Capacitor leakage is the small, unwanted DC current that sneaks through a capacitor's dielectric material when a voltage is applied. While modern ceramic and film capacitors exhibit leakage in the picoamp range (effectively zero for most hobbyist work), aluminum electrolytic capacitors are notorious for measurable leakage. If you are repairing vintage audio gear, rebuilding a linear power supply, or sorting through a bin of salvaged components, knowing how to accurately measure this leakage is the difference between a reliable build and a smoking failure.
This guide gives you the exact numerical targets, the bench setup required to measure microamps safely, and a hard decision tree for when to keep, reform, or trash a component.
The Direct Answer: What is a "Good" Leakage Reading?
A "good" reading is any value at or below the manufacturer's specified maximum leakage current after a standard settling time (usually 2 to 5 minutes). For standard aluminum electrolytic capacitors, the industry rule of thumb formula is:
Ileak ≤ 0.01 × C × V (or 3 µA, whichever is greater)
Where C is capacitance in µF, and V is the rated DC voltage.
Worked Numeric Example:
You are testing a 470 µF, 50V electrolytic capacitor.
Calculation: 0.01 × 470 × 50 = 235 µA.
Therefore, a reading of 235 µA or less after 5 minutes of applied voltage is a pass. If your meter reads 450 µA and climbing, the dielectric is degraded.
For high-reliability or "low-leakage" series capacitors (like the Nichicon UHE or Panasonic FR lines), the multiplier drops to 0.002 or 0.003. Always check the specific datasheet, but the 0.01CV formula is your reliable baseline for general-purpose parts.
Meter Setup and Bench Configuration
You cannot measure capacitor leakage accurately by just slapping a multimeter in resistance mode across the leads. You must apply the rated DC voltage and measure the resulting current in series.
If you are probing a 400V bulk capacitor in a switched-mode power supply to verify it is discharged before removal, your multimeter must be CAT III 1000V or CAT IV 600V rated (e.g., Fluke 87V) to protect against transient spikes. However, the actual leakage measurement described below is strictly an out-of-circuit bench test using a low-voltage DC supply. Never attempt to measure microamp leakage in-circuit on live mains-derived boards.
DMM Configuration
- Meter: A true-RMS DMM with a dedicated microamp (µA) range (e.g., Fluke 87V, Brymen BM235).
- Dial Position: Set to the µA/mA position, then press the range/function button to select DC µA.
- Lead Jacks: Black lead in COM. Red lead in the µA/mA jack. Never use the 10A jack; its resolution is too low and will just read 0.00 for leakage currents.
The Test Circuit
Connecting a DMM directly in series between a power supply and an empty capacitor is a fast way to blow your meter's internal µA fuse due to inrush current. You must build a simple test rig:
- DC Power Supply: Set to the capacitor's rated voltage (e.g., 25V for a 25V cap).
- Current-Limiting Resistor: A 1 kΩ, 1/4W resistor in series with the positive supply line. This limits inrush to safe levels while the cap charges.
- DMM: Placed in series after the current-limiting resistor, measuring current flowing into the capacitor's positive lead.
Step-by-Step Measurement Procedure
- Discharge and Isolate: Ensure the capacitor is fully discharged using a bleed resistor (e.g., 100Ω 5W) and removed from the circuit. Out-of-circuit testing is mandatory; parallel PCB traces will ruin your reading.
- Wire the Rig: Connect the DC supply positive → 1 kΩ resistor → DMM Red lead. DMM Black lead → Capacitor Positive (+). Capacitor Negative (-) → DC supply ground.
- Apply Power: Turn on the bench supply. The DMM will initially spike (charging current) and then begin dropping rapidly.
- The Waiting Game: Start a timer. You must wait a minimum of 2 minutes (5 minutes is better for large >1000µF caps). This allows the dielectric absorption to settle.
- Record the Reading: At the 5-minute mark, read the DC µA value on the display. This is your true leakage current.
Expected Readings: Good vs. Bad Values
Use this spec-sheet-style reference table to evaluate your 5-minute readings based on common capacitor sizes.
| Capacitor Spec | Calculated Max (0.01CV) | Good Reading (Pass) | Bad Reading (Fail) |
|---|---|---|---|
| 100 µF, 16V | 16 µA | < 16 µA | > 25 µA |
| 470 µF, 25V | 117.5 µA | < 117 µA | > 150 µA |
| 1000 µF, 50V | 500 µA | < 500 µA | > 750 µA |
| 4700 µF, 35V | 1645 µA (1.6 mA) | < 1.6 mA | > 2.5 mA |
Note: If your calculated max is less than 3 µA (e.g., a 10µF 16V cap), the pass threshold defaults to 3 µA.
Common Mistakes That Skew Readings
According to technical guidelines from manufacturers like Nichicon and testing primers on All About Circuits, bench errors are the primary cause of false failures.
When voltage is first applied, the capacitor draws heavy charging current, followed by "dielectric absorption" current as the internal chemical layer aligns. If you read the meter at 10 seconds, you are measuring absorption, not leakage. Fix: Strictly enforce the 5-minute timer.
If you hold the capacitor body or touch the bare metal probe tips with your fingers while taking the final reading, your body's resistance (roughly 10kΩ to 100kΩ depending on skin moisture) creates a parallel path. The meter will read your body's leakage, not the capacitor's. Fix: Use alligator clips or PCB test hooks, and step back from the bench when recording the final number.
An ESR meter (like the Atlas ESR70) measures Equivalent Series Resistance at high frequency (usually 100 kHz). It is fantastic for finding dried-out, high-ESR capacitors in switching power supplies, but it cannot measure DC dielectric leakage. A cap can have perfect ESR but terrible DC leakage. Use the right tool for the right job.
Decision Tree: Keep, Reform, or Trash?
Stop guessing. Use this exact decision path based on your 5-minute µA reading to determine the component's fate.
| Condition (at 5 mins) | Verdict | Action Required |
|---|---|---|
| Ileak ≤ Calculated Max | KEEP | Capacitor is healthy. Install in circuit. |
| Ileak is 2x to 5x Max (but dropping slowly) | REFORM | Dielectric is weak but recoverable. Connect cap to bench supply set to rated voltage via a 10 kΩ, 2W resistor. Leave for 2 to 4 hours. Re-test. If it passes, keep. |
| Ileak is > 5x Max OR reading is climbing | TRASH | Dielectric breakdown is imminent. Throw it away. Default Replacement: Buy a modern, low-ESR 105°C rated equivalent, such as the Panasonic FR series or Nichicon PW series. |
| Physical bulging, venting, or leaked electrolyte | TRASH | Do not test. Internal pressure has already compromised the foil. Clean the PCB pads with isopropyl alcohol and replace. |
By adhering to the 0.01CV formula, enforcing the 5-minute settling time, and utilizing a current-limited bench setup, you eliminate the guesswork from capacitor salvage and repair. When in doubt, follow the decision tree: if it fails the math and won't reform, bin it and drop in a fresh Panasonic FR.
References: For deeper reading on multimeter safety categories during initial circuit discharge, consult the Fluke CAT Ratings Guide. For manufacturer-specific leakage multipliers, refer to the Nichicon Aluminum Electrolytic Technical Guidelines.
