On this page
Two resistive tracks do not create redundancy merely by occupying the same card. Their supply, ground, wiper mechanics, connector pins and receiver circuits determine which failures are common and which are observable. The controller may compare ratios, slopes, sums or bounded envelopes, but those rules require two clearly defined transfer functions under loaded conditions. Interface work assigns electrical dependencies and preserves raw measurements; the OEM remains responsible for diagnostics, reaction and functional safety.
System boundary
Two drawing-defined resistive tracks and contacts, their terminals, connector, harness, excitation, receiver channels and raw plausibility inputs. Pedal mechanics, propulsion request, diagnostic coverage, fault reaction and functional-safety compliance remain entirely with the system developer.
System integration decisions
- Declare each channel law and allowed relationship over travel.
- Identify shared elements that defeat assumed independence.
- Define receiver loading and diagnostic bias for both channels.
Specify two transfer laws in physical coordinates
For each channel, provide output versus traceable pedal or wiper position, approach direction and temperature. State whether slopes are same, opposite or intentionally scaled. Include endpoints, transition regions and receiver reference. Do not define plausibility only at a handful of nominal voltages. Tolerance bands must remain interpretable over continuous travel and under the actual input loads.
Calculate a bounded channel relationship
One possible monitor uses a calibrated linear combination, but coefficients belong to the system design.
G(x)=V_2(x)-a V_1(x)-b; |G(x)| <= T(x)
- V_1 and V_2 are time-aligned loaded channel outputs.
- a and b define the approved relationship for the chosen architecture.
- T(x) is a position-dependent tolerance including agreed contributors.
Both channels are valid and synchronized; the expression does not establish diagnostic coverage or safe response.
Illustrate how sampling delay consumes relationship margin
If channel slope is 2 V per unit travel and pedal speed is 0.6 unit per second, a 5 ms sampling offset represents about 6 mV difference before electrical tolerances. The example shows timing propagation only. Actual slopes, speeds, filtering and thresholds are system inputs. A static bench comparison cannot reveal this dynamic contribution.
Map common and independent elements honestly
List shared substrate, mounting, wiper carrier, supply, return, connector shell, harness bundle, reference and software task. Then list truly separate tracks, contacts, pins, inputs and acquisition resources. Physical separation on the ceramic does not make the common mechanics independent. The safety architect determines whether the architecture and diagnostics meet required coverage; the card supplier should not claim redundancy performance.
Evaluate loading and bias for each ECU state
Calculate normal acquisition, sleep, wake, open-load and short-diagnostic states. A pull-up on one input can couple through contamination, protection or shared card features. Input impedance variation can change channel slope differently. Record connector-pin voltages and currents at defined travel positions. Preserve raw channel data before scaling or plausibility logic.
Keep dual-channel allocations explicit
A dependency table prevents an assumed independent feature from hiding inside a shared assembly.
| Element | Shared or separate | Decision owner |
|---|---|---|
| Pedal and carrier | Typically shared mechanics | Pedal designer |
| Printed tracks | Separate functions on one substrate | Card drawing authority |
| Supply and return | Architecture-specific | Electrical architect |
| ADC, timing and logic | Controller-specific | ECU and safety owner |
Study signatures without promising diagnostic coverage
A single-channel open may be visible as a rail value; a shared-ground shift can move both channels; a contact bridge may create a localized relationship error; mechanical carrier slip can preserve a plausible channel relationship while reporting the wrong pedal position. These examples show why plausibility is not absolute position truth. The system team performs formal hazard, fault and diagnostic analysis.
Acquire synchronized mechanics and raw channels
Record reference travel, both contact positions where accessible, supplies, returns and raw ECU inputs during slow sweeps, reversals, dynamic motion and endpoint dwell. Include specified electrical, mechanical and environmental corners. Inject faults only under an authorized system plan. ChipSimple can support card drawing review and card-level measurement; it does not validate propulsion control or functional safety.
Release channel relationships with configuration identifiers
Control track artwork, wipers, carrier, connector, harness, ECU circuit, calibration coefficients, sampling and threshold revisions. Reopen after any common or independent element changes. Archive raw tolerance envelopes and assumptions so a software update does not silently rely on obsolete card or mechanism behavior. A passing nominal correlation is not evidence of fault tolerance.
Tolerance analysis should preserve channel covariance. Track dimensions printed on one substrate can move together, while separate contacts and input circuits can vary independently. Worst-case independent stacking may be excessively conservative for one decision, whereas assuming perfect correlation can hide a hazardous corner. The system analyst chooses a justified dependence model and validates it with paired data. Plot the channel relationship against physical position, temperature, supply and direction instead of reporting one correlation coefficient. At endpoints, ensure both footprints remain on allowed surfaces even if their electrical relationship appears plausible. Mechanical coverage and electronic plausibility are complementary checks, not substitutes.
Electrical tests should include reference-channel monitoring so a shared supply disturbance is not misclassified as simultaneous card movement. Where channels use separate references, capture both. Analyze relationship margin at low, middle and high travel and through reversals. Retain position alignment and acquisition timestamps. Component-level resistance checks remain useful for manufacturing control, but system plausibility validation must occur at the controller boundary with the intended harness, contact mechanics and software configuration.
Change control should include connector pin assignments and diagnostic software calibration. A harness revision that joins returns or changes shielding can alter common-cause behavior without changing either printed track.
For environmental validation, examine whether condensation, contamination or differential expansion can couple nominally separate tracks. Preserve channel-specific and shared-interface observations so corrective action targets the actual dependency.
Any modification to channel routing, contact geometry or acquisition requires renewed paired verification at the defined controller boundary.
Provide both channel laws and the complete dependency map
Dual-channel review needs raw electrical and mechanical definitions.
- Pedal travel, carrier, wiper footprints, forces and stop allocation.
- Both track functions, tolerances, terminals and substrate datums.
- Supply, return, harness, connector and receiver schematics.
- Sampling, plausibility method, fault assumptions and validation ownership.
The drawing-upload form loads as you reach this section.