Membrane Switch Pinouts: Can One Key Prove the Map?

Give physical positions and electrical nets different columns

Use a fixed installed view to identify a tail contact and its receiving connector position. Keep that physical identity separate from the net name and key function. A useful interface row reads: tail contact T1, connector position J1-1, common net COM, drawing revision and view. These are illustrative identifiers; the chosen component drawing supplies the real numbering. Do not force every tail contact number to equal the connector number when the approved routing says otherwise.

Also distinguish the conductive face from the electrical order. Molex's 15266 series chart lists four-circuit 152660029 with same-side Type A contacts and 152660030 with opposite-side Type D contacts. That identifies a contact-layout difference in named FFC jumpers. It is not a rule that Type D reverses a custom membrane tail's pinout, nor a statement that either jumper belongs in your assembly.

Hirose's FH34(S) operation notes show conductive surfaces facing up for the illustrated top-contact specification. This is a part-specific orientation example, not an instruction for every ZIF connector. Read the actual selected-part view before assigning the physical positions. Review dimensional mating separately in the HYR tail-thickness guide; matching thickness does not settle net identity.

Worked example: Reset passes even when the assignment is reversed

Assume four physical positions, always numbered 1 through 4 in the same reference view. Assume three ideal normally-open switches: Start connects common position 1 to position 2, Stop connects 1 to 3, and Reset connects 1 to 4. There are no other components or paths, and only one key is pressed at a time. This is an invented circuit for reasoning, not the circuit in the product photograph.

Now introduce one deliberate fault: assign each original position p to position 5 minus p. The reference positions themselves do not change names. Common moves from 1 to 4; Start's other endpoint moves from 2 to 3; Stop's from 3 to 2; Reset's from 4 to 1. Start therefore closes positions 3 and 4 instead of 1 and 2. Stop closes 2 and 4 instead of 1 and 3. Reset still closes the same unordered pair, 1 and 4.

A check limited to continuity between 1 and 4 while pressing Reset would produce the same ideal result for the approved assignment and this fault. Both assignments also leave every pair open when all keys are released. The comparison below makes that limited discrimination visible. It does not predict what an actual host controller would do, and the reassignment does not imply that a physical connector can be inserted backward.

Write expected states before reading the test report

For this small passive example, list the six unordered pairs: 1–2, 1–3, 1–4, 2–3, 2–4 and 3–4. In the released state, all six are expected open. With Start alone, only 1–2 is expected closed; with Stop alone, only 1–3; with Reset alone, only 1–4. Six pairs across these four states create 24 expected entries. This enumeration is an illustrative coverage worksheet, not a requirement that every real assembly use 24 measurements.

Recording both expected connections and expected nonconnections gives the reviewer a basis to challenge unintended paths. The project's engineer still chooses the actual coverage, resistance and isolation limits, safe stimulus, settling time, fixture and measurement method. A continuity beeper's threshold is not automatically the product's acceptance limit. Components or alternate paths may make the simple open/closed model inappropriate.

Keep the pair worksheet separate from lifetime, insulation, environmental and contact-resistance qualification. A correctly mapped new specimen does not establish those properties. If the project uses a matrix, assign rows and columns under the released circuit and review simultaneous-key behavior separately rather than transferring this common-line example.

Check the fixture against a reference it did not create

A broader scan does not solve a shared-reference mistake. If the tail drawing and fixture harness both inherit the same mirrored numbering, their records can agree without matching the released receiving connector. Independently trace fixture channels to physical connector positions using the selected component drawing and controlled assembly view. Keep the observed wiring record and the expected net table identifiable.

Ask the review owner to inspect the fixture revision, channel labels, harness routing and any software mapping. Identify which document governs when they disagree. A useful sample record links specimen ID, tail drawing, connector order number, fixture revision, expected pair state, observed reading and disposition. An unlabeled meter photograph does not supply that chain.

After the passive map is accepted, the equipment owner can separately verify physical key ID against the actual controller net and intended response in a safe, approved setup. Do not use unexpected operation of live equipment as a pinout diagnostic. Powered behavior, interlocks and safety decisions remain outside a loose membrane-switch continuity result.

Challenge five realistic handoff gaps

These are proposed review scenarios, not HYR field cases or frequency estimates. Use the one that matches the project's uncertainty, and name the evidence that would resolve it.

• Only the symmetric key is checked: Reset's 1–4 pair passes in both illustrative assignments. Review coverage for the other keys before accepting the map.
• Contact face is treated as numbering: a contacts-up instruction is mistaken for a Pin 1 definition. Resolve the selected-part physical view and net table separately.
• Fixture and drawing share the same wrong reference: a broad scan agrees with itself. Independently trace fixture channels to the receiving connector's controlled positions.
• Only intended closures are recorded: unintended pair connections have no stated expectation. Have the engineer define relevant nonconnection and isolation checks.
• A mapping change is hidden in an artwork revision: a moved legend is accepted without checking its key ID and assigned net. Review the changed dependency before transferring old sample evidence.

Send one mapping record with the RFQ

For an initial quotation, send the information you have and label unknowns. For sample acceptance, resolve the physical reference, exact circuit and coverage plan. Give electrical engineering ownership of the expected net map, manufacturing or quality ownership of the observed fixture record, and the equipment owner authority over controller behavior. These are proposed responsibilities to agree on, not a claim about HYR staffing or an already approved workflow.

Review the record again when a tail route, contact face, connector, circuit assignment, key legend or fixture changes. Not every geometry change changes a net, but it should receive an explicit dependency decision. Preserve the accepted configuration and distinguish unchanged evidence from evidence that must be repeated. The useful outcome is a traceable map, not a larger collection of unnamed pass screenshots.

Use the secure engineering RFQ to share the keypad view, connector drawing, key-to-position table and unresolved coverage question. HYR's quotation discussion can then be scoped to the membrane-switch assembly. No sample test, production coverage or finished-equipment approval is asserted by this editorial example.