Why an FSR Baseline Shifts After Assembly

When an FSR unloaded reading changes after installation, first find the step at which it changed. A compressed pad, housing rib, adhesive ridge or tightened bezel may apply a resting load. Cable strain, a connector fault or a readout-circuit change can produce a similar symptom. Do not label the shift as sensor preload—or remove it in firmware—until the mechanical stack and electrical path have been compared under the same measurement conditions.

A custom FSR pressure sensor should be reviewed as part of its mechanical stack. The same film can produce a different unloaded reading when the actuator, support surface, fasteners and cable route change. A bare-sensor measurement is useful, but it does not prove the installed zero condition. Tekscan’s sensor-integration guidance treats mechanical mounting and readout-circuit selection as separate integration checks. That general guidance does not supply specifications or acceptance limits for this Baoshengda part.

Confirm that the baseline moved in the installed stack

Use the same readout circuit, supply condition, connector and sampling method for every comparison. First record the unloaded value with the sensor lying flat and untouched. Then place it on the intended support without the actuator or cover. Add the actuator, close the housing, tighten the fasteners in the planned sequence and secure the cable. Log the unloaded value after each step.

This sequence shows where the shift begins. If the value changes when the cover is set in place but before screws are tightened, the actuator or foam may already be touching the active zone. A shift appearing only after final screw torque points to support bending or a compressed sealing lip. A shift appearing when the cable is clipped down points to strain at the sensor transition.

Let the assembly rest for the same interval before each reading. FSR response can depend on load history and dwell. Comparing an immediate reading from one step with a settled reading from another can hide the actual cause.

Separate mechanical preload from electrical offset

A zero shift does not identify its own cause. Use one controlled change at a time. Keep the circuit connected while removing the mechanical stack, then keep the mechanical stack fixed while checking the cable, connector and electronics.

For the mechanical branch, loosen the fasteners in reverse order and watch whether the baseline returns. Remove the actuator or foam without moving the cable. Place a thin witness film or removable marking layer over the sensor to see whether the unloaded cover leaves a contact mark. Check for adhesive squeeze-out, a raised liner edge, a housing rib, a screw boss or debris under the sensor.

For the electrical branch, leave the installed stack untouched. Record the raw output, supply, reference or feedback resistor, sampling settings and connector position before reseating anything. Then inspect terminal orientation, reseat the connector and move the cable only within its planned service loop. A step that follows cable movement calls for a termination or strain check; it does not by itself prove that the sensing film is faulty.

If the equipment maker’s circuit permits a safe substitution, its electrical engineer can disconnect the sensor under the approved procedure and apply a known reference element at the sensor-side connector. Compare the readout before and after the substitution without changing firmware, supply or sampling. A stable reference with a shifting installed-sensor signal directs attention back toward the sensor and its mounted contact stack. A shifting reference directs attention to the connector-to-board path or electronics. If the reference is introduced at the board instead, the cable has been bypassed; note that difference in the record. Do not interpret a reference-resistor value as a calibrated force.

Do not compensate both branches at once. Changing a firmware offset while also trimming foam makes the next sample hard to diagnose because the evidence no longer shows which change worked.

Check actuator footprint, support and hard-stop geometry

An FSR responds to the pressure delivered through the part above it. The actuator may be a molded button, foam pad, rubber feature, plastic rib, cushion layer or compliant skin. Its footprint should land within the intended sensing zone and remain controlled when the housing tolerance moves.

A small hard actuator can create a strong local signal, but it is sensitive to lateral misalignment and may concentrate load. A wider soft pad spreads the load and tolerates position variation, but it can leave residual compression or bridge adjacent zones. If a foam layer is used, record its nominal thickness, compressed thickness and contact area in the section drawing instead of calling it only a pad.

The rear support matters just as much. A gap or flexible wall under the sensor can bow when the enclosure is closed. A raised adhesive seam can create a local high point. A hard stop should limit excess travel without pressing the active zone at rest. These are installed-stack decisions, so the sensor supplier cannot infer them from the film outline alone.

Keep cable restraint away from the sensing transition

The film-to-cable transition needs a stable path into the enclosure. A tie, clip or tight connector pocket can pull on the strip and change how it lies against the support. The effect may appear as a baseline shift even though no external object is touching the sensing zone.

Mark the cable exit, first bend, strain-relief point and connector envelope on the assembly drawing. Leave enough service loop for connection without pushing the cable back toward the active area. Confirm that the cable does not cross a moving cover, sharp edge or screw path. During the staged test, secure the cable only after the mechanical zero has been recorded to isolate its effect.

If the design uses several sensing zones, log every channel while changing one cable or housing condition. A shared shift is a clue to a common supply, connector, cable or housing condition, not a diagnosis by itself. A shift isolated to one zone calls for a closer check of local actuator contact, support flatness and its conductor path. Preserve every channel’s raw reading before making a threshold change.

Approve a mounted zero before setting thresholds

A useful sample test contains both unloaded and loaded evidence. After the installed baseline stabilizes, apply the intended actuator at the agreed locations and force levels or fixture steps. Release the load and confirm that the reading returns to the accepted unloaded band. Repeat the sequence after opening and refitting the housing.

Record at least these states:

• bare sensor on a flat support;
• sensor placed in the fixture without the actuator;
• actuator or foam added without final fastener load;
• housing closed and fasteners tightened in sequence;
• cable restrained and connector fully mated;
• light, normal and high application loads using the intended contact geometry;
• unloaded recovery after a held load and after housing refit.

An FSR is normally used for relative contact, threshold or trend information unless the complete mechanical and electrical system has its own calibration plan. Do not turn one bench resistance into a universal force claim. The equipment team owns the readout circuit, firmware filtering, threshold margin, safety logic and final validation. Its approval record should name the sensor and housing revisions, the unloaded dwell interval, the raw-output method and the loaded/released fixture. Accept the mounted zero only if the OEM’s own limits hold across the agreed assembly and refit sequence; this page does not set those limits.

Send an evidence package, not only a new threshold

For a useful troubleshooting review, send the sensor drawing, active-zone map, section stack, actuator dimensions, foam or cover material, support surface, adhesive layout, fastener positions and tightening sequence. Include the cable route, connector, supply/readout circuit and a short table showing the unloaded value after each assembly step. Add the measurement units or ADC scale, supply and reference settings, dwell interval, sample IDs, the point at which any reference element was substituted, and the result before firmware correction.

Photos should show the sensor before the cover is installed and the finished cable restraint. If one zone shifts, mark that zone on both the sensor image and housing drawing. If all zones shift together, include the common connector and circuit conditions before assuming the printed sensors changed together.

Use the Request Quote page to send that evidence with the target quantity and sample plan. The next decision should be concrete: remove a resting load, change one support or restraint feature, or reproduce the electrical offset under a fixed mechanical stack before a revised threshold is approved.