AMC1300 measures shunt current with nominal ±250 mV differential input and fixed gain 8.2. The shunt converts current to voltage; the isolated amplifier sends a differential signal to the low side. Errors can arise from resistance, sensing location, drift, supplies, or ADC reception; complete-product current readings alone cannot establish isolator damage.
Calculate measurement range and thermal loss simultaneously
Shunt voltage is VSH＝I×RSH and dissipation PSH＝IRMS²×RSH. At assumed 10 A with 5 mΩ, voltage is 50 mV, DC dissipation 0.5 W, and ideal differential output approximately 0.41 V. This illustrates conversion; check maximum continuous, peak, reverse, and fault-pulse currents against input range instead of equating the normal point with full scale.
Higher shunt resistance increases signal but also loss and heating. Lower resistance reduces heat but increases relative input-offset and copper-drop contributions. An assumed 1 mV input-referred error with 5 mΩ gives 0.2 A current error. This is error propagation, not guaranteed AMC1300 offset.
| Selection Input | Inspection Item | Output |
|---|---|---|
| Current Range | Forward and reverse, continuous, pulse, and fault | Normal and fault shunt voltages |
| Shunt resistor | Resistance tolerance, temperature drift, power, and pulse | Hot resistance and error budget |
| Sampling chain | Gain, input offset, and filtering | ADC differential and common-mode range at ends |
| Mechanical Environment | Copper layer, connection, airflow, and adjacent heat sources | Temperature Rise and Layout Constraints |
Sampling line drawn independently from shunt terminal
Separate high-current connections from small-signal sensing. Route a short pair from specified Kelvin terminals or near the resistor's effective endpoints into the input network, without load current in sensing copper. Remote sensing along large copper paths includes extra copper drop as shunt signal, potentially invalidating calibration as temperature changes.
Keep input filtering symmetric with compact local loops away from switching nodes and drive circuits. Evaluate differential/common-mode behavior; unequal resistors/capacitors can convert common mode to differential error. Soldering and copper layout affect shunt cooling, thermoelectric voltages, and hot resistance; layout cannot be treated purely as an accuracy postprocessing issue.
Differential outputs must also meet ADC common-mode conditions
Differential output equals input times gain, but individual OUTP/OUTN voltages relative to low-side ground must also fit receiver limits. Single-ended ADCs need suitable conversion stages, with resistor ratios, amplifier noise, and output swing budgeted. Reading one output and subtracting fixed bias can sacrifice common-mode rejection and add drift.
AMC1300 and AMC1300B have different error grades; record maximum, typical, and drift conditions separately. The B version cannot fix wrong shunt wiring or saturated outputs. Identify coefficients as input-referred or current engineering units to avoid applying gain 8.2 or shunt ratios twice in software.
Keep supplies and instruments referenced to their respective isolation sides
Decouple high/low sides to their own grounds and check startup sequence, missing supplies, and repowering. The isolator does not provide high-side power automatically; isolated-supply ripple/parasitic capacitance can affect common-mode transients. Instrument connections must not accidentally short isolated grounds; verify with the approved test structure.
PCB pads, vias, contamination, and mounting hardware affect isolation paths; package parameters do not certify the whole board. Layout inputs need actual potentials, insulation requirements, materials, and environment. Product requirements determine spacing/tests; ±250 mV input range cannot establish safe high-side potential.
Verify separately under cold, hot, and dynamic conditions
| phase | Experiment | Record |
|---|---|---|
| Zero Point | No load or known zero input under specified connections | Output, dual power supplies, temperature, and bias |
| Ratio | Multiple current points and independent reference comparison | Shunt drop, differential output, and ADC codes |
| Thermal stability | Fixed current running from cold to stable | Resistor temperature, reading drift, and supply changes |
| Dynamic | Controlled load step and reverse change | Bandwidth, delay, overshoot, and saturation recovery |
If shunt drop changes, check actual current and hot shunt behavior first. If the shunt is stable but output drifts, check isolator supply/temperature. If analog output is stable but codes change, check ADC reference, range, and software. Change one condition at a time and retain raw/calibrated data so filtering does not temporarily conceal drift.
Make shunt mounting and conversion chains reviewable
For four-terminal shunts, connect current and sense terminals according to package definitions, matching pads/stencils to the actual part. Equal-resistance parts from different manufacturers can have different sensing points, power derating, and thermal resistance; resistance alone cannot approve substitution. For two-terminal replacements, reassess potential gradients across high-current pads.
Screws, links, and busbars in field wiring heat up and can warm the shunt. A distant board sensor does not represent the resistor body; establish measurement-to-actual-temperature relationships before compensation models. Compare cold and hot errors separately instead of zeroing only at startup.
Convert ADC codes to low-side differential voltage, divide by 8.2 for input voltage, then divide by shunt resistance for current. Include receiver-stage scale/bias if present. Verify signs with known positive, zero, and negative inputs, retaining units at each step to avoid thousandfold errors from millivolts/volts or milliohms/ohms.
During dynamic measurement, chain delay and filtering can postpone software detection of overcurrent. System tasks determine whether measurement and fast protection share a path. Good static accuracy does not establish arbitrary short-circuit response; response time and actual actuators need separate justification.
If noise tracks power-switch edges, compare common-mode transients, isolated supply, and input-loop coupling first. In an approved low-risk setup, hold differential input constant while toggling nearby activity and observe output disturbance. Do not hide transients with heavy low-pass filtering while claiming the original bandwidth.
Calibration deliverables should identify reference-current source, shunt lot, sensing layout, temperature, both supplies, and receiver configuration. After shunt or copper-path changes, do not reuse old scale/temperature coefficients unconditionally.
Can shunt temperature drift be ignored after room-temperature calibration?
No. Self-heating and nearby components change resistance, and shunt temperature may differ from ambient. Verify actual current and mounting, recording compensation-model applicability where needed. For high-impedance divided-voltage sensing, see AMC1311 input structure; its 2 V voltage input is not interchangeable with this page's low-voltage differential current input.
Review hot behavior and sensing locations for shunt substitutions
Equal 5 mΩ parts may have different terminals, derating, and heat distribution. Substitution inputs should include at least package, four-terminal definitions, temperature drift, pulses, and mounting; outputs are conversion and temperature-rise reviews. Similar room-temperature multimeter resistance alone cannot approve substitution.
Record zero current and known positive/negative points together, confirming the actual high-current path is unchanged. If zero moves as the load warms, compare thermoelectric voltages, shunt temperature, and receiver offset before adding a fixed correction. Establish application accuracy only after separate cold, hot, and dynamic verification.
Source material and applicable scope
- Texas Instruments, “AMC1300 Precision, ±250-mV Input, Reinforced Isolated Amplifier”; Rev. D; checked on 2026-10-17; relevant sections: Electrical Characteristics, Analog Input, Application and Implementation, Power Supply, Layout. Supports ±250 mV differential input, gain, and shunt-sensing application conditions.
- Texas Instruments, “AMC1300 product information”; online page with no fixed revision identified; checked on 2026-10-17; relevant sections: Features and Description. Supports gain/error distinctions between the basic and B versions.
