 14 in the example.
• A photo of every opening sits in the record, so the next reviewer sees the as-tested condition.
• The label is readable at the door handle and states code, barrier condition, date and signer.
• The sheet cites the arc flash study revision it was checked against.
• Any qualified person asked can say in one sentence what the label grants: shock protection, covers on, this configuration.

Mistakes that void the validation

• Testing a kinder condition than the label claims. Tested door closed, labeled for open-cover work: the label lies. Test the condition you print.
• The bare-hand check. Your own finger is wider than the 12 mm probe, so a hand that reaches nothing proves little — the slimmer probe goes where a hand cannot. Buy or rent the real probe; it is standard IEC 61032 hardware.
• Counting components instead of openings. Ninety-five percent finger-safe components leaves one open path in twenty, and one path is a fail. Count and test openings, and percentages never come up.
• Printing "no PPE" on the label. The study relieves PPE below 1.2 cal/cm² at its stated working distance; the label grants access. Print the tested condition and let each document do its own job.
• Letting it go stale. A new gland, a fresh cutout, a cover off and back on — each one resets the panel to unvalidated until the list is re-run.

When the same openings keep failing — field-drilled holes oversize for their glands, hinge gaps worn wide — fix it at the source. The enclosure selector sets the size and rating for a replacement body, and factory cutouts on a custom stainless enclosure sized to your glands and display mean the next validation matches the drawing instead of chasing it.

Enclosures for this job

Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.

Questions