The total capacitance in series is always less than the smallest individual capacitor in the string. For two capacitors, the formula is \(C_{total} = (C_1 \times C_2) / (C_1 + C_2)\). For three or more, you use the reciprocal sum: \(1/C_{total} = 1/C_1 + 1/C_2 + ... + 1/C_n\). If you place two identical 100µF capacitors in series, the total capacitance drops to exactly 50µF, while the voltage rating doubles.
This counterintuitive behavior—where adding components reduces the overall value—is the exact opposite of series resistors. Below, we break down the topology, walk through a high-voltage design using real component values, and detail the exact failure modes you will encounter on the bench.
The Series Capacitor Topology: Nodes, Math, and Behavior
In a series configuration, capacitors are connected end-to-end, forming a single path for charge displacement. Let us define a basic two-capacitor string with three distinct nodes:
- Node A (Input): The positive terminal of \(C_1\), connected to the voltage source.
- Node B (Junction): The electrical connection between the negative terminal of \(C_1\) and the positive terminal of \(C_2\). This node is floating with respect to ground and holds the intermediate voltage.
- Node C (Output/Ground): The negative terminal of \(C_2\), connected to the return path.
Because the same charging current flows through the entire string, the charge \(Q\) stored on each capacitor is identical (\(Q = C_1V_1 = C_2V_2\)). However, the voltage divides inversely proportional to the capacitance. The smaller capacitor absorbs the larger share of the voltage drop.
Behavior Matrix: What Changes When One Element Shifts?
| Event at \(C_1\) | Effect on Total Capacitance | Effect on Node B Voltage | System Consequence |
|---|---|---|---|
| \(C_1\) value decreases | Decreases | Increases (takes more voltage) | Risk of overvoltage on \(C_1\) if near its rating. |
| \(C_1\) shorts internally | Becomes exactly \(C_2\) | Drops to Node A voltage | \(C_2\) now takes 100% of the source voltage; likely catastrophic failure if \(V_{source} > V_{rating(C2)}\). |
| \(C_1\) opens internally | Drops to ~0 (stray pF) | Floats / Undefined | Circuit is broken; no DC blocking or energy storage occurs. |
| \(C_1\) leakage current increases | No change in nominal C | Drops (voltage bleeds off) | \(C_2\) takes excess voltage; high risk of dielectric breakdown in electrolytics. |
Series vs. Parallel: Why Choose Series Capacitance?
Why wire capacitors in series when it reduces your total capacitance? The primary reason is voltage stacking. When designing power supplies for tube amplifiers, motor drives, or defibrillators, you often encounter rail voltages that exceed the maximum rating of commercially available, reasonably priced capacitors.
Rule of Thumb: Use parallel to increase capacitance at a fixed voltage. Use series to increase the voltage rating at the cost of capacitance.
If you need 50µF at 400V, you could buy a specialized 50µF 450V snap-in capacitor, which might cost $15 and have a 12-week lead time. Alternatively, you can use two standard 100µF 250V capacitors (costing $2 each, readily available from SparkFun or DigiKey) in series. The series combination yields 50µF and theoretically withstands 500V.
However, series wiring introduces a critical vulnerability: leakage current mismatch. Real-world electrolytic capacitors act like imperfect capacitors with a high-value resistor in parallel. If \(C_1\) has higher leakage than \(C_2\), the DC voltage will not split evenly. \(C_2\) will absorb more than half the total voltage, potentially exceeding its 250V rating and venting explosively. This is why series electrolytics must be paired with voltage-balancing resistors.
Design Walkthrough: Building a 400V Filter Bank
Let us design a safe, reliable series capacitor bank for a 350V DC rail. We need approximately 50µF of filter capacitance.
Component Selection:
- Capacitors: Two Panasonic EEUEE2E101 (100µF, 250V, radial electrolytic).
- Balancing Resistors: Two 220kΩ, 1W metal film resistors (e.g., Vishay PR01 series).
Calculating the Balancing Resistors:
The resistors must bleed off enough current to swamp the capacitor's internal leakage current, forcing the voltage to divide evenly based on the resistor ratio (which is tightly matched at 1% or 5%) rather than the unpredictable leakage.
- Estimate maximum leakage current for the 100µF 250V cap. The standard formula is \(I_{leak} = 0.01 \times C \times V\) or 3µA, whichever is greater. Here, \(0.01 \times 100 \times 250 = 250\mu A\).
- The bleeder current through the resistors should be at least 5 to 10 times the maximum leakage current. Let us target ~1mA of bleeder current per resistor.
- Using Ohm's Law for half the rail voltage (175V per cap): \(R = V / I = 175V / 0.001A = 175k\Omega\).
- We select the next standard E24 value up: 220kΩ. This yields a bleeder current of ~0.8mA, safely above the 250µA leakage.
- Check power dissipation: \(P = V^2 / R = 175^2 / 220,000 = 0.139W\). A standard 0.25W resistor is technically fine, but we specify 1W resistors for thermal headroom and long-term reliability inside a hot chassis.
Wire each 220kΩ resistor in parallel with its respective capacitor. This guarantees that Node B sits at exactly half the total DC voltage, protecting both dielectrics. For a deeper theoretical breakdown of series networks, refer to the All About Circuits DC textbook.
Breadboard Testing: Step-by-Step Verification
Before soldering this into a high-voltage chassis, verify the topology and balancing on a breadboard using a low-voltage DC source (e.g., a 9V battery or bench supply set to 12V). Never breadboard high voltage.
- Discharge: Short the leads of both capacitors with a 1kΩ resistor to ensure they start at 0V. Never short them directly with a screwdriver, as the high di/dt can damage the internal foil connections.
- Insert Components: Place \(C_1\) and \(C_2\) on the breadboard. Connect the negative lead of \(C_1\) to the positive lead of \(C_2\) (Node B). Watch your polarity.
- Add Resistors: Plug the 220kΩ balancing resistors into the same breadboard rows as the capacitor leads, ensuring one resistor spans \(C_1\) and the other spans \(C_2\).
- Apply Low Voltage: Connect +12V to Node A and GND to Node C.
- Measure Node B: Use a multimeter (like a Fluke 87V) to measure the DC voltage at Node B relative to ground. It should read exactly 6.0V (±0.2V). If it reads 8V or 2V, your resistors are not connected properly, or one capacitor has a massive internal short.
- Measure Total Capacitance: Power down, discharge the bank through a 1kΩ resistor, and switch your multimeter to capacitance mode. Probe Node A and Node C. You should read approximately 50µF. (Note: Electronics Tutorials notes that cheap multimeters may struggle to read series capacitance accurately if the equivalent series resistance (ESR) is high; an LCR meter is preferred).
Failure Modes: What Breaks at the Extremes?
Understanding how series strings fail is critical for designing protective circuitry.
The Short Circuit Extreme
If \(C_1\) suffers a dielectric puncture and shorts internally, Node A and Node B become electrically identical. The total capacitance of the string instantly becomes the value of \(C_2\) (100µF). More dangerously, \(C_2\) is now exposed to the entire source voltage. If this is a 350V rail and \(C_2\) is rated for 250V, \(C_2\) will rapidly overheat, vent electrolyte, and potentially rupture. This cascading failure mode is why series strings in critical applications often include overvoltage crowbar circuits or individual fuses.
The Open Circuit Extreme
If a lead breaks or an internal foil connection opens at \(C_1\), the series path is severed. The total capacitance drops to zero (practically, it drops to the stray parasitic capacitance of the PCB traces, usually less than 2pF). In a DC blocking application, the signal path is dead. In a power filter, the downstream circuit loses all energy storage, leading to massive voltage ripple and immediate brownout of the load.
Frequently Asked Questions
Does total capacitance in series increase or decrease?
It always decreases. Adding capacitors in series increases the total distance between the effective outer plates (the dielectric thickness adds up), which inversely reduces the capacitance. The total value will always be lower than the smallest single capacitor in the chain.
How do you find total capacitance in series with different values?
Use the reciprocal formula: \(1/C_{total} = 1/C_1 + 1/C_2 + 1/C_3\). For example, if you put a 10µF and a 40µF capacitor in series, the math is \(1/10 + 1/40 = 4/40 + 1/40 = 5/40\). Inverting that fraction gives \(40/5 = 8\mu F\). Notice that 8µF is less than the smallest capacitor (10µF).
Can I mix ceramic and electrolytic capacitors in series?
Technically yes, but it is highly discouraged in DC applications. Ceramic capacitors (like X7R or C0G) have virtually zero DC leakage current, while electrolytics have significant leakage. If placed in series across a DC voltage, the electrolytic will slowly leak, shifting the entire DC voltage burden onto the ceramic capacitor. Furthermore, ceramics exhibit severe capacitance loss under DC bias (a 10µF X7R might act like 2µF at 25V), making the total series capacitance unpredictable.
Why do my series capacitors measure lower than the calculated value?
If your calculated value is 50µF but your multimeter reads 42µF, you are encountering component tolerance and Equivalent Series Resistance (ESR). Electrolytic capacitors typically have a tolerance of -20% / +80%. Additionally, basic multimeters measure capacitance by timing an RC charge curve; the added ESR of two capacitors and breadboard contact resistance slows the charge rate, causing the meter to under-report the value. For precise readings, use a dedicated LCR meter at 120Hz.
