Membrane Switch Circuit Resistance: Set an RFQ Limit

Measure the closed path, not an undefined contact

Label the two pins used for the closed-circuit measurement. In a simple switch-to-ground input, the relevant path can include two printed routes, the switching contact and connections between the keypad and receiver ground. A bench result taken at the tail may exclude a connector that is present in the equipment. Keep those results separately named.

Ask for resistance in ohms under agreed actuation and electrical conditions, with the fixture contribution addressed. Record the key, sample revision, force or displacement condition, measurement timing, temperature and test current or voltage as applicable. These are proposed RFQ fields, not a reproduced standardized procedure.

ASTM's record for F1680-07a(2014) identifies circuit-resistance measurement and marks the method withdrawn in 2023 with no replacement. If it appears on an old drawing, have the requirement owner resolve the exact historical reference and measurement plan. The catalog description supplies neither a universal pass limit nor an active replacement method.

Read guaranteed input limits before choosing a limit

For a pull-up resistor connected to the supply and a closed switch path to ground, a first DC screen is Vpressed = Vsupply × Rpath / (Rpull + Rpath). This ideal divider neglects receiver leakage and other connected networks. Include those contributions in the actual design model; do not equate this calculation with system verification.

Use the selected receiver's guaranteed limits at the applicable supply and temperature, rather than its typical switching point. As one bounded example, TI's SN74LVC1G04 data sheet specifies a low-level input maximum of 0.8 V and a high-level input minimum of 2.0 V for a 3.0–3.6 V supply. These are not generic microcontroller limits, and this ordinary CMOS inverter is not proposed here as a complete direct-switch interface.

Choose a design target inside the guaranteed region. In the unloaded divider, a chosen low-voltage target gives Rpath ≤ Vtarget × Rpull / (Vsupply − Vtarget). The result is a preliminary budget shared by the defined path, not an automatic drawing tolerance. Account for component tolerances, leakage, ground differences, aging assumptions and measurement uncertainty before approving an allocation.

Worked example: neither pull-up passes the whole screen

This is a hypothetical purchasing example, not a customer case or measured result. Assume a passive switch-to-ground path of 3.0 kΩ, a supply spanning 3.0–3.6 V and two proposed pull-ups: 10 kΩ and 47 kΩ. Use the example logic limits above only to demonstrate the comparison. The assumed path resistance is deliberately illustrative, not a typical or acceptable membrane-switch rating.

At the 3.6 V supply corner, the unloaded pressed-state calculation gives 0.831 V with 10 kΩ and 0.216 V with 47 kΩ. The first misses the 0.8 V low-level requirement even before adding other effects. The second clears this one calculation, but that does not approve it.

Now consider release. Assume, separately, a hypothetical total leakage of 30 µA sinking from the input node to ground while the switch is open. This is an assumed whole-interface load, not the TI part's specified input current. With Vreleased = Vsupply − Ileak × Rpull and the 3.0 V supply corner, the options give 2.70 V and 1.59 V respectively. The larger resistor now misses the 2.0 V high-level requirement.

The two rows answer different corners and use simplified models; they are not a full tolerance simulation. Their useful conclusion is that neither proposal is ready on the stated assumptions. The buyer must revisit the path budget, leakage budget or input architecture with the electronics owner, then repeat both state checks. Simply asking for a stronger pull-up does not resolve the complete problem.

If the buyer additionally chooses a hypothetical 0.6 V low-state target at 3.6 V, the unloaded resistance ceilings are 2.0 kΩ with 10 kΩ and 9.4 kΩ with 47 kΩ. These are worksheet outputs only. The failed released-state check prevents treating the larger number as an approved supplier allowance.

Keep DC margin separate from bounce and settling

A steady reading does not describe every press, release or sampling instant. TI's Debounce a Switch brief highlights the joint selection of filtering, resistance, input leakage and power. Its guidance supports reviewing those interactions; it does not assign a universal debounce interval to an HYR keypad.

Do not add a large capacitor to an ordinary CMOS input and assume that a clean-looking slow transition is acceptable. TI specifies input-transition restrictions for ordinary logic and explains that true Schmitt-trigger inputs use different rising and falling thresholds. Select and verify the actual receiver and conditioning circuit before transferring any threshold or timing assumption.

For the equipment check, capture the receiver-node waveform and the accepted logical event under agreed press and release conditions. Identify the firmware revision, scan interval, debounce behavior and any hold or repeat function. A waveform that eventually reaches the right voltage may still settle after the controller's decision point. Matrix scanning and multiple simultaneous keys require their own state analysis.

Five ways an apparently good sample can miss the requirement

Use these failure modes to locate the missing evidence. They are review scenarios, not reported HYR defects.

• Boundary mismatch: the report measures contact-only resistance while the equipment depends on the complete tail and connector path.
• Pressed-state miss: path resistance and the pull-up leave the receiver above its guaranteed low-level region.
• Released-state miss: leakage and a weak pull-up prevent a valid high level after the key is released.
• Timing miss: bounce or settling crosses the controller's sampling window even though the final DC level is acceptable.
• Uncontrolled retest: a changed fixture, actuation condition or measurement time makes two reports look different without isolating a part change.

Turn the budget into a supplier acceptance record

Issue one controlled requirement containing the circuit revision, measurement endpoints, test conditions and limit. Ask the supplier to state which path elements its result includes and how fixture contributions are handled. If the complete-system budget includes a customer connector or PCB route, identify who measures that remaining portion.

Record initial acceptance separately from any checks after cycling or environmental exposure. Preserve the agreed measurement conditions and decide whether acceptance uses an absolute limit, an allowed change or both. The electrical budget should explain that choice; a general endurance statement does not set it.

For an engineering review, send the relevant input schematic, controlled pin map and proposed inspection record through the secure RFQ. Request clarification of unresolved measurement boundaries before comparing prices. The OEM remains responsible for controller design and complete-equipment validation.