Table of Contents
- How do photocurrent and output headroom set feedback resistance?
- Why can a lower-bias-current amplifier still be the wrong replacement?
- Account for dark current and board leakage separately
- How should the TIA be validated before an amplifier purchase is released?
- Turn the measurement limits into an RFQ specification
Low offset current is a weak starting point for a photodiode amplifier. The circuit must convert the detector’s current into a useful voltage without losing the smallest signal or saturating at the brightest condition, and selecting an operational amplifier by offset alone misses the detector capacitance, leakage and feedback behavior that often control the result.
Selecting a transimpedance amplifier for a photodiode means qualifying the detector, amplifier, feedback network and board layout together. The purchasing specification should be a bounded optical-to-electrical signal-chain requirement, not simply a request for a “precision op amp.”
How do photocurrent and output headroom set feedback resistance?
Choose feedback resistance from the maximum expected detector current and the available output-voltage excursion, then include tolerances and non-signal currents. For a simple inverting TIA, the signal excursion has magnitude approximately Iphoto × RF while the amplifier remains in its linear operating region.
Suppose the circuit uses a 1.8 V reference, produces a decreasing output as photocurrent rises, and must remain above 0.2 V. The available negative-going excursion is 1.6 V. With a maximum 20 µA photocurrent, the initial resistance ceiling is 1.6 V / 20 µA = 80 kΩ, before adding design margin.
A 100 kΩ feedback resistor would demand a 2.0 V excursion and cannot meet that assumed output limit. A 75 kΩ candidate produces a 1.5 V excursion, leaving 0.1 V nominal margin. That small margin still needs to absorb component tolerance, reference error, amplifier behavior and any additional current. The example fixes the arithmetic; it does not set a design target.
Check the amplifier’s specified output swing at the actual load and supply. If an analog-to-digital converter follows the TIA, also compare the output with the converter’s allowed input range and acquisition requirements. The SAR ADC settling guide addresses that next interface.
Why can a lower-bias-current amplifier still be the wrong replacement?
Lower input bias current can reduce one error term, but a replacement can change bandwidth, noise, input capacitance and loop stability. The best choice depends on the detector and the complete feedback network.
Analog Devices’ transimpedance stability analysis explains how input capacitance and amplifier response interact with the feedback network, and how a feedback capacitor can restore stability at a bandwidth cost. The source supports the mechanism; the required component values depend on the actual design.
Compare at least these parameters at the intended operating conditions:
| Parameter | What changes the selection | What a headline value can conceal |
|---|---|---|
| Input bias and leakage | Smallest photocurrent and temperature range | Room-temperature performance may not represent the hot limit |
| Input voltage/current noise | Detector impedance, gain and measurement bandwidth | A low single-frequency noise value does not give integrated output noise |
| Input capacitance and gain-bandwidth behavior | Detector capacitance, layout and feedback network | A faster device may require different compensation |
| Output swing and recovery | Maximum light, reference level and following load | Saturation recovery can dominate a pulsed measurement |
| Supply and common-mode range | Photodiode bias and reference configuration | A matching package does not prove a matching operating range |
Use the general op amp replacement checks for pinout and DC compatibility, then add the detector-specific checks here.
Account for dark current and board leakage separately
Detector dark current, amplifier input bias and board leakage can all produce output error, but they have different dependencies. Separate budgets make it possible to see which term changed.
For illustration, an additional 5 nA of net input current through 75 kΩ creates 0.375 mV of output shift. Whether that is acceptable depends on the smallest useful signal and the error budget.
Record photodiode bias, temperature, ambient-light exclusion and settling time during a dark measurement. A changing baseline may come from the detector, the amplifier, contamination or measurement setup. A single offset correction at room temperature cannot establish performance across all those conditions.
Noise also needs a defined bandwidth. Compare integrated noise over the application’s measurement band, with filtering and sampling conditions stated. Do not add unrelated datasheet maxima as if they were uncorrelated RMS noise contributions, or compare two measurements taken with different bandwidth settings.
How should the TIA be validated before an amplifier purchase is released?
Validate the candidate with the intended detector capacitance, feedback network, layout and load, then measure small-signal response and recovery from the largest expected input. Simulation helps identify candidates, but it does not cover every board leakage path or model limitation.
The acceptance run should include a dark baseline, several calibrated optical or equivalent electrical input levels, and a step response that reveals ringing and settling. Test the bright-input limit and recovery after overload where that condition is possible in service. Record temperature and supply conditions for each result.
An electrical current-injection fixture can isolate amplifier behavior, but it does not necessarily reproduce a real detector’s capacitance, bias dependence or optical response. Record both the fixture definition and why its result is relevant. If the detector changes, revisit the input model even when the amplifier ordering code remains the same.
Turn the measurement limits into an RFQ specification
For industrial signal-chain sourcing, provide the full amplifier and detector codes, supply range, photocurrent range, required bandwidth, temperature envelope and allowed output error. Identify which feedback parts and layout are already fixed.
The Analog Devices brand page can support manufacturer-level sourcing, while the engineering comparison should stay at the exact device and configuration level. A supplier’s pin-compatible suggestion becomes useful only when its data supports the specified operating point.
Approve the replacement against the measured signal-chain requirements. If it needs a new feedback capacitor, gain resistor or layout change, release that revised circuit as a controlled configuration rather than describing the IC alone as a drop-in substitute.
Frequently Asked Questions (FAQ)
Can an electrical current source replace all optical qualification?
No. Current injection can isolate amplifier behavior, but it may not reproduce detector capacitance, bias dependence or optical response. Document its limits and retain detector-level tests where needed.
Should two photodiodes with the same area use the same compensation?
Not automatically. Compare capacitance at the operating bias, dark current, responsivity and the required bandwidth before carrying over the feedback network.
Is room-temperature dark-offset calibration sufficient?
Only if it meets the complete application requirement. Temperature-dependent detector current, amplifier bias and board leakage may require additional characterization or compensation.
