On this page
A current-sensing resistor can produce an accurate reading of a current that it has already changed. Increasing its resistance improves the voltage available to the receiver, but also increases the electrical burden introduced into the load path. For a low-ohmic printed resistor interface, specify these as separate requirements: how well the instrument reads the connected circuit, and how closely that circuit still represents the original operating condition.
Key design decisions
- State whether the measurement targets the current with the sensing resistor installed or the current before insertion.
- Include every added series element in the disturbance budget, even when Kelvin leads exclude its voltage from the reading.
- Check the intersection of the burden limit and the receiver-error limit before selecting the resistor value.
Distinguish a built-in sensor from a temporary insertion
If a sensing resistor is part of the released circuit, its voltage drop belongs in the original supply and load design. The measurement then describes that intended circuit. If it is added temporarily to investigate a circuit that normally lacks it, the test creates a different electrical path. The target quantity must distinguish those cases before an accuracy percentage has a useful meaning.
A calibrated voltage receiver and accurately known resistance can determine the instrumented current very well. That calibration does not establish that the load voltage, current distribution or controller state remained unchanged. Retain the original and instrumented connection drawings, including the location of the current return, so the measurement report does not silently substitute one operating state for the other.
Calculate the disturbance for an explicitly resistive circuit
Consider a fixed ideal source voltage Vs feeding an ohmic load path of resistance R0. Before insertion the current is I0 equal to Vs divided by R0. Add total series resistance Ra, including the sensing element and newly introduced current leads or contacts. The current becomes Im equal to Vs divided by R0 plus Ra. The fractional reduction relative to I0 is Ra divided by R0 plus Ra.
For a maximum permitted fractional reduction epsilon, the total added resistance must not exceed epsilon R0 divided by one minus epsilon. This is an exact result for the stated fixed-voltage, fixed-resistance comparison. It is not a correction formula for an arbitrary regulated electronic load, switching converter or temperature-changing heater.
I0 = Vs/R0; Im = Vs/(R0 + Ra); (I0 - Im)/I0 = Ra/(R0 + Ra); Ra,max = epsilon R0/(1 - epsilon)
- Vs is the unchanged ideal source voltage in volts; R0 is the original series load-path resistance in ohms.
- Ra is all resistance added to the current path in ohms, not only the Kelvin-sensed element.
- epsilon is the allowed reduction relative to the original current, between zero and one.
Steady ohmic operation with unchanged source voltage and original load resistance. Reactive effects, active load regulation and thermal movement are excluded.
Keep the instrument reading and circuit error separate
Assume a 5 V source and an original 10 ohm load path. The original current is 0.5 A. A 0.10 ohm inserted sensing resistor with no other added resistance reduces current to approximately 0.495050 A. Its voltage is approximately 49.505 mV, while load voltage falls to approximately 4.95050 V. A perfect receiver reporting 0.495050 A is correct about the connected circuit, even though the insertion reduced the original current by approximately 0.9901 percent.
If another 0.05 ohm is added through fixture leads and contacts, the current becomes approximately 0.492611 A. The receiver still senses only the 0.10 ohm element and sees approximately 49.261 mV. The circuit disturbance has increased to approximately 1.4778 percent. Excluding those extra drops from the voltage measurement is useful, but does not remove their physical effect on the current.
Calculate the resistance needed for the receiver signal
Let the receiver's remaining input-referred offset magnitude be bounded by Voff after the intended zeroing procedure. For an installed current I, the corresponding current error magnitude is Voff divided by the sensed resistance Rsh. Allocating a fractional offset error e at the minimum installed current Imin requires Rsh to be at least Voff divided by e Imin.
This bound allocates offset only. Resistance calibration uncertainty, temperature change, receiver gain error, noise and common-mode effects require their own treatment. Do not spend the complete system error allowance on this one term. Use the current after insertion for the minimum-current condition; substituting a larger original current would make the signal estimate optimistic.
As an illustrative allocation, take a 20 microvolt offset bound, 0.5 percent allowed offset contribution and a minimum installed current of 0.10 A. The sensed resistance must be at least 0.040 ohm. The associated minimum signal is 4 mV. These are design inputs for the comparison, not specifications of a particular resistor or instrument.
Intersect the two limits rather than choosing by signal alone
For the original 10 ohm path, limiting current disturbance to 0.5 percent allows no more than approximately 0.050251 ohm total added resistance. If added leads and contacts account for 0.010 ohm, the sensing element is limited to approximately 0.040251 ohm. Combined with the 0.040 ohm signal lower bound, the remaining nominal interval is only about 0.000251 ohm wide, before resistance tolerance or temperature movement is allocated.
If those extra connections instead contribute 0.020 ohm, the upper bound for the element falls to approximately 0.030251 ohm. The two requirements are then incompatible. A more accurate resistor value alone cannot fix the missing interval. Reducing receiver offset, reducing insertion resistance outside the sensed element, relaxing a justified requirement or selecting a different measurement architecture changes the feasibility problem.
| Added unsensed resistance | Sensed-resistance upper bound from burden | Sensed-resistance lower bound from offset | Outcome before other allowances |
|---|---|---|---|
| 0 ohm | 0.050251 ohm | 0.040000 ohm | Nonempty nominal interval |
| 0.010 ohm | 0.040251 ohm | 0.040000 ohm | Very narrow nominal interval |
| 0.020 ohm | 0.030251 ohm | 0.040000 ohm | No feasible resistance value |
Draw two different resistance boundaries
The measurement boundary encloses the voltage-sensed region whose resistance converts voltage to current. The disturbance boundary encloses everything newly added to the original current path. A four-terminal connection can make the first boundary precise while leaving substantial resistance in the second. Record both instead of labeling the complete fixture simply as Kelvin connected.
This distinction also matters when moving the sensing element to a return conductor. The inserted voltage can change the load's local ground relative to other signal connections. A communication cable or additional return path may then bypass part of the intended current measurement. Verify the complete connectivity and receiver common-mode conditions; a convenient low-voltage sensing location is not automatically equivalent to the original grounding arrangement.
Stop the simple correction when the load changes state
A regulated load may maintain current while terminal voltage falls, until a compliance or operating limit is reached. A converter may draw more input current as its input voltage falls. A heater may change resistance after its temperature responds. None follows the fixed-R0 example automatically. Characterize the load response around the proposed insertion condition or use a model appropriate to its control state.
Even when the resistive model is applicable, recovering I0 from Im requires the original resistance and all added resistance to remain known. Multiplying the measured result by a calculated correction does not restore the load voltage during the test or undo a state transition that already occurred. Report corrected estimates separately from observed current, together with the model and its input uncertainty.
Verify the allocation at the actual measurement conditions
Measure the added path resistance between the defined disturbance boundaries and check the receiver residual near the relevant common-mode voltage. Repeat the comparison at the intended current extremes and thermal state. Where a lower-burden independent method is available, compare operating voltage and load behavior as well as current indication.
Preserve the chosen sensing resistance, added connection estimate, error allocations and before-and-after circuit state in the drawing review. This produces an actionable decision about the measurement interface. It does not substitute for evaluating the printed element's power density, terminal integrity or transient behavior, which remain separate parts of the component and application review.
Send the current-sensing and disturbance boundaries
Provide the original circuit and the proposed measurement insertion so signal requirements and load disturbance can be reviewed together.
- Original and instrumented current-path drawings, including return paths and exact Kelvin sense positions.
- Source behavior, load characteristic, minimum installed current and permitted operating-point disturbance.
- Receiver offset/noise/gain allocations and relevant common-mode and thermal conditions.
- Sensed resistance range plus every added lead, connector and contact contribution.
- Required waveform range, operating-state comparison records and the intended current-reporting definition.
The drawing-upload form loads as you reach this section.