A DC to DC voltage multiplier is an inductorless switching circuit that uses capacitors and semiconductor switches to step up, invert, or multiply an input DC voltage. In a real PCB layout or breadboard build, deploying a multiplier fundamentally changes your bill of materials and EMI profile: it eliminates the need for bulky magnetic components (inductors or transformers), shrinking the physical footprint and removing the primary source of radiated magnetic EMI, but it trades that advantage for higher output voltage ripple and a strict current ceiling that rarely exceeds 100mA.
Instead of storing energy in a magnetic field, these circuits transfer discrete packets of charge through a network of capacitors. If you picture a bucket brigade passing water to put out a fire, the capacitors are the buckets and the semiconductor switches (diodes or MOSFETs) are the people handing them along. By rapidly switching the capacitors between parallel charging and series discharging configurations, the circuit stacks the input voltage to achieve the desired output.
Core Topologies and Output Expectations
The term "voltage multiplier" is often used interchangeably with "charge pump," though strictly speaking, a charge pump is the broader category of switched-capacitor converters, while a multiplier refers specifically to topologies that scale the voltage by an integer factor (2x, 3x, 4x) or invert it. Below is a reference matrix of the most common configurations you will encounter in both discrete designs and integrated circuits.
| Topology | Ideal Vout Formula | Typical Max Iout | Common IC / Implementation | Primary Limitation |
|---|---|---|---|---|
| Voltage Doubler (1-Stage) | 2 × Vin | ~50mA to 100mA | ICL7660 (doubler config), TPS60400 | Output impedance scales with switching frequency |
| Voltage Tripler (2-Stage) | 3 × Vin | ~20mA to 40mA | Discrete Dickson Charge Pump | Compounding diode forward voltage (Vf) drops |
| Quadrupler (3-Stage) | 4 × Vin | ~10mA to 20mA | Discrete Cockcroft-Walton | Severe voltage droop under dynamic loads |
| Voltage Inverter (1-Stage) | -1 × Vin | ~100mA to 250mA | MAX660, LT1054, ADM660 | Generates negative rail; limited by cap ESR |
Worked Numeric Example: The 5V to 18V Quadrupler
Let us look at what happens when theory meets the workbench. Suppose you need to generate a ~20V bias rail for an LCD display from a standard 5V USB supply, and your load only draws 10mA. You decide to build a discrete 3-stage Cockcroft-Walton quadrupler (which yields 4x voltage gain) using 1N5819 Schottky diodes and 1µF X7R ceramic capacitors.
The Ideal Math:
Vout(ideal) = 4 × Vin = 4 × 5V = 20.0V.
The Real-World Math:
In a discrete diode-based multiplier, every stage introduces a forward voltage drop (Vf). For a 1N5819 Schottky diode at low currents, Vf is approximately 0.3V. Because a quadrupler passes current through four diode junctions in its charging/discharging paths, we must subtract these drops.
However, we are not done. The capacitors have Equivalent Series Resistance (ESR), and the 10mA load will cause voltage droop between the switching cycles. Assuming a switching frequency of 10kHz and a total effective ESR and ripple loss of roughly 0.6V at 10mA, the final measurable output on your multimeter will be:
Vout(final) = 18.8V - 0.6V = 18.2V.
This 1.8V discrepancy (nearly 10% of your ideal target) is exactly why discrete multipliers are rarely used for precision voltage references. If you need exactly 20.0V, you must either increase the input voltage to 5.45V to compensate for the drops, or follow the multiplier with a low-dropout linear regulator (LDO).
Where You Meet This in Practice
While inductive boost converters dominate high-power applications, voltage multipliers are the undisputed champions of specific low-power, high-voltage, or noise-sensitive niches. According to semiconductor design literature, you will frequently encounter these circuits in the following scenarios:
- SiC and GaN Gate Drivers: Modern silicon carbide (SiC) MOSFETs often require a negative gate bias (e.g., -4V) to prevent false turn-on from Miller capacitance coupling. A simple inverting charge pump takes the 12V or 15V driver supply and generates the required -4V negative rail without needing an isolated transformer.
- EEPROM and Flash Programming: Many microcontrollers and external memory chips require a 12V pulse to write to EEPROM cells, even when the MCU is running at 3.3V. An on-chip or external 4-stage charge pump steps the 3.3V up to the necessary programming voltage for the few milliseconds required to burn the fuse.
- Geiger-Müller Tubes and Photomultipliers: Radiation detection and low-light optical sensors require extreme bias voltages (300V to 1000V+) but draw virtually zero current (microamps). A 10-stage or 15-stage Cockcroft-Walton multiplier driven by a high-frequency oscillator is the standard, cost-effective way to generate these lethal but low-energy voltages from a 5V battery.
- Op-Amp Negative Rails: In mixed-signal audio or sensor circuits, running a single-supply op-amp into the ground rail causes clipping and crossover distortion. An inverting charge pump (like the ICL7660) generates a clean -5V rail from a +5V supply, allowing the op-amp output to swing symmetrically around true 0V.
Voltage Multiplier vs. Inductive Boost Converter
The most common mistake hobbyists and junior engineers make is confusing a switched-capacitor voltage multiplier with an inductive boost converter. While both step up DC voltage, their underlying physics and practical limits are entirely different.
A boost converter stores energy in the magnetic field of an inductor. When the internal MOSFET switch opens, the inductor's collapsing magnetic field forces current through a diode into the output capacitor, stepping up the voltage. Because magnetic fields can store massive amounts of energy, boost converters can easily deliver amps of current (e.g., stepping 3.7V from a LiPo up to 12V at 2A to drive an LED strip). The trade-off is the physical size and cost of the inductor, and the high-frequency magnetic EMI it radiates, which can interfere with nearby RF antennas or sensitive ADC readings.
A voltage multiplier, conversely, relies on the electric field inside capacitors. Capacitors are physically smaller, cheaper, and do not radiate magnetic EMI. However, the energy density of a capacitor is vastly lower than that of an inductor. Furthermore, the output impedance of a multiplier is inversely proportional to the switching frequency and the capacitance value. To get 1A of current out of a capacitor-based multiplier, you would need impractically massive capacitors and switching frequencies that would melt the MOSFET gates due to switching losses.
Choose a Voltage Multiplier when: Your load requires less than 50mA, you need to invert a voltage, you have strict EMI constraints (e.g., near a 2.4GHz WiFi antenna), or you need to generate voltages above 100V at microamp currents.
Choose an Inductive Boost Converter when: Your load requires more than 100mA, you need high efficiency (>90%) at moderate power levels, or your input-to-output voltage ratio is non-integer (e.g., 3.3V to 5.1V).
A Critical Note on Capacitor Selection (The DC Bias Trap)
When building or repairing a voltage multiplier, never blindly substitute capacitors based solely on their printed microfarad value. Multipliers rely on precise charge transfer. If you use Y5V or Z5U dielectric ceramic capacitors, you will run into the DC bias effect: a 10µF Y5V capacitor subjected to 10V of DC bias will actually lose up to 70% of its capacitance, dropping to just 3µF. This will cause your multiplier's output impedance to skyrocket and the voltage to collapse under load.
Always specify X7R or X5R multilayer ceramic capacitors (MLCCs) for the flying and reservoir capacitors in a charge pump. While X7R caps still exhibit some DC bias derating, it is far more predictable and stable across temperature ranges, ensuring your 5V to 18V quadrupler actually delivers the 18.2V you calculated on the bench.
