On this page
A smaller in-phase reading does not necessarily mean that a ceramic sensor or its printed resistor network has lost sensitivity. A cable, filter or excitation reference can rotate the signal phase while its amplitude remains unchanged. Retaining both synchronous components makes that distinction measurable before a resistor trim conceals the underlying change.
System boundary
An AC-excited sensor, ceramic resistor signal-conditioning network, amplifier, reference and synchronous detector. This review does not establish a sensor accuracy or company lock-in measurement capability.
Integration interfaces
| Interface | Required input | Thick film role | Validation owner |
|---|---|---|---|
| Sensor and printed network to detector input | Excitation frequency, gain, filtering and input amplitude range. | Printed resistors participate in gain and impedance paths before demodulation. | Analog designer. |
| Excitation source to synchronous reference | Reference frequency, phase origin and timing continuity. | A resistor change cannot repair a lost or inconsistent reference. | Acquisition engineer. |
| Detector components to reported measurement | X, Y, normalization, settling and calibration records. | Network gain is evaluated separately from phase projection. | System calibration owner. |
Integration risks
| Risk | Control or verification | Validation owner |
|---|---|---|
| A phase shift is trimmed out as an apparent sensitivity error. | Compare amplitude and phase reconstructed from both components. | Analog engineer. |
| Peak and RMS conventions are mixed between instruments. | Verify scaling using a known coherent input before calibration. | Measurement owner. |
| A detector returns a stable-looking number while its front end clips. | Inspect the pre-demodulation waveform and valid input range. | Acquisition engineer. |
System integration decisions
- Keep in-phase and quadrature data together rather than storing only one projected reading.
- Declare peak or RMS normalization and the reference phase convention.
- Separate gain changes, phase rotation and loss of coherent reference before adjusting the network.
Recognize what the in-phase output measures
A synchronous detector compares a signal with a periodic reference. Its in-phase output is a projection onto that reference direction, not an unconditional amplitude measurement. This distinction matters when an AC-excited bridge or other sensor feeds an amplifier whose resistor network sets the gain but whose complete input path also contains capacitance and delay.
First identify the physical measurement boundary. Is the requested quantity the sensor voltage at its terminals, the amplified voltage at the detector, or a calibrated physical input? Record the excitation and reference connection alongside it. Moving the reference pickoff or changing a cable can change phase without changing the printed resistor values. Those configurations must not share an unexplained calibration adjustment.
Declare a two-component normalization
Use a declared convention in which a coherent sinusoid has RMS amplitude R and relative phase phi, with X = R cos(phi) and Y = R sin(phi). Then R = sqrt(X squared + Y squared), and phi = atan2(Y, X). The two-argument angle function retains the quadrant, unlike a simple ratio followed by an ordinary arctangent.
These expressions assume that both component channels have been scaled consistently. A hand-built multiplier implementation can have another factor before normalization, and peak amplitude differs from RMS amplitude for a sinusoid. Check the actual implementation rather than copying a displayed value from another instrument. Preserve the sign convention for the quadrature reference so that positive phase means the same thing in software and the test record.
X = R cos(phi); Y = R sin(phi); R = sqrt(X^2 + Y^2); phi = atan2(Y, X)
- X and Y: normalized in-phase and quadrature components in volts RMS.
- R: coherent sinusoidal amplitude in volts RMS.
- phi: signal phase relative to the declared reference, in radians in the formula.
Equal-frequency coherent sinusoid, settled detector, consistent channel scaling and a linear unclipped front end. This is not total broadband RMS.
Test phase rotation before changing resistor gain
Consider an assumed detector input of 10 millivolts RMS. At zero relative phase, X is 10 millivolts and Y is zero. At 30 degrees, the same amplitude gives X approximately 8.660 millivolts and Y 5.000 millivolts. A system retaining only X would report a 13.40 percent reduction, even though the reconstructed amplitude remains 10 millivolts.
If an engineer increases resistor gain to restore X to 10 millivolts in that second condition, the necessary factor is about 1.1547. The actual amplitude would rise to about 11.547 millivolts. When phase returns to zero, the error becomes visible as excessive sensitivity. This assumed example is a diagnostic counterexample, not an expected drift or acceptable error for any supplied resistor network. It shows why the phase state belongs in the calibration record.
Compare component motion in controlled experiments
Apply a known coherent input through the same path used by the sensor, within the permitted electrical range. Change the reference phase deliberately while holding input amplitude fixed. The component pair should rotate with nearly constant reconstructed amplitude after settling. Then hold phase fixed and change the input amplitude; the pair should scale along the same direction.
These two experiments test different behaviors. If the amplitude varies strongly during a phase rotation, investigate channel scaling, quadrature error, offset or nonlinear behavior before attributing the result to the ceramic network. Include an input-off or appropriate zero condition, but retain the raw components: subtracting an unexplained magnitude baseline is not equivalent to correcting two component offsets.
| Observed pattern | First interpretation to test | Useful controlled comparison |
|---|---|---|
| X decreases while Y increases and R stays similar | Phase rotation | Change reference phase at fixed input |
| X and Y scale together at stable angle | Amplitude or gain change | Apply two known coherent amplitudes |
| Component pair rotates continuously | Frequency difference or reference timing issue | Verify reference coherence and timestamps |
| R changes during an imposed phase sweep | Channel mismatch or nonlinear response | Check raw components and pre-detector waveform |
Do not treat frequency mismatch as a stable phase offset
When the input and reference frequencies differ, their relative phase changes with time. In a two-component record, the vector can rotate instead of remaining at a fixed angle. Averaging or narrow filtering can then reduce the reported components even though the input waveform remains present. A single late reading does not distinguish this from a gain loss.
Record reference lock, frequency source and acquisition timing with a sufficiently informative interval of X and Y. A reset in a waveform generator or a changed clock relationship can invalidate an earlier phase correction. Restore the specified reference relationship before recalibrating the resistor network. If the application intentionally measures an offset frequency, define the resulting beat and analysis method explicitly rather than applying a static projection formula.
Keep front-end validity outside the amplitude calculation
Reconstructing R cannot recover information lost through amplifier clipping, input protection conduction or inadequate analog bandwidth. Inspect the waveform ahead of demodulation and confirm its range during the largest intended sensor signal and disturbance. A coherent interference component at the reference frequency is not rejected merely because a lock-in method is used.
Also distinguish the measured fundamental from the complete waveform. Harmonics and broadband noise can contribute to an ordinary RMS measurement without contributing equally to the selected coherent component. If the excitation is nonsinusoidal or a switching detector is used, document which harmonics the method admits. Agreement between two numbers is meaningful only when both instruments measure the same quantity and boundary.
Allocate resistor-network changes to the demonstrated error
After establishing a valid detector, a stable phase-independent amplitude error can justify investigation of the analog gain path. Compare the actual resistor ratios, source loading and amplifier response at the operating frequency. A phase-only change instead directs attention toward impedance, filtering, timing and reference placement. Changing absolute resistance can itself alter a parasitic time constant while leaving the nominal DC ratio unchanged.
Keep the application frequency and connected components with the network drawing. A DC resistance acceptance result does not settle an AC signal-path question, but neither does an AC anomaly prove a defective printed resistor. Record the evidence that assigns the change to its stage so that the supplier receives a focused network requirement rather than a request to compensate an unidentified system error.
Provide a reproducible synchronous measurement record
For a ceramic resistor-network review, provide the terminal mapping, intended gain, excitation frequency, source impedance and detector boundary. Attach the coherent input used for verification, X and Y records, amplitude convention and phase origin. State any component offset correction and the settling rule applied after changes. Include the waveform evidence that establishes linear operation.
ChipSimple can review the drawing-defined printed network against the supplied electrical interface. The customer's system owner remains responsible for reference generation, detector implementation and complete sensor calibration. Repeat the phase-versus-amplitude comparison after changes to cable, filtering, amplifier, reference routing or firmware, rather than assuming an unchanged resistor drawing guarantees an unchanged measurement chain.
Review the ceramic network in its AC signal path
Send the actual excitation and detector interface with the resistor drawing.
- Terminal-level network and source/load impedances.
- Excitation frequency, amplitude and reference origin.
- Raw X and Y, phase and amplitude conventions.
- Pre-detector waveform and settling observations.
- Required measurement range and allocated network contribution.
The drawing-upload form loads as you reach this section.