A voltage doubler circuit is a passive charge pump network that uses diodes and capacitors to produce a DC output voltage approximately twice the peak of the AC input, without requiring a heavy, expensive step-up transformer. If you feed it 12V AC (RMS), you get roughly 32V DC out. It is the go-to topology for generating bias voltages for photomultiplier tubes, CRT anodes, or high-voltage gate drives when your current requirements are strictly in the milliamp range.
This guide breaks down the half-wave (Greinacher) voltage doubler topology, contrasts it with active switching alternatives, and walks through a real-world breadboard build with exact component values and failure-mode analysis.
The Half-Wave Voltage Doubler Topology Explained
The most common DIY voltage doubler is the half-wave series doubler, often called the Greinacher circuit. It consists of two diodes and two capacitors arranged to pump charge during alternating half-cycles of the AC input.
Topology Node Labels
- Node IN: The AC input source (e.g., secondary winding of a step-down transformer).
- Node N1: The junction between Capacitor 1 (C1) and Diode 1 (D1). This node clamps to the negative peak of the AC waveform.
- Node N2: The junction between D1, Diode 2 (D2), and Capacitor 2 (C2). This is the pumping node.
- Node OUT: The positive DC output, measured across C2 relative to ground.
- Node GND: The common ground, tied to the bottom of C1 and the cathode of D2.
How it pumps: During the negative half-cycle of the AC input, D1 is forward-biased and charges C1 to the peak input voltage (Vpk). During the positive half-cycle, the AC source voltage adds to the voltage already stored in C1. This combined voltage forward-biases D2, pumping charge into C2. Because C2 sees the AC peak plus the C1 peak, it charges to 2 × Vpk (minus two diode forward voltage drops).
Why Choose a Passive Doubler Over a Boost Converter?
When you need higher voltage, the immediate alternative is an active switching boost converter (like the MT3608 or a TPS61230). Why use a passive diode-capacitor network instead? The decision hinges on your current draw and EMI tolerance.
| Criteria | Passive Voltage Doubler | Active Boost Converter (IC) |
|---|---|---|
| Output Current Capability | Very Low (< 20mA typical) | High (Hundreds of mA to Amps) |
| Voltage Regulation | Poor (Voltage sags heavily under load) | Excellent (Closed-loop feedback) |
| EMI / Switching Noise | Zero (Passive rectification only) | High (MHz switching nodes require filtering) |
| Component Count & Cost | 4 components (< $0.50 total) | 6+ components ($2.00 - $5.00 total) |
| Startup Time | Slow (Limited by RC time constant) | Fast (Soft-start circuitry) |
Choose the voltage doubler when: You need high voltage (30V-100V+) at microamp to low-milliamp currents, you are highly sensitive to switching noise (e.g., audio preamps, sensor biasing), or you want a purely analog, radiation-hardened solution.
Choose a boost converter when: You need to power motors, LEDs, or logic rails that draw >50mA and require tight voltage regulation regardless of load transients.
Design Walkthrough: Real Component Values and Behavior
Let's design a doubler to generate a ~30V DC bias rail from a standard 12V AC (RMS) wall transformer. The peak voltage of 12V AC is roughly 17V. Our theoretical maximum output is 34V DC.
Component Selection
- D1, D2: 1N4007. Rated for 1A and 1000V Peak Inverse Voltage (PIV). The PIV across D2 will be 2 × Vpk (~34V), so a standard 1N4001 (50V PIV) would technically survive, but the 1N4007 costs the same and provides a massive safety margin against transient spikes.
- C1, C2: 22µF, 50V Aluminum Electrolytic. C1 must withstand Vpk (~17V). C2 must withstand 2 × Vpk (~34V). A 50V rating gives us a 20% derating margin on C2, which is critical for electrolytic longevity.
Behavior Table: What Changes When Elements Change
| Variable Changed | Effect on Output Voltage (Vout) | Effect on Ripple |
|---|---|---|
| Increase C1 / C2 value | Slight increase under load (less sag) | Decreases ripple significantly |
| Increase Load Current | Massive drop (poor load regulation) | Increases ripple linearly |
| Increase AC Frequency | Slight increase (caps charge faster) | Decreases ripple significantly |
| Use Schottky Diodes | Increases by ~0.8V (lower Vf drop) | No direct effect on ripple |
Failure-Mode Contrast: Series vs. Parallel Extremes
Understanding what breaks at the extremes requires looking at the series/parallel nature of the charge pump. C1 and C2 charge in a parallel-like sequence from the AC source but discharge in series into the load.
- Short D1: C1 charges to the positive peak, but the pumping action to N2 is destroyed. D2 will conduct on every positive half-cycle. Vout drops to 1 × Vpk (half-wave rectifier behavior).
- Open C1 (Series Failure): The charge pump is broken. No energy is stored to add to the positive half-cycle. Vout drops to 0V.
- Open C2 (Parallel/Filter Failure): The pump still operates, but there is no output reservoir. Vout becomes a pulsing DC waveform peaking at 2 × Vpk but dropping to 0V between cycles. Your multimeter will read an unpredictable average, while a scope shows massive 100% ripple.
- Short C1: The AC source is essentially placed directly across D2 during the positive half-cycle. If your AC source is not current-limited, D2 will likely overheat and fail short, potentially taking out the transformer secondary winding.
Step-by-Step Breadboard Testing
Testing high-voltage multipliers requires strict verification steps. Do not skip the de-energize checks.
- Prepare the Source: Plug in your 12V AC wall transformer. Verify the output with a multimeter set to AC voltage. You should read between 11.5V and 12.5V AC. Do not use a DC bench supply; the doubler requires alternating polarity to pump charge.
- Build the Network: Insert D1 and D2 into the breadboard, ensuring the cathode stripes point toward Node N2 and Node OUT respectively. Insert C1 and C2, strictly observing the polarity stripe (negative leg to GND for C2, negative leg to Node IN for C1).
- Pre-Flight Continuity Check: With the AC source unplugged, use your multimeter in continuity mode. Probe from Node OUT to GND. You should read an open circuit (OL) once C2 charges slightly from the meter's test current. If it reads a dead short, check for a backwards C2 or a solder bridge.
- Energize and Measure DC: Plug in the AC source. Set your DMM to DC Voltage. Probe Node OUT (red) and GND (black). You should read approximately 31V to 33V DC (accounting for the ~0.7V drop across each silicon diode).
- Measure Ripple (Optional): Connect an oscilloscope. Set the input coupling to AC. Probe Node OUT. With no load, ripple should be < 50mV. Connect a 10kΩ load resistor across OUT and GND (drawing ~3mA). The ripple should increase, but the DC average should remain above 28V.
- De-energize and Bleed: Unplug the AC source. Warning: C2 will hold a 30V charge for minutes or hours. Use a 1kΩ power resistor to safely bleed Node OUT to GND before touching the breadboard.
Voltage Doubler Circuit FAQ
Can a voltage doubler circuit work with a DC input?
No. A passive diode-capacitor voltage doubler relies entirely on the alternating polarity of an AC waveform to sequentially charge and transfer energy between capacitors. If you apply pure DC, C1 will charge to the DC voltage once, D1 will reverse-bias, and the circuit will halt. To double a DC voltage, you need an active charge pump IC (like the ICL7660) that uses internal MOSFETs to artificially 'chop' the DC into an AC square wave before feeding it into a capacitor network.
How do I calculate the ripple voltage in a voltage doubler?
The peak-to-peak ripple voltage ($V_{ripple}$) in a half-wave doubler is dictated by the load current, the AC frequency, and the output capacitance. The formula is:
V_ripple = I_load / (f × C2)
Where I_load is in Amps, f is the AC frequency in Hertz, and C2 is in Farads. For example, if you draw 5mA (0.005A) from a 60Hz source using a 22µF (0.000022F) capacitor, your ripple will be: 0.005 / (60 × 0.000022) = 3.78V peak-to-peak. To reduce this ripple, you must either increase C2 or increase the input frequency (which is why switch-mode charge pumps run at 100kHz+).
Why is my voltage doubler circuit output lower than expected?
If your theoretical math says 34V but your multimeter reads 22V, you are experiencing the harsh reality of charge pump internal impedance. The most common culprits are:
1. Excessive Load: Voltage doublers have terrible load regulation. Even a 20mA load can cause the output to sag by 30% or more.
2. High ESR Capacitors: Cheap electrolytic capacitors have high Equivalent Series Resistance. The ESR acts as a hidden resistor in series with the pump, dropping voltage under load.
3. Diode Reverse Recovery: If you are running the doubler at high frequencies (>1kHz), standard 1N4007 diodes are too slow. They will conduct backwards during the switching transition, bleeding charge back into the source. Switch to UF4007 (ultra-fast) or Schottky diodes for high-frequency applications.
