Skip to main content

Worked Example: Rugged Control Box

This architecture study develops Exercise 1: a small sealed control box with 24 VDC input, Ethernet, CAN, four sensors, one motor output, and a debug port. We'll run each interface through the decision paths, then turn the result into the kind of paperwork the templates expect.

Illustrative composite — not a case report

The box below is a teaching scenario built on Exercise 1, not a report of any single project. The reasoning is the point, and every connector fact is stated by reference to the deep dives and decision paths, which carry the sources.

Read this as reasoning, not a shopping list

Like the Selection Packet and the M12 example, this page teaches how to decide and document. Part numbers, counts, and lengths marked J- or TBD are placeholders. A real project replaces every one with an exact, datasheet-verified value. Don't carry any rating from this example into your hardware.

Scenario​

A shoebox-class sealed controller mounted on a machine frame in a light-industrial bay: splash and dust, occasional washdown mist nearby, no submersion. It powers from the plant's 24 VDC supply, talks Ethernet to the line controller, sits on a small CAN segment with two other nodes, reads four DC sensors on the frame, drives one small 24 VDC gearmotor, and needs a service port a technician can reach without opening the enclosure. Cables are unplugged whenever the frame section is serviced — by whoever is on shift.

The quiet headline decision: every interface below lands on one circular family (M12-class), on purpose. This can simplify sourcing, but the exact coding, assembly, seal, and torque requirements still differ. Nine similar ports create a wrong-port risk that labels alone cannot close. The mating matrix below is a release gate, not a claim that the present placeholders are safe to build.

Requirements​

InterfaceRequirementNotes
24 VDC inputBox supply, current TBD from the power budget (assumed within a power-coded M12 class below — verify against the exact datasheet)Plant 24 VDC distribution; source-side protection TBD
EthernetOne run to the line controller; rate TBD from the controller spec — 10/100 assumed below, to be confirmedShielded run, washdown-adjacent
CANMulti-drop segment, this box + 2 nodes; bus power TBDTermination location must be designed, not discovered (§8.5)
Four sensorsDiscrete DC sensors on the frame, per-circuit current TBD vs. exact part ratingsSwapped by shift technicians — unplug/replug service model
Motor outputOne small 24 VDC gearmotor, stall current TBD from the motor datasheetNo separate feedback run in this scenario
Debug / service portReachable without opening the box; sealed when unused; low mating-cycle dutyUsed a few times a year, capped otherwise
Environment (all)Sealed-when-mated target per plant spec (exact IP class TBD); unmated state must also be protectedCaps are part of the design, not an accessory

Connector selection table​

InterfaceRequirement summaryCandidate familiesSelected familyStatus
J1 — 24 VDC inputSealed DC power inletM12 power-coded; bare cable gland; DT-styleM12 L-coded candidate (§8.1); exact device and cordset TBDPending power budget
J2 — EthernetSealed industrial EthernetM12 D-coded; M12 X-coded; sealed RJ45M12 D-coded, 4-pin (§8.1)Pending rate confirmation
J3 — CANSealed multi-drop fieldbus dropM12 A-coded 5-pin; hardwired glandM12 A-coded, 5-pin ("some CAN" rides A-coded — §8.1); topology per §8.5Candidate; exact mate and pinout open
J4–J7 — SensorsFour sealed DC sensor circuitsM12 A-coded 4-pin; M8; glands into a junction boxM12 A-coded, 4-pin (Industrial sensor path)Candidate; exact mate and pinout open
J8 — Motor outputSealed small-motor DC powerM12 power-coded; M23-class; DT-styleM12 T-coded candidate, distinct from J1; no shared power-cordset typePending stall-current check
J9 — Debug/serviceSealed, occasional-use service portSealed M12 A-coded 8-pin; USB-C behind a cover; Micro-D behind a hatchM12 A-coded, 8-pin, with a chained screw cap (Debug/service path)Candidate; exact mate and pinout open

Rejected connectors​

InterfaceRejected optionWhy rejected
24 VDC inputCable gland + hardwired tailFails the service model — every box swap becomes a wiring job, done by whoever is on shift; a gland is the right answer only for a run that never unplugs, and this run does unplug during service
24 VDC inputMIL-DTL-38999No program requirement to justify the cost/lead-time class (defense path logic in reverse)
24 VDC / motorM12 K-codedAC power coding (§8.1) — wrong lane for a 24 VDC system
24 VDC / motorA-coded pins doing power dutyA-coded connectors can carry power within their exact ratings. Here the supply and stall-current budgets remain open; retain the power-coded candidates until those budgets are checked
EthernetM12 X-coded "to be safe"D-coded is the simpler candidate for the assumed 10/100 link. X-coded can also support lower rates where the equipment, cordset ecosystem, or upgrade plan justifies it
EthernetSealed/rugged RJ45Workable, but breaks the one-family standardization and adds a second seal/latch system to maintain (Rugged Ethernet path checklist concerns)
SensorsM8Viable technically; rejected as a standardization decision — one shell size across the box (same reasoning as the M12 example)
SensorsGlands into a junction boxService model again — sensor swaps become enclosure work inside a sealed box; the M12 example rejects glands on the same grounds
Motor outputM23-classRight family for a real servo axis with feedback; oversized for one small DC gearmotor — becomes the answer if the motor grows (see "What would change")
All sealed runsDT-style sealed automotiveGood parts in their lane (budget path); rejected because the industrial cordset ecosystem here is M12 — fighting the ecosystem means custom pigtails
Debug/serviceBare USB-C on the panelThe debug path allows USB-C only behind a cover; a bare consumer port on a washdown-adjacent panel fails §12.4's boundary

Pinout​

Assignments below are illustrative structure only — every pin function, and especially every wire color, is verified against the exact device/cordset datasheet and recorded in the ICD before release. Wire-color conventions are example-only, never design authority (Source Notes).

ConnectorPinSignalDirectionNotes
J1 (24 VDC in)TBD from selected L-coded device drawingSupply circuit(s), associated returns, and FE per that deviceInSeparate circuit assignments; no contact paralleling or current-sharing credit
J2 (Ethernet)1–4TX+ / RX+ / TX− / RX− per the D-coded cordset datasheetBidirShield continuity per §5.7; category per the cordset spec
J3 (CAN)1–5Shield / bus power / bus return / CAN_H / CAN_L per the device and protocol documentationBidirMulti-drop: this port is one drop — T-piece and end-termination per §8.5
J4–J7 (sensors)1–4+V / signal / 0 V (+ spare or second signal) per each sensor datasheetInSame discipline as the M12 example
J8 (motor)TBD from selected T-coded assembly drawingOne motor feed and its return; remaining positions per approved assemblyOutOne contact per current path; stall/fault current and unused-position treatment must be verified
J9 (service)1–8UART TX/RX, boot/enable strap, logic ground, reservedBidirLogic-level only — no power export; capped when unused

Power pinout release gate​

The table above allocates functions; it is not a numbered wiring instruction. Coding is not permission to parallel contacts. For J1, select the exact L-coded device and cordset, then preserve its distinct supply/return and FE assignments. For example, Schmalz's L-coded device documentation distinguishes sensor and actuator supplies and their returns; it does not describe two interchangeable positive and two interchangeable return contacts.1 For J8, select a T-coded assembly whose individual current paths support the motor's load, stall, and protection-clearing conditions.

Before a cable can be built, attach mating-face drawings showing pin numbers, pin 1, key orientation, contact gender, and the two ends' wire mapping. Record current and fault duty per contact. Neither connector is released while these fields remain TBD.

Port-to-cable mating review​

The proposed contact genders below are architecture choices, not checked part selections. “Plug” and “receptacle” describe housings; pin/socket describes the contacts. Verify both in the exact drawings.

PortIntended cableProposed panel contacts / cable contactsCoding and positionsStatus
J1 supplyP1 plant supplyPins / sockets; live cable contacts recessedL, per selected device including FEOpen
J2 EthernetP2 controllerSockets / pinsD, 4Open
J3 CANP3 bus dropPins / sockets; confirm bus ecosystemA, 5Open
J4–J7 sensorsP4–P7 respective sensorsSockets / pinsA, 4Open
J8 motorP8 motorSockets / pins; driven panel contacts recessedT, per selected assemblyOpen
J9 debugP9 service cableSockets / pinsA, 8Open

Every cable class is compared with every panel class below. The sensor group represents all four cables and all four ports, including their 12 unintended permutations.

Cable → panelJ1J2J3J4–J7J8J9
P1 supplyIntendedBlockBlockBlockBlockBlock
P2 EthernetBlockIntendedBlockBlockBlockBlock
P3 CANBlockBlockIntendedCheck ABlockCheck A
P4–P7 sensorsBlockBlockCheck AFunction checkBlockCheck A
P8 motorBlockBlockBlockBlockIntendedBlock
P9 debugBlockBlockCheck ACheck ABlockIntended
  • Intended: confirm the complete mating pair, wiring, ratings, and permitted service state.
  • Block: different coding is the intended mechanical exclusion. Verify exact parts and any adapters; do not treat this matrix as test evidence.
  • Check A: position count and proposed gender alone are not proof of exclusion. Check drawings and actual samples. A possible mate must be electrically safe under its resulting pin mapping, or the architecture must change.
  • Function check: identical sensor ports can accept the wrong sensor. Establish whether each swap is harmless. Where it can command hazardous motion, use physical exclusion or a validated identification/interlock strategy before enabling the motor; labels and a service checklist alone do not establish safety.

For each unintended mate, record the physical result, resulting net connections, electrical consequence, mitigation, and verification evidence. Wrong supply-to-motor connection could energize the motor or backfeed the controller; incompatible A-coded mappings could put supply voltage on data or logic. These are design risks to resolve, not proven faults in unspecified parts.

Cable drawing​

One row class per run; the cable drawing template carries the full per-cable version with evidence columns.

WireGaugeColorPair / shieldEnd AEnd BNotes
J1 power conductorsPer cordset spec (sized to the power budget)Per cordset — recorded, not assumedUnshielded powerJ1 plug (molded)Plant 24 VDC dropVerify current at temperature vs. the exact assembly
J2 EthernetPer cordset category spec—Shielded, pairing per specJ2 plug (molded)Line controllerShield termination both ends per §5.7
J3 CAN dropPer CAN cable spec—Shielded pair + powerJ3 plugBus T-pieceDrop length within protocol limits (§8.5)
J4–J7 sensor runsPer cordset spec—UnshieldedSensor M12J4–J7Off-the-shelf molded cordsets, stocked lengths
J8 motor runSized to stall current—Per motor/EMC needsJ8 plugMotor leadsAnchored against gearmotor vibration
  • Cable jacket: oil-resistant per the plant environment spec; drag-chain-rated only where a run actually moves (none in this scenario — record that, too)
  • Labels: every cable labeled at both ends with the J-number; label spec in the ICD
  • Length tolerance: per the cable drawing template's tolerance field, set from routing reality, not guessed
  • Test requirements: continuity + hipot/IR per the harness inspection checklist and the acceptance standard the program names; design-level qualification evidence, where the program requires it, structured per the qualification plan template

Backshell / cap notes​

Molded cordsets carry their own strain relief and sealing, so this box needs no field backshells — the accessory budget goes to caps and torque instead.

InterfaceAccessoryNotes
All receptaclesSealing screw caps, chainedThe unmated state is part of the sealing design (§8.3); verify the receptacle's unmated rating; exposed contacts and the panel penetration are separate sealing boundaries
J9 service portCap plus a log disciplineThe port is capped 360 days a year — the cap is the interface most of the time
AllTorque per each part's datasheetCoupling-torque values are manufacturer-specified (§8.3); finger-tight is not a spec
CAN segmentTerminating M12 plug at each bus endTermination lives at the bus ends, not "wherever" — §8.5

ICD entry​

One worked entry (J2) in the ICD template's shape; the real package carries one per interface.

  • Interface name: J2-RCB-ETH — control box to line controller, industrial Ethernet
  • Connector P/N: TBD — M12 D-coded 4-pos panel receptacle, front-mount, shielded
  • Mating connector P/N: TBD — molded D-coded cordset, shielded, length per routing
  • Contact P/Ns: n/a (molded assembly) — field-repair kit P/N TBD, temporary-article rule per the M12 example §6
  • Backshell P/N: n/a (molded); panel-side sealing per receptacle datasheet
  • Dust cap P/N: TBD, chained
  • Keying / polarization: D-coding is the key (§8.1); no same-coding neighbor on this panel face
  • Electrical limits: per the exact receptacle/cordset datasheets — recorded here with revision, never quoted from memory
  • Shielding: 360° continuity through the receptacle to the enclosure bond point (§5.7)
  • Environmental assumptions: sealed-when-mated per the exact assembly's tested rating; capped when unmated; plant spec TBD governs
  • Mating cycles: per datasheet (A4 carries the family-level orientation)
  • Torque / assembly notes: coupling torque per datasheet; torque tool TBD; record value + tool in the build record
  • Cable requirements: category/shield per the cordset spec; bend radius per datasheet; no drag-chain duty
  • Test / inspection requirements: harness inspection checklist + program acceptance standard

Risks and mitigations​

RiskMitigation
Nine same-size circular connectors (six interface types) on one small box — wrong-port mating attemptsClose every cell of the port-to-cable mating review above; verify actual parts and resulting net connections. Labels and panel photos supplement mechanical or electrical mitigation
A-coded pins drafted into power duty during a "quick fix"Verify actual load, fault duty, and assembly ratings; do not infer suitability or exclusion from A-coding alone (§8.1)
CAN termination forgotten or duplicatedTerminating plugs are BOM line items with J-numbers, not accessories; topology drawing in the package (§8.5)
Receptacles left uncapped after serviceChained caps; capped-state photo in the close-out checklist; the service port's cap logged like a tool
Motor stall current outgrows the selected power codingStall check is a release gate (requirements table); escalation path pre-named (M23-class — see below)
Ethernet rate requirement changes after releaseThe D-vs-X decision and its trigger are recorded (§8.1) — a rate change reopens J2, not a debate
Cordset vendor change alters pinout/wire colorsColors are example-only by rule; the ICD pins functions to pin numbers with datasheet revisions (Source Notes)

What would change if…​

  • The power budget outgrows the M12 power class — J1/J8 move up a family (the high-current path takes over, derating curve first), and the one-family standardization is consciously traded away, in writing.
  • The motor becomes a servo axis — feedback appears, and the whole J8 question transfers to the motor + feedback cable path: drive-ecosystem cordsets, M23-class connectors, and the EMC discipline that comes with them.
  • The box moves into direct washdown — re-verify every mated and unmated rating against the plant spec (§8.3) and re-run the sealed enclosure feedthrough path for the panel itself.
  • A defense/aero customer appears — the requirement set, not preference, moves the external interfaces toward 38999-class hardware, and the Selection Packet shows what that documentation grade looks like.
  • Someone proposes one big multipole for everything — that's the removable machine module path's problem statement; run it honestly before deciding: use the path for the service reality, and §5's anatomy for the blind-mate and float mechanics a docking interface leans on.

Documentation bundle​

The following items remain open in this study. Complete them before treating it as a released design:

  • Requirements table with sources and status closed out
  • Port-to-cable mating matrix closed with exact parts, consequences, and evidence
  • J1/J8 numbered pinouts and per-contact normal/stall/fault-current budgets closed without unapproved paralleling
  • Comparison matrix rows for the contested decisions (J1 power coding, J2 D-vs-X)
  • Pinout + cable drawing per run, evidence columns filled
  • ICD entry per interface (J1–J9), datasheet revisions recorded
  • Design review checklist pass — the caps, torque, and termination rows are the ones this box lives or dies by
  • Harness inspection checklist on the built cables
note

These examples illustrate structure and process, not a released design. The datasheet, applicable standard, and program requirements decide the actual values.

Footnotes​

  1. Schmalz, SCTSi IOL operating instructions, §8.6, L-coded power connection (device-specific example, not a universal pinout). https://pimmedia.schmalz.com/MAM_Library/Dokumente/Bedienungsanleitung/30/3030/303001/30300101598/dd27ee2a2ed2_BAL_30.30.01.01598_en-EN.pdf ↩