On this page
Begin with separate signal paths from each printed track to the receiver, then mark where their supplies, returns, connector bodies, mechanical carrier, substrate, ADC reference and software processing become common. Two visible tracks are not automatically independent. A common return drop can move both readings together, leaving their difference almost unchanged.
Key design decisions
- Draw the two channels and every shared node.
- Calculate how shared impedance moves both outputs.
- Distinguish sensor ground from receiver reference.
Draw the two channels and every shared node
Begin with separate signal paths from each printed track to the receiver, then mark where their supplies, returns, connector bodies, mechanical carrier, substrate, ADC reference and software processing become common. Two visible tracks are not automatically independent. A common return drop can move both readings together, leaving their difference almost unchanged. A shared shaft offset can likewise make the pair agree about the wrong position. The dependency map should name which faults create differential evidence and which require an independent observer. It must not present two channels as a functional-safety claim or assign diagnostic coverage without the system owner’s analysis.
Calculate how shared impedance moves both outputs
Model a common conductor or contact as Rcom carrying the sum of applicable channel and bias currents. The effective excitation becomes Vs minus Icom Rcom before each channel’s gain and offset are applied. As an illustration, a 20 mV rise on a shared return can shift two nominally 2.5 V outputs in the same direction while a channel-to-channel residual stays nearly constant. Whether the shift is equal depends on topology and receiver reference. Solve the actual network, including where sense grounds join, rather than subtracting a guessed common voltage from every node.
yi = Gi(x)[Vs - Icom Rcom] + Oi
- yi: reported output of channel i
- Gi(x): channel transfer at mechanical coordinate x
- Vs: source at the declared reference
- Icom Rcom: drop in the shared path
- Oi: channel-specific offset
Quasi-static circuit with explicitly located shared impedance; dynamic filtering and sampling need a time-domain extension.
Distinguish sensor ground from receiver reference
A voltage is meaningful only between named points. If both channels are measured relative to a receiver ground downstream of a connector, a shift between that ground and the card may be invisible to local subtraction. If ADC conversion shares the same moving reference as sensor excitation, some supply changes may cancel ratiometrically while other offsets remain. Document the source, return, shield, housing and ADC reference nodes. Avoid calling a conductor “ground” without locating its joins and load currents. This boundary decision determines which perturbation should be injected and where independent voltage observations belong.
Keep electrical correlation separate from shared mechanics
A common wiper carrier, shaft, bearing or card datum can move both contact positions together. The two outputs may remain perfectly correlated even though the physical input is displaced. Introduce an independent position reference when testing this class of fault. Conversely, a carrier skew may affect the two footprints differently and create a residual. Record paired mechanical coordinates where accessible, and compare them with raw electrical values before normalization. The card can support separated tracks and terminals, but the mechanism and diagnostic architecture determine whether those features yield useful independence.
Challenge common and channel-specific paths separately
Hold mechanical position fixed while varying source within the supplied range, adding a controlled return drop or loading one channel according to the approved circuit. Then keep electrical conditions stable and perturb mechanical position or carrier alignment. Sample both channels and reference nodes synchronously so a time skew is not mistaken for a differential fault. Use bounded, safe perturbations defined by the equipment owner; the article is not a universal fault-injection procedure. Reserve independent runs for validation after coefficients and plausibility bands have been selected.
Interpret agreement and disagreement with the dependency map
A residual excursion can reveal channel-specific gain, offset, contact or timing behavior, but it does not identify the cause automatically. A stable residual while both absolute values shift directs attention to common excitation, reference or mechanics. One channel changing alone focuses the electrical and contact path for that channel. Both channels changing with an independent position reference stable suggests a common electrical influence. Preserve absolute values, residual, supply, return and position together; a residual-only log discards the information needed to diagnose common causes.
| Observation | Supported interpretation | What is not yet excluded |
|---|---|---|
| Both outputs move, residual stable | Common influence is plausible | Shared mechanical error versus shared electrical reference |
| One output moves | Channel-specific path is implicated | Receiver input versus track or contact |
| Residual changes during motion only | Timing or dynamic contact may contribute | Sampling skew versus real geometry difference |
| Pair agrees, independent position disagrees | Correlation did not detect common position error | Mechanism datum or shared carrier shift |
Release dependencies, not a generic redundancy label
The circuit interface record should identify separate and common conductors, terminal assignments, source and return joins, receiver input impedances, reference nodes, sample alignment and the expected mathematical relationship between channels. The mechanical record should identify the common carrier and independent position observation used for validation. Define the coordinate range over which any plausibility equation applies, including saturation and endpoint behavior. Changes to connector pinning, grounding, ADC reference, filtering, carrier, shaft or curve coefficients trigger review.
Provide the evidence needed for a system-owned plausibility decision
An RFQ package should include both raw transfer curves, card schematic, terminal map, shared supply and ground budget, receiver circuit, sampling timing and mechanical-reference data. State which common causes the surrounding system must detect and which card-level differences are intended to support that detection. Supply owner-defined residual limits and actions separately from component acceptance. Thick-film track layout can be evaluated for geometry, separation and electrical behavior under the reviewed conditions; it cannot by itself establish diagnostic independence, safety integrity or complete accelerator or throttle compliance.
Keep diagnostic thresholds outside the manufacturing drawing when they depend on controller timing or system reaction. The drawing may carry the channel relationship and measurable electrical limits, while the receiver specification owns filtering, confirmation time, fallback behavior and event handling. This separation prevents a component revision from silently changing system logic.
Send the shared supply and ground effects on apparent redundancy inputs
Provide the dimensions, circuit and validation registers needed to evaluation common excitation and return network against two reported channel voltages and their correlation.
- two transfer curves, circuit schematic, supply and ground budgets, connector topology, sampling timing and system-owner plausibility criteria
- Relevant tolerance limits and raw observations for coordinate reference drop, shared return source/load impedance and connector resistance.
- Definition of independently observed supply and ground at the sensor boundary, motion or exposure progression, fixture and receiver receiver loading.
- Allowed functional error, validation ownership, unresolved assumptions and necessary technical basis format.
The drawing-upload form loads as you reach this section.