Differential & Instrumentation Voltage Sensing with INAs
← Back to: Current Sensing & Power / Energy Measurement
This page shows how to use differential and instrumentation amplifiers to measure small voltage differences reliably on noisy, high or remote nodes. It focuses on picking the right topology, checking range and accuracy, and applying filters and layout rules so your ADC always sees a clean, trustworthy signal.
Role & Use-Cases for Differential / INA Voltage Sense
When you need to measure a small voltage difference in the middle of noisy, remote or high-common-mode environments, it is time to use a differential amplifier or instrumentation amplifier. These front ends trade extra circuitry for strong common-mode rejection, predictable gain accuracy and reduced sensitivity to ground noise, cable resistance and shared return paths across the power system.
Within the current sensing and energy measurement stack, a differential / INA front end can watch bus voltage, shunt drops, remote loads and phase nodes with one consistent architecture. Deep topics such as zero-drift behaviour, ultra-fast protection and digital energy accumulation are covered in dedicated pages; here the focus stays on the analog differential voltage-sense building block.
Bus Voltage + Shunt Drop
One differential / INA stage can sense both the bus rail and the shunt drop, feeding ADC channels that reconstruct voltage and current. This is common in DC buses where space and channels are limited but combined power and health monitoring is still required.
Remote Load & Kelvin Sense
Remote sense leads bring back the true load voltage at the far end of long, resistive traces or cables. A differential / INA front end measures this voltage and reports it to a regulator or controller so line-loss can be compensated without over-stressing the source.
Motor Phases & Battery Taps
Instrumentation amplifiers can monitor motor phase voltages, inverter legs and battery-stack tap points where common-mode levels are high and edges are fast. Their wide input range and strong CMRR keep the small differential information usable for control and diagnostics.
Rail Comparison & Current Sharing
By measuring small voltage differences between rails, shunts or sense nodes, differential front ends can highlight imbalances and support current-sharing or droop-monitoring schemes. The same block can drive slow supervisory thresholds or feed more advanced digital control loops.
Classic Differential & Instrumentation Amplifier Topologies
Differential and instrumentation amplifiers cover a spectrum from simple single-op-amp resistor networks to three-op-amp INAs and fully differential amplifiers that drive high-resolution ADCs. Choosing between them is mostly a trade-off between complexity, cost, input impedance, CMRR and bandwidth. This section sketches the classic topologies before later sections quantify noise, accuracy and layout behaviour.
3-Op-Amp Instrumentation Amplifier
The classic three-op-amp INA splits the job into two stages: a pair of input amplifiers that buffer and amplify each input, followed by a differential amplifier that removes common-mode voltage. A single gain resistor typically controls the overall gain, while the internal resistor network preserves matching and CMRR. This structure supports very high input impedance and clean, predictable gain for precision sensing.
Three-op-amp INAs are the default choice when you connect to high-impedance sources, long sensor leads or critical measurement points that must remain accurate over temperature and lifetime. For ultra-low drift or functionally safe current sensing, see the dedicated zero-drift and high-side current-sense pages where the same concept is extended with chopper techniques and protection features.
Single-Diff Amplifier with Resistor Network
A single-op-amp differential amplifier uses four resistors around a general-purpose op amp to subtract and scale two input voltages. When the resistor ratios are tightly matched and the PCB layout is symmetric, it can achieve respectable CMRR at very low cost. However, the input impedance is lower than that of a true INA and the performance is far more sensitive to resistor tolerance and parasitics.
This topology is well suited to short on-board Kelvin sense connections, auxiliary test nodes or cost-constrained channels where the environment is controlled. In demanding, high-noise or high-temperature systems, it is usually safer to reserve discrete differential stages for secondary measurements and keep primary functional monitoring on dedicated INAs.
Fully Differential Amplifier Front End
A fully differential amplifier accepts a differential input and delivers a controlled differential output with a defined common-mode level, often set via a VOCM pin. It bridges the gap between sensing and data conversion by driving high-resolution, high-speed ADCs with matched differential signals and enough bandwidth to keep distortion and settling errors under control.
FDA front ends are most attractive when you need to combine precise differential sensing with wide bandwidth, such as in multi-kilohertz or megahertz energy measurement, motor control, digital power and industrial data-acquisition nodes. ADC-specific topics such as jitter, SFDR and anti-alias filtering are handled in ADC and clocking pages; here the FDA is treated purely as the analog front-end block.
Key Selection Metrics & Trade-offs
Choosing a differential or instrumentation amplifier front end is mostly about how much common-mode you must reject, how accurately you must scale the useful differential signal and how much noise and bandwidth the system can tolerate. This section highlights the metrics that matter most in real power and current sensing paths, and how they trade off against each other.
CMRR & Gain Accuracy
Common-mode rejection ratio describes how much of the shared bus and supply noise leaks through to the output when both inputs move together. A CMRR of 80 dB means common-mode ripple is attenuated by a factor of ten thousand under the test conditions, but real performance depends on frequency, gain settings and layout symmetry. High CMRR is essential on switching buses with tens or hundreds of millivolts of ripple.
Gain accuracy defines how closely the amplifier converts the differential input into the expected output. In a current-sensing path, shunt tolerance and front-end gain error add directly, so a 1 percent shunt and 0.5 percent gain amplifier may already consume most of a 1.5 percent total error budget. Critical rails often target tighter gain errors, while supervisory or secondary channels can accept looser limits.
Both differential and common-mode imperfections translate into output error. A small mismatch between inputs appears as pure differential error and is multiplied by the gain. Residual common-mode noise is converted into differential error by finite CMRR. Even simple back-of-the-envelope calculations based on worst-case ripple and CMRR help decide whether a given amplifier is appropriate for the bus environment.
Input Range, Headroom & Output Swing
Datasheets usually specify both a differential input range and a common-mode input range. The device must operate safely inside both at the same time, and the allowed window can narrow at extreme temperatures or supply voltages. Relying only on a headline “0 to 60 V common-mode” figure is risky unless you verify how much differential voltage is supported over that entire domain.
Every amplifier also needs headroom from its supply rails. Input stages may require a few hundred millivolts from ground or VCC before they behave linearly, and the output swing often stops short of the rails even on so-called rail-to-rail devices. These limits become visible when the shunt drop or bus voltage wants to drive the front end close to its extremes.
Matching the output swing to the ADC input range is a system-level step. Maximum shunt current multiplied by gain must keep the amplifier output inside the ADC input window with some margin for overload, and the smallest useful signal should occupy enough codes to stand above noise. It is usually better to leave headroom around both ends of the ADC full-scale than to “fill every code” and risk clipping on spikes.
Noise, Bandwidth & Settling Time
Noise is often specified as a voltage density in nanovolts per square root hertz. Integrated over the effective bandwidth, it sets a floor on the smallest differential voltage you can reliably resolve. Wider bandwidth admits more noise, but overly narrow bandwidth can hide important dynamics or delay the response of control loops and protection thresholds.
Bandwidth targets differ by application. Slow energy metering and state-of-charge tracking often want clean measurements of averaged values with kilohertz-level bandwidth, whereas high-speed current loops and drive control may require tens of kilohertz or more. Simple RC filtering at the inputs or outputs can help balance the trade-off between noise, switching artefacts and usable signal content.
Settling time becomes critical when the front end feeds a multiplexed ADC. After each channel switch, the amplifier and any RC filtering must settle within the error budget before the sample is taken. Comparing the available per-channel time against the datasheet settling curves at the intended gain and loading helps prevent intermittent accuracy or missing-code issues in multi-channel measurement systems.
Error Sources & Matching Techniques
A differential or instrumentation amplifier front end rarely fails because of a single dramatic error. Instead, several modest contributions from the amplifier core, the resistor network and the PCB layout add up to the total measurement uncertainty. Building an explicit error budget helps decide where to spend money on precision components and where simple layout discipline is enough.
Core Amplifier Error Sources
The amplifier itself introduces offset, bias and gain errors. Input offset voltage is effectively an equivalent differential signal that is multiplied by gain and rides on top of the true shunt or sense voltage. Input bias currents flowing through series and source resistances generate additional small voltage drops, which become more important in high-impedance or heavily protected front ends.
Finite open-loop gain, process variation and temperature drift translate into gain error and its change over temperature. While modern INAs offer tightly controlled gain options and low drift, discrete differential stages built from general-purpose op amps typically vary more over time. In a formal error budget it is useful to keep offset-related and gain-related terms separate, because they may be calibrated or trimmed in different ways.
Resistor Matching and Network Design
Resistor networks define the gain and common-mode rejection in both three-op-amp INAs and single-op-amp differential amplifiers. Absolute tolerance affects gain accuracy, while matching between pairs of resistors directly sets the achievable CMRR. Poor matching can quickly erode a good amplifier core and bring common-mode noise back into the measurement.
Practical techniques include using integrated resistor networks with well-controlled ratios, choosing thin film resistors for critical pairs and locating external gain-setting resistors close to the device pins to minimise parasitics. When the front end is gain-programmable, each setting may deserve its own budget for resistor-induced gain error and CMRR degradation, especially in automotive or industrial temperature ranges.
PCB Routing, Leakage and Coupling
Layout often determines how much of the theoretical performance survives on the board. Asymmetric routing, shared return paths and extra copper in one leg of a differential pair all introduce small series impedances that convert common-mode disturbances into differential error. Close proximity to switching nodes can capacitively couple ripple into the sense lines, degrading both CMRR and noise.
Leakage paths from contaminated surfaces, solder flux residue and moisture are another subtle source of error, especially around high-impedance nodes. Guard rings, clean solder mask openings and conservative creepage distances around input pins reduce this risk. A later layout section focuses on placement and routing details; here the goal is simply to recognise that board geometry belongs in the error budget.
Example Error Budget for a 50 mV Differential Signal
Consider a shunt that produces a typical 50 mV differential signal at nominal current and an INA with a gain of 20, giving a 1 V nominal output. Suppose the design targets a total measurement error of about plus or minus one percent at this operating point. One way to approach the problem is to split that one percent across the main contributors and verify each term against realistic component choices.
A simple starting point might allocate roughly 0.5 percent to shunt tolerance, 0.3 percent to INA gain error including any external gain resistor, 0.2 percent to input offset and its drift, and 0.2 percent to layout- and CMRR-related errors. Remaining room can be reserved for ADC quantisation, reference tolerance and temperature effects. The exact numbers will change by project, but the structure of the budget remains similar.
Working through such an example encourages designers to treat matching, layout symmetry and component class as levers instead of afterthoughts. If life-time stability or re-calibration strategy is critical, an additional layer of budgeting for long-term drift can be added and tied into the separate aging and zero-drift reference material, rather than overloading the current-sense front-end specification.
Input Filtering, Line Impedance & Stability
Differential and instrumentation amplifier front ends almost always need some filtering between the noisy power path and the ADC. The goal is to tame switching edges and radio-frequency interference without destroying common-mode rejection or making the amplifier unstable. This section outlines practical RC structures and highlights how long cables and series impedance interact with INA input capacitance.
Input RC & RFI Filters
Simple RC networks in front of a differential or instrumentation amplifier can remove a large portion of switching edges and radio-frequency interference before they reach the ADC. The basic building block is a pair of small series resistors feeding the INA inputs, a capacitor between the lines to limit differential bandwidth and optional, well-matched capacitors from each input to ground to shape the common-mode path.
Preserving CMRR means treating both inputs as symmetrically as possible. Series resistors should share the same value and layout, the differential capacitor should sit between the two filtered nodes and any capacitors to ground should be matched and placed close to the device pins. In practice the differential capacitor dominates the useful signal bandwidth, while very small capacitors to ground mainly bleed away higher-frequency noise and RFI.
Effective RFI filtering rarely comes from a single large capacitor. Instead, a combination of modest resistance, a differential capacitor sized for the intended measurement bandwidth and a few picofarads to tens of nanofarads to ground can greatly reduce the impact of megahertz harmonics and radio pick-up. Exact values depend on the ADC sampling rate and required signal bandwidth, so it is safer to start from the datasheet guidance and refine the network based on bench measurements.
Line Impedance, Cable Effects & Stability
Long cables, large series protection resistors and the input capacitance of an INA or ADC sampling network together form additional poles in the signal path. If these poles sit near the closed-loop bandwidth of the amplifier, they can erode phase margin and cause overshoot, ringing or even oscillation in response to step changes. The risk grows with higher gain, higher bandwidth and more aggressive filtering.
Practical mitigation techniques include placing a small differential capacitor across the inputs, keeping series resistances as low as protection requirements allow and, in demanding cases, buffering very long or high-impedance lines with a local amplifier before handing the signal to the INA. Many datasheets include explicit stability recommendations versus source impedance and load capacitance, and those curves are a good starting point for deciding whether a given cable and filter combination is safe.
Truly high-speed protection problems, such as sub-microsecond overcurrent detection and surge handling, deserve their own treatment in fast current-sense and protection pages. Here the focus stays on preventing slow and medium-speed measurement channels from becoming marginally stable due to unintended line impedance and capacitive loading.
Layout & Kelvin Routing for Differential Voltage Sense
Good performance from a differential or instrumentation amplifier depends as much on how the board is routed as on the datasheet numbers. Kelvin connections, clean reference points and separation between measurement paths and high-current returns all help keep small differential signals from being polluted by shared impedance and switching noise. This section focuses on general layout rules for voltage sensing.
Kelvin Connections & Star Points
Kelvin routing means that the measurement leads connect directly to the component terminals whose voltage you care about, rather than joining the main current path somewhere along a long copper pour. On a shunt resistor that carries high current, the main path enters and leaves on wide traces or planes while a pair of thinner Kelvin traces pick off the voltage at dedicated pads near the resistor body and run back to the INA inputs.
The same principle applies to bus bars, connectors and motor terminals. Measurement points should sit where the system actually sees the voltage, not just where it is convenient to attach a via or test pad. Keeping the sense loop physically separate from the high-current loop and avoiding shared segments of copper reduces the risk that dynamic load currents will appear as extra error in the measurement.
A well-chosen star point connects the measurement reference back to the power return at a single, controlled location. This star node ties together the INA reference, ADC ground and other sensitive circuits, while the noisy power currents follow their own paths. Thinking of the measurement loop as a spectator that touches the power path at only one point helps enforce clean Kelvin routing.
Shielding, Guard & Reference Ground
Differential pairs are most effective when both conductors see nearly the same environment. On the PCB this means routing the two traces close together, with similar length and geometry, over a continuous reference plane. Avoid running one trace near switching nodes, gate-drive lines or digital clocks while the other trace enjoys a quiet route, because this undermines the common-mode cancellation the INA relies on.
Around high-impedance input pins, guard rings can intercept leakage currents that would otherwise flow across the board surface. These guards are copper traces driven to a similar potential as the protected node, placed around the pads and routed between them and any potentially contaminated surfaces. Guards are most helpful in high-resistance bias networks and very low current measurements where even nanoampere leakage is significant.
Separating a quiet analog ground area from the noisy power ground and connecting them at a single tie point near the converter or controller helps keep switching currents out of the measurement paths. Detailed multi-layer grounding strategies and isolation clearances are handled in system-level and zero-drift layout pages; here the emphasis is on giving the differential voltage-sense front end a clean local reference and short, symmetric routes back to it.
Application Patterns & System Hooks
Differential and instrumentation amplifiers rarely live in isolation. They sit at key points in the power and control path, watching bus voltage, shunt drops, phase nodes and droop signals. This section highlights a few common application patterns, and shows where to hand off detailed current measurement, protection and digital telemetry to dedicated companion pages.
Bus + Shunt Combo
In a bus plus shunt combo the same INA family device observes both the bus voltage and the shunt drop. One channel measures the small differential voltage across the shunt for current, while a second channel or output monitors the bus-to-ground voltage for undervoltage and overvoltage supervision.
- Check that both channels stay within the allowed common-mode range at all operating points.
- Share a clean reference and layout region so the bus channel does not pick up switching noise.
- Coordinate current accuracy and bus threshold needs with the dedicated current-sense pages.
Deeper current and protection design: see High-Side Shunt Current Sense Amp and Fast Current Sense for Protection.
Remote Sense for Regulators
Remote sense moves the measurement point from the regulator pins to the load terminals. A differential pair or INA front end tracks the true load voltage and feeds it back to the control loop or ADC, so the regulator can compensate cable and plane drops across the power path.
- Route the sense pair as a tight, quiet Kelvin connection to the load, not to the regulator socket.
- Verify the INA input range and headroom at both light and heavy loads, including any overshoot.
- Loop-compensation details stay with the LDO/DC–DC and loop-stability pages, referenced from here.
More on remote sense and compensation: see LDO Output Remote Sense and Loop Compensation & Stability.
Phase Voltage Sense
Phase-voltage sensing uses INAs or differential amplifiers to watch motor phases or multi-phase converter nodes. The goal is to monitor waveform quality, phase sequence and per-phase stress, feeding information to control algorithms, fault monitors or logging channels.
- Choose front ends with enough common-mode range and dv/dt immunity for switching phase nodes.
- Use careful RC filtering so real phase voltage is preserved but sharp switching spikes are controlled.
- Keep phase-sense differential pairs clear of gate-drive and SW nodes to maintain CMRR and bandwidth.
Motor control and inverter details live in Motor Phase Voltage Sense and related drive-stage pages.
Current Sharing & Droop Check
In current-sharing or droop-based systems, small voltage differences between output nodes reveal how current divides among parallel regulators or modules. A precision differential front end measures these millivolt-level drops and flags imbalance or unexpected current flow directions.
- Budget for very small differential signals and low offset drift over temperature and lifetime.
- Use strict Kelvin routing to each module’s output so shared copper does not hide real imbalance.
- Link long-term accuracy and recalibration strategy to the zero-drift current-sense pages.
For low-drift architecture and digital monitors, see Zero-Drift Current Sense and Digital Current Monitor (I²C/SPI).
Vendor & Part-Class Mapping for Differential / INA Voltage Sense
This page focuses on voltage-sense front ends: classic differential amplifiers, precision instrumentation amplifiers (INAs), wide-common-mode sense amplifiers used as voltage monitors, and fully differential amplifiers (FDAs) that drive ADCs. The table below groups representative part classes from the seven key vendors. Current-sense–branded families and zero-drift shunt monitors are only mentioned briefly and are covered in depth on the dedicated low-side, high-side and zero-drift current-sense pages.
| Vendor | Part Class | Example Part Numbers | Positioning Notes for Differential / INA Voltage Sense |
|---|---|---|---|
| Texas Instruments | Precision INAs (general-purpose) | INA828, INA821, INA826 | High CMRR, low offset and flexible gain with a single external resistor. Good default choice for remote voltage sense, low-frequency phase-voltage monitoring and sensor bridge interfaces where the common-mode stays within ±10–40 V and precision matters more than bandwidth. |
| Texas Instruments | High-voltage difference amplifiers | INA149, INA597, INA597-Q1 | Resistive-diff-amp structures with very wide common-mode capability, useful for bus-voltage monitoring on industrial rails, battery stacks and AC-line sense where pure current measurement is handled by shunt pages. |
| Texas Instruments | FDA drivers for ADC front-ends | THS4551, THS4552, OPA1632 | Fully differential amplifiers with low distortion and well-controlled output common-mode. A good match for high-resolution SAR / delta-sigma ADCs when this page’s INA chain must feed a truly differential converter input. |
| STMicroelectronics | General-purpose precision op amps as diff-amps | TSZ124, TSZ181, TSV914 (with external resistor network) | Zero-drift and low-noise op amps that are commonly used to build single-diff or 3-op-amp INA structures with external resistors. Good for board-level Kelvin sense and remote regulator sense where you want to customize gain and bandwidth. |
| STMicroelectronics | Wide-CM current/voltage sense INAs | TSC2010/11/12, TSC101, TSC888 | Marketed as current-sense amplifiers with wide common-mode ranges. When used with suitably sized shunts or resistive dividers they also provide robust bus or phase-voltage sense; detailed current measurement aspects are handled by high-side pages. |
| NXP | Integrated AFEs with PGAs | N-AFE family (e.g. NAFE113x/123x) | Multi-channel universal-input AFEs that integrate low-drift PGAs, multiplexers and high-resolution ADCs. Appropriate when differential voltage sense is one of several measurement functions (temperature, pressure, loop signals) and you want a single combined front-end. |
| NXP | Classic instrumentation-amp reference designs | Discrete INA circuits built from general-purpose op amps | NXP application notes show how to realise instrumentation-amplifier topologies using their precision op amps. This is useful if you need custom supply ranges or integration into existing signal-chain ICs rather than a monolithic INA. |
| Renesas | Zero-drift programmable-gain INAs | ISL2853x, ISL2863x | Rail-to-rail, low-offset zero-drift INAs with selectable gains. They suit precision differential voltage sense in industrial and instrumentation systems where one device must cover several ranges and where long-term drift is critical. |
| Renesas | General-purpose diff-amp building blocks | ISL28134, ISL28233 and similar precision op amps | Precision op amps frequently used in Renesas reference designs to form difference amplifiers and custom INAs around sensors, bridges and remote sense points—useful when a discrete topology is required instead of a fixed monolithic INA. |
| onsemi | High-voltage diff amps / sense amps | NCS21x/22x, NCV21x (automotive) | Intended for current sensing, but the wide common-mode range and AEC-Q qualified variants make them attractive as compact bus and phase-voltage monitors in automotive and industrial power stages. |
| onsemi | Low-noise op amps for custom INAs | NCS333, NCS325/326 | Precision op amps that can be configured as single-diff amplifiers or 3-op-amp INAs with external resistors, suitable for local Kelvin voltage sense and sensor interfaces. |
| Microchip | Instrumentation / diff amplifiers | MCP6N11/2, MCP6N16, MCP6D22 | Zero-drift and chopper-stabilised INAs/diff amps aimed at low-frequency precision sensing. They are natural candidates for bus, shunt-voltage and remote-sense measurement around MCU-based power stages. |
| Microchip | FDA and ADC drivers | MCP6H01, MCP6L91 and related FDA-capable op amps | Used as single-ended or differential drivers for Microchip’s own ADCs; suitable when this voltage sense chain must feed higher-speed data acquisition. |
| Melexis | Integrated magnetic / position sensors with INAs | MLX90371/72, MLX90393 and similar front-ends | Melexis focuses on smart magnetic and position sensors where the INA and diff front-end is built into the sensor IC. Use these when voltage sense is part of a broader magnetic or position feedback loop, not as a stand-alone board-level INA. |
Note: the above list is indicative rather than exhaustive and is organised by part class, not by specific current-sense or zero-drift shunt amplifiers, which are covered in their own pages.
BOM & Procurement Notes for Differential / INA Voltage Sense
This section turns the electrical requirements of a differential / INA voltage-sense stage into copy-ready BOM and RFQ fields. The goal is that a supplier or distributor can see the intended measurement point, accuracy, bandwidth and qualification level from the line item alone, without reverse-engineering the schematic.
Core Electrical & Topology Fields
| Field in BOM / RFQ | What to Write | Why It Matters |
|---|---|---|
| Sense node & function | e.g. “Bus + shunt combo”, “Remote sense for 5 V rail”, “3-phase voltage sense”, “Droop / current-sharing check”. Indicate whether the INA sees a shunt, a phase node, or a remote load point. | Vendors understand if you need high-voltage diff, high CMRR in noisy motor phases, or low-offset bridge-style sensing, and can recommend the right INA / diff-amp family. |
| Differential input range | State the minimum and maximum differential voltage you want to measure (e.g. ±50 mV, ±500 mV) and whether the signal is unipolar or bipolar. | Sets the required gain, offset performance and noise floor. Helps avoid parts that saturate or lack resolution at your smallest useful differential. |
| Common-mode range at sense point | Specify the common-mode voltage window at the measurement nodes (e.g. “0–80 V bus”, “−2 V to +12 V motor phase”, “1.5 V around mid-supply reference”). | Determines whether you need a precision INA, a high-voltage diff amp, or a wide-CM current-sense device repurposed as a voltage monitor. |
| Accuracy & drift targets | Give a total error budget over temperature (e.g. “≤ ±0.5 % of reading, −40 to 125 °C”) and, if relevant, CMRR at mains frequencies (e.g. “CMRR ≥ 100 dB @ 50/60 Hz”). | Enables the vendor to choose between general-purpose INAs and zero-drift / chopper parts and to check that CMRR and gain accuracy align with your current-sense and ADC pages. |
| Bandwidth & settling requirements | State required small-signal bandwidth (e.g. “DC–50 kHz”), settling time to a given error (e.g. “<5 µs to 0.1 %”) and ADC sample rate / mux switching frequency if applicable. | Prevents overspecifying a fast, noisy part for slow measurements, or under-specifying bandwidth for motor-phase and transient monitoring. |
| Output format & load | Indicate whether you need single-ended or fully differential output, and what ADC or comparator input it must drive (input range, reference level, input impedance, sampling capacitor). | Guides towards classic single-ended INAs, FDA drivers or integrated AFEs, and ensures output swing and drive strength are adequate. |
Package, Environment & Functional-Safety Fields
| Field in BOM / RFQ | What to Write | Why It Matters |
|---|---|---|
| Package & height limits | Specify preferred package families and maximum mounted height (e.g. “SOIC-8 / TSSOP-14, max 1.2 mm”) and whether you allow SOT-23 or DFN for dense layouts. | Avoids proposals that cannot be assembled on the existing PCB stack-up or violate creepage / clearance constraints in high-voltage corners. |
| Temperature & qualification grade | State the required temperature range and grade (e.g. “−40 to 125 °C industrial”, “−40 to 150 °C AEC-Q100 Grade 0”), and whether an automotive or industrial qualification is mandatory or optional. | Lets vendors pick the correct suffixes (-Q1, -A, etc.) and avoid parts that will later fail qualification reviews. |
| Supply & power constraints | Give allowed supply rails (single or dual), quiescent current budget per channel and any startup sequencing constraints (e.g. “runs from 5 V rail, IQ ≤ 1 mA/ch, no negative rail available”). | Filters out parts that need dual rails or exceed standby power limits in low-power or always-on monitors. |
| Diagnostic & functional-safety support | Indicate if you require documented diagnostics or safety data: “ASIL-capable preferred, FMEDA / FIT data required”, “supports open/short detection via external network”, or “no explicit safety docs required”. | Encourages vendors to propose parts with safety manuals, FIT data and app notes that simplify ISO 26262 or IEC 61508 compliance, or to confirm when a commodity INA is acceptable. |
| Second-source & lifecycle expectations | Add notes such as “dual-source desirable across TI / onsemi / Microchip” or “no niche parts with near-EOL status”. If relevant, state expected product lifetime (e.g. “10+ years automotive production”). | Reduces the risk of selecting parts with fragile supply chains or short lifecycles for key sense paths that are hard to redesign late in the project. |
When you send an RFQ, including the above fields alongside a preferred part class (for example “TI INA828-class precision INA, or equivalent” or “Renesas ISL2853x-class zero-drift INA”) usually yields much more consistent proposals than just writing “instrumentation amplifier, 5 V”.
Request a Quote
FAQs on Differential / INA Voltage Sensing
This FAQ collects twelve practical questions engineers often ask when they design differential and instrumentation amplifier voltage sensing stages. It gives quick, concrete answers on when to pick each topology, how to plan accuracy and bandwidth, and how to avoid layout and filtering mistakes around noisy power rails.
1. When should I use a full instrumentation amplifier instead of a simple differential amplifier for voltage sensing?
Use a full instrumentation amplifier when the source impedance is high, the sense nodes sit far from the ADC or you expect large and varying common mode noise. The INA gives high input impedance, trimmed gain and guaranteed CMRR. A simple differential amplifier is fine for short local traces and modest accuracy.
2. How much CMRR do I really need when sensing a small differential voltage on a noisy bus?
When you measure tens of millivolts on a rail that carries hundreds of millivolts of ripple, CMRR sets how much of that ripple leaks into the reading. For many industrial buses, around eighty to ninety decibels across the noise band is enough. Precision metering and tight current sharing checks usually benefit from one hundred decibels or more.
3. How do resistor matching and PCB layout impact the CMRR of a discrete differential amplifier?
In a discrete differential amplifier the theoretical CMRR assumes perfectly matched resistor ratios and symmetrical routing. Real world tolerances, temperature drift, unequal trace lengths and stray capacitance all convert common mode into differential error. Using precision resistor networks, tight physical grouping and closely coupled differential traces keeps the practical CMRR close to the datasheet value.
4. What input range and headroom checks should I run before finalizing a differential or INA front end?
Before locking the front end, sweep minimum and maximum common mode and differential voltages against the amplifier input range and supply rails. Check output swing limits into the planned load, including gain error and overdrive conditions. Finally review cold start, brown out, surge and fault cases so the inputs never clip or latch up in the field.
5. How do I size the input RC filters without destroying bandwidth or stability?
Start from the highest useful signal frequency and pick a differential cutoff somewhat above it so the waveform is preserved. Choose equal series resistors and a single differential capacitor, then add small common mode capacitors for radio frequency interference. Keep series resistance moderate and review the amplifier stability guidance, especially when long cables and input capacitance are present.
6. What is the best way to sense bus voltage and shunt voltage with the same differential front end?
A common approach is to use a multi channel instrumentation amplifier or a current sense amplifier plus an extra differential stage. One path measures the shunt drop and the other samples the bus to ground. Keep both channels within their common mode limits and share a clean reference and layout region to avoid coupling heavy switching noise.
7. How do long cables, connectors and line impedance affect instrumentation amplifier accuracy?
Long cables and connector resistance add extra series impedance and voltage drop, so the voltage at the sensor end may differ from the local node you think you are measuring. They also introduce capacitance that interacts with the instrumentation amplifier input. Using Kelvin connections, local buffering and carefully chosen series resistors keeps errors under control.
8. When is it acceptable to use a non zero drift differential amplifier in energy measurement paths?
A standard differential amplifier is often fine for short term logging, efficiency checks and internal diagnostics where occasional recalibration is easy. For billing grade meters and multi year energy accumulation, offset and gain drift quickly dominate the error budget. Those applications usually call for a zero drift architecture and a defined recalibration strategy.
9. How can I share one instrumentation amplifier among multiple measurement channels using a mux?
You can place a low leakage analog multiplexer in front of a single instrumentation amplifier and switch different sensor pairs, but you must treat settling time and charge injection with care. Allow enough time after each switch event for the inputs and output to settle before the ADC samples, and keep source impedance and wiring consistent between channels.
10. What practical layout rules matter most for low noise differential voltage sensing near switching regulators?
Route the differential pair as a short, tightly coupled path that avoids switch nodes, gate drive traces and transformer windings. Keep the instrumentation amplifier in a quiet analog ground island with a single connection back to power ground. Use guard rings or shielding copper around high impedance nodes and place filter components tight and symmetric around the inputs.
11. Which datasheet parameters should I highlight in the BOM so suppliers propose suitable INA options?
Call out the common mode range, gain or programmable gain, gain error, input offset and drift, CMRR over the relevant frequency band, noise density, bandwidth, quiescent current and supply rails. Add package, temperature range, qualification grade and any diagnostic or safety document needs so distributors can filter their portfolio to realistic instrumentation amplifier candidates.
12. How do I choose between a precision INA, a wide common mode sense amplifier and an FDA for my ADC front end?
If you need accurate low frequency sensing with moderate common mode voltage, a precision instrumentation amplifier is usually the best default. When the sense nodes ride on very high or switching common mode, a wide common mode sense amplifier is safer. Choose a fully differential amplifier when the ADC expects a symmetrical differential drive and high dynamic performance.
