Engineering diagnostic independence in a dual-track resistive circuit

System boundary

Measured mechanical input through shared and independent motion transfer, two resistive paths, terminals, wiring, bias networks, acquisition channels and comparison logic

Partition faults before drawing the schematic

Create a boundary diagram that marks every item as common, separated or conditionally independent. The external mechanism may drive both contacts while the tracks, conductors and input channels are separate. Alternatively, the same wiper carrier, substrate and connector may remain common. These distinctions determine which failures a second signal can expose.

List physical faults, electrical faults and data-processing faults independently. Include jam, backlash, loss of contact force, track discontinuity, conductor bridge, terminal open, pin-to-pin short, supply change, ground shift, converter saturation, stale data and incorrect calibration. The system owner decides relevance and consequence; the circuit design supplies observable behavior.

Choose channel relationships for diagnosability

The two channels may rise together, move in opposite directions or use unequal slopes and offsets. Each relationship has implications for wiring faults, range use and comparison logic. Complementary outputs can make some common shifts more visible, while identical outputs can simplify direct comparison. Neither choice is universally superior.

Define valid channel envelopes across position, temperature, supply and tolerance. Avoid a single constant mismatch limit when the expected difference changes with travel. Preserve sufficient electrical separation between valid behavior and detectable faults, including at endpoints where clipping or dead bands can reduce diagnostic margin.

Calculate a position-dependent comparison residual

Represent the expected relation between channels with a stated model. For a local linear region, use e=y2-(a y1+b), where a and b are derived from the released channel relationship rather than fitted silently to each unit. Compare the residual with bounds that include allowed component, acquisition and environmental effects.

As an illustrative calculation, if the released relation is y2=4.80-0.75y1 volts and y1 is 2.00 volts, the expected y2 is 3.30 volts. A measured 3.22 volts gives e=-0.08 volt. Whether this is acceptable depends on the defined regional bound and uncertainty; the example is not a product limit or diagnostic threshold.

Expose common-cause paths explicitly

A shared shaft can jam both contacts while their electrical values remain mutually plausible. A shared return can move both inputs together. Adjacent connector pins can bridge. Contamination can spread across both tracks. Software can apply the same incorrect scale to each channel. These cases require additional observation, physical separation, independent references or a bounded argument that the common cause is controlled elsewhere.

Record separation distances, conductor routing, terminal assignment, supply and ground ownership, ADC allocation and data paths. Where independence is not practical, label the common element instead of implying it has been removed. Functional-safety classification and regulatory acceptance are outside this page and remain with the integrator.

Define synchronization and diagnostic state

Channel comparison is meaningful only when samples represent the same mechanical state. State sample timing, multiplexing delay, filter delay, update rate and time alignment. During fast motion, a small delay can appear as disagreement even when both tracks are correct. During a jam, perfectly aligned channels can agree on the wrong position.

Specify startup, calibration, normal, degraded and fault states. Define debounce, confirmation time, recovery and latching behavior. Retain raw channel values and the diagnostic decision so later investigation can distinguish a physical disagreement from processing or communication behavior.

Use fault signatures rather than a single mismatch flag

A residual outside its envelope is an observation, not a root cause. Review absolute channel ranges, slopes, supply, reference, contact continuity and mechanical position. Inject one controlled change at a time and identify the earliest node that departs from expectation.

The following matrix organizes investigation. Actual diagnostic coverage must be demonstrated against the application fault set and must not be inferred from the presence of two channels.

Account for detection claims by fault

Maintain a fault list with cause, affected channels, expected signature, detection method, detection time, system response, evidence and owner. A percentage without a defined denominator is not useful. If coverage is calculated, identify whether the denominator counts fault modes, weighted risk cases or test injections.

Separate detected, safely tolerated, externally controlled and unresolved cases. Do not count a fault twice because it triggers two diagnostic flags. Do not claim that an electrical comparison detects loss of mechanical coupling unless an independent mechanical reference is present.

Validate normal envelopes and injected faults

First characterize both channels over travel, direction, speed, temperature and supply conditions with synchronized mechanical reference. Then inject representative opens, shorts, resistive connections, supply shifts, reference shifts and timing offsets. Mechanical fault work should include bounded jam, backlash or contact-load conditions where safe and feasible.

For every injection, record the injection point, expected channel behavior, observed raw values, diagnostic transition, latency and recovery. Freeze pass criteria before the run. Prototype results demonstrate only the tested configuration and cannot establish lifetime, production capability or safety approval.

Control revisions across both paths

Changes to track curves, wipers, substrate layout, connector pinout, shared return, protection, ADC configuration or comparison software can alter independence and detection margin. Route such changes through a common interface record rather than separate drawing approvals that never reassemble the architecture.

Revisit the common-cause analysis when packaging, mounting or harness routing changes. Preserve diagnostic test vectors so revised hardware and software can be compared with the same fault intent. Residual system risk and fallback behavior remain integrator decisions.

RFQ inputs for a dual-track circuit

Submit mechanical travel and coupling, proposed track relationships, resistance or voltage ranges, excitation, return, terminals, connector and acquisition topology. Mark every shared and separated feature. Provide temperature, vibration, contamination and motion duty.

Include the system fault list, diagnostic objectives, timing, allowed envelope, response state and validation ownership. Identify any required separation, pin-fault assumption or independent reference. Unresolved safety or regulatory requirements must be handled through the responsible system process, not assumed from component geometry.