Silicon photodiodes convert incident light into electrical current with high speed, excellent linearity, and predictable temperature characteristics. Unlike cadmium sulfide (CdS) photoresistors that suffer from memory effects and slow response times (often >10ms), or phototransistors that trade bandwidth for internal gain, a silicon photodiode gives you nanosecond response times and a strictly linear current-to-light relationship. If you are building a pulse oximeter, a high-speed optical encoder, or a precision light meter, this is the component you reach for.
Symbol, Pinout, and Safe Default Part Numbers
On a schematic, the silicon photodiode symbol is identical to a standard PN junction diode, but with two arrows pointing inward toward the junction, indicating incident photons. The pins are the Anode (A) and Cathode (K).
In through-hole 5mm clear epoxy packages, the cathode is the short leg and corresponds to the flat edge of the epoxy dome. In surface-mount packages (like the popular BPW34), the cathode is marked by a dot, a notch, or a thick silkscreen line on the PCB footprint. Always verify the cathode orientation before soldering; reversing the bias in a photoconductive circuit will forward-bias the diode, clamping your signal to ~0.4V and ruining your measurement.
Here are the safe default part numbers for 90% of hobbyist and commercial bench projects, complete with their critical ratings:
| Part Number | Manufacturer | Active Area | Peak Wavelength | Max Reverse Voltage | Dark Current (Typ) | Best Use Case |
|---|---|---|---|---|---|---|
| BPW34 | Vishay / Osram | 7.5 mm² | 850 nm (Broadband) | 60V | 1 nA @ 10V | General visible/IR, pulse oximetry, light meters |
| SFH203 | Osram | 1.0 mm² | 900 nm | 5V | 1 nA @ 5V | IR remote receivers, beam-break sensors |
| BPX65 | TT Electronics | 0.78 mm² | 850 nm | 60V | 2 nA @ 10V | Harsh environments (TO-18 metal can package) |
| S1223 | Hamamatsu | 5.8 mm² | 960 nm | 30V | 0.1 nA @ 10V | Precision UV/Vis spectroscopy, ultra-low light |
Photovoltaic vs. Photoconductive Biasing Regions
How you bias a silicon photodiode dictates its speed, noise floor, and linearity. You must select the operating region based on your application's bandwidth requirements.
1. Photovoltaic Mode (Zero Bias)
The diode is connected directly across the inputs of an op-amp with no external voltage applied. The light generates a small forward voltage (up to ~0.4V). Because there is zero external reverse voltage, the depletion region is narrow, resulting in high junction capacitance. This makes the circuit slow (kilohertz range), but it completely eliminates dark current, giving you the lowest possible noise floor.
2. Photoconductive Mode (Reverse Bias)
A positive voltage (typically 3V to 15V) is applied to the cathode, while the anode is held at a virtual ground. This widens the depletion region, drastically reducing junction capacitance and allowing for megahertz-level bandwidth. The trade-off is the introduction of dark current (leakage current that flows even in total darkness), which adds shot noise to your signal.
| Parameter | Photovoltaic (Zero Bias) | Photoconductive (Reverse Bias) |
|---|---|---|
| Voltage Across Diode | 0V to +0.4V | -3V to -60V (Cathode positive) |
| Junction Capacitance | High (e.g., 70pF for BPW34) | Low (e.g., 15pF for BPW34 @ 10V) |
| Response Time | Slow (µs to ms) | Fast (ns) |
| Dark Current | Zero | Non-zero (nA range) |
| Primary Use Case | DC light measurement, low-light precision | High-speed pulses, optical communications |
Complete Transimpedance Amplifier (TIA) Circuit
A photodiode outputs current, not voltage. To interface it with a microcontroller ADC, you need a Transimpedance Amplifier (TIA) to convert the current to a usable voltage. Below is a complete, bench-tested design for a 100kHz bandwidth TIA using a BPW34 and an OPA380 op-amp, optimized for a single 5V supply.
Component List & Values
- U1: OPA380 (Texas Instruments) - Single-supply, low-input-bias-current op-amp optimized for TIAs.
- D1: BPW34 Silicon Photodiode.
- Rf (Feedback Resistor): 1 MΩ (0.1% tolerance, metal film). Sets the gain: 1 µA of light = 1V output.
- Cf (Feedback Capacitor): 2.7 pF (C0G/NP0 ceramic). Prevents high-frequency oscillation caused by the diode's parasitic capacitance.
- R1, R2 (Bias Divider): 10 kΩ each. Creates a 2.5V virtual ground reference for the non-inverting input.
- C_bypass: 100 nF ceramic across the 5V and GND pins of the OPA380.
Wiring Steps
- Connect the Cathode (K) of the BPW34 directly to the +5V rail. This applies a 2.5V reverse bias relative to the virtual ground, dropping the diode's capacitance to roughly 25pF.
- Connect the Anode (A) of the BPW34 to the Inverting Input (-) of the OPA380.
- Wire the 1 MΩ Rf and 2.7 pF Cf in parallel between the OPA380's Inverting Input (-) and its Output pin.
- Connect the junction of the R1/R2 voltage divider (2.5V) to the Non-Inverting Input (+) of the OPA380.
- Route the OPA380 Output to your microcontroller's ADC pin. The output will sit at 2.5V in total darkness and drop toward 0V as light intensity increases.
Failure Modes and Multimeter Testing
Silicon photodiodes are rugged, but they do fail. The most common failure mode on the bench is shunt resistance degradation caused by moisture ingress through a cracked epoxy lens, which manifests as an artificially high dark current. In high-power laser applications, they suffer from Catastrophic Optical Damage (COD), where the silicon junction literally melts from localized thermal runaway.
Here is how to test a suspect photodiode with a standard digital multimeter (DMM):
- Forward Bias Test: Set your DMM to Diode Test mode. Place the red probe on the Anode and the black probe on the Cathode. A healthy silicon photodiode will read between 0.35V and 0.55V. If it reads 0.0V, the junction is shorted.
- Reverse Blocking Test: Swap the probes (Red to Cathode, Black to Anode). The meter should read OL (Over Limit). If it reads a low voltage or a resistance value, the diode is leaky or shorted and must be discarded.
- Photovoltaic Light Test: Set the DMM to the lowest DC millivolt (mV) range. Connect Red to Anode, Black to Cathode. Shine a bright LED flashlight directly onto the lens. The meter should immediately jump to 100mV - 350mV depending on the light intensity. If the voltage sluggishly drifts or doesn't exceed 10mV under a bright beam, the diode has degraded.
Frequently Asked Questions
Can I use a silicon photodiode to measure exact lux or PAR for grow lights?
Not directly out of the box. Silicon photodiodes have a spectral response that peaks in the near-infrared (around 850nm-950nm), whereas human vision (lux) peaks at 555nm (green), and Photosynthetically Active Radiation (PAR) spans 400-700nm. If you shine a standard white LED on a bare BPW34, the diode will heavily over-report the IR bleed from the LED's phosphor. To measure lux, you must place a specialized photopic correction filter (like the Hoya CM-5000) over the diode to block IR and shape the response curve to the CIE 1931 standard. For PAR, you need a blue-red enhanced diode and a specific calibration factor.
Why is my photodiode TIA circuit oscillating or ringing on the oscilloscope?
This is almost always caused by ignoring the feedback capacitor ($C_f$). The photodiode's junction capacitance ($C_d$), combined with the op-amp's input capacitance, creates a pole in the feedback loop that destroys the phase margin. When you use a high-value feedback resistor (like 1MΩ or 10MΩ) to get high gain, this pole causes the circuit to ring or break into high-frequency oscillation (often in the 1MHz - 10MHz range). Adding a small capacitor (typically 1pF to 10pF) in parallel with the feedback resistor introduces a zero that cancels the pole, restoring stability. Use C0G/NP0 ceramic capacitors for $C_f$; X7R capacitors are microphonic and voltage-dependent, which will inject noise into your signal.
What is the difference between a PIN silicon photodiode and a standard PN junction?
A standard PN photodiode has the P-type and N-type silicon directly adjacent to each other. A PIN photodiode inserts a wide, undoped Intrinsic (I) silicon layer between the P and N regions. This wider intrinsic layer does two things: it increases the volume where photons are absorbed (improving quantum efficiency in the near-IR spectrum), and it physically separates the P and N plates, drastically reducing the junction capacitance. If you are building a 10Mbps fiber optic receiver or a high-speed LIDAR pulse detector, you must use a PIN diode. For simple ambient light sensing or slow beam-breaks, a standard PN junction like the BPW34 is cheaper and perfectly adequate.
