The Short Answer: Farads and the Sub-Unit Reality
The base unit of capacitance is the Farad (F), named after Michael Faraday. By definition, one Farad represents the capacitance when one coulomb of charge causes a potential difference of one volt. However, in practical electronics and electrical work, a one-Farad capacitor is physically massive—roughly the size of a D-cell battery or larger, typically reserved for supercapacitor memory backup or extreme audio applications.
Because the Farad is so large, we almost exclusively work with sub-units. When you ask 'what unit is capacitance measured in' on a bench or jobsite, the real answer depends on the circuit domain:
- Microfarads (µF or mfd): $10^{-6}$ Farads. The standard unit for power supply filtering, motor run/start circuits, and audio crossovers. (Typical range: 1 µF to 10,000 µF).
- Nanofarads (nF): $10^{-9}$ Farads. Common in timing circuits, EMI filtering, and snubber networks. (Typical range: 1 nF to 999 nF).
- Picofarads (pF): $10^{-12}$ Farads. Used in RF oscillators, high-frequency bypassing, and precision timing. (Typical range: 1 pF to 999 pF).
Meter Setup and Safety Categories (CAT Ratings)
Before you touch a component, your digital multimeter (DMM) or dedicated LCR meter must be configured correctly. Measuring capacitance requires the meter to output a small DC test voltage to charge the component and measure the time constant. Here is the exact setup for a standard professional DMM like the Fluke 87V or Brymen BM235.
Meter Configuration Block
- Dial Position: Rotate the dial to the capacitance symbol (two parallel lines, one straight and one curved, or simply labeled 'CAP' / 'F').
- Lead Jacks: Black lead into COM. Red lead into the V/Ω/F (Voltage/Ohms/Farads) jack. Do not use the current (A or mA) jacks, as they contain internal shunt resistors that will skew the reading or blow the meter's fuse.
- Range Setting: Start in Auto-Range. If the meter 'hunts' (clicks repeatedly between ranges) on small values, press the RANGE button to manually lock it to the µF or nF scale.
- Zeroing: Touch the probe tips together and press the REL (Relative) or NULL button. This subtracts the inherent parasitic capacitance of your test leads (usually 50pF to 100pF), which is critical when measuring values under 1 nF.
Probe Placement and the Mandatory Discharge Step
Capacitors store energy. A 45 µF motor run capacitor charged to 240V AC (which peaks at ~340V DC) holds enough energy to destroy your meter's input protection and deliver a painful shock. Never measure capacitance on a live circuit, and never assume a capacitor is discharged just because the power is off.
- De-energize and Lockout: Turn off the breaker and verify zero voltage with your DMM in AC/DC voltage mode first.
- Discharge Safely: Use a 20kΩ, 5-watt wirewound resistor attached to insulated alligator clips. Place it across the capacitor terminals for 5 to 10 seconds. (Never use a flathead screwdriver to short the terminals; the instantaneous current spike can vaporize the screwdriver tip, damage the capacitor's internal dielectric, and trip the panel breaker).
- Isolate the Component: For electrolytic capacitors, desolder and lift at least one leg from the PCB. Measuring in-circuit will yield the combined capacitance of all parallel components on that rail.
- Probe Placement:
- Polarized (Electrolytic/Tantalum): Red probe to the Anode (+), Black probe to the Cathode (-). While some modern DMMs can read reverse-polarity, it introduces measurement error.
- Non-Polarized (Ceramic/Film/HVAC Run): Probe placement does not matter. Place one lead on each terminal.
Expected Readings: Good vs. Failing Capacitors
What does a good reading look like numerically? A healthy capacitor will read within its stated tolerance (usually ±5%, ±10%, or ±20%) of the printed nominal value. A failing capacitor will typically read low (dielectric drying out) or read as a dead short (dielectric puncture).
| Component Type | Nominal Value | Tolerance | Good Reading Range | Bad Reading (Replace) |
|---|---|---|---|---|
| HVAC Run Capacitor | 45 µF | ±6% | 42.3 µF to 47.7 µF | < 40 µF, > 50 µF, or 'OL' (Open) |
| Power Supply Filter | 1000 µF | ±20% | 800 µF to 1200 µF | < 750 µF (High ESR/Dried out) |
| Ceramic Bypass | 100 nF (0.1 µF) | ±10% | 90 nF to 110 nF | 0.00 nF (Short) or 'OL' (Cracked) |
| Timing Capacitor | 10 pF | ±5% | 9.5 pF to 10.5 pF | Any drift > 11 pF |
Note on ESR: Standard DMMs measure total capacitance by timing a charge cycle. They cannot measure Equivalent Series Resistance (ESR). A 1000 µF capacitor might read 980 µF on your Fluke, but if its ESR has spiked from 0.05Ω to 2.0Ω due to heat degradation, it will fail under load. For power supply repair, you must use a dedicated ESR meter (like the MESR-100) to test in-circuit health.
Four Mistakes That Yield Misleading Readings
If your meter is giving you numbers that defy physics, you are likely falling victim to one of these bench errors:
- Body Capacitance (The pF Problem): When measuring small ceramic capacitors (e.g., 22 pF), holding the component between your fingers introduces your body's parasitic capacitance (often 50pF to 200pF) in parallel. Fix: Use a breadboard or PCB vise to hold the component, or use SMD test tweezers.
- In-Circuit Parallel Paths: Measuring a 100 nF capacitor while it is still soldered to a board might yield a reading of 4.7 µF. You aren't measuring the capacitor; you are measuring the entire decoupling network on that VCC rail. Fix: Always lift one leg for out-of-circuit verification.
- Dielectric Absorption: If you discharge a large film capacitor, remove the short, and wait a minute, it will 'rebound' and generate a few volts on its own. If you try to measure capacitance immediately after a rebound, the meter's internal test voltage will conflict with the rebound voltage, causing erratic readings. Fix: Discharge, wait 60 seconds, discharge again, then measure.
- Ignoring Test Lead Offset: If you measure a 10 pF capacitor without pressing the REL button to zero out your leads, your meter will read 85 pF (10 pF cap + 75 pF leads). Fix: Short the tips and press REL before every batch of sub-nanofarad measurements.
Decision Path: Selecting the Right Replacement
When your measurements confirm a capacitor is dead, you need a replacement. Do not guess. Follow this decision tree to select the exact part.
| Parameter | Decision Rule | Action |
|---|---|---|
| Capacitance (µF) | Must match the original nominal value exactly (within the original tolerance). | If original is 45 µF, you MUST buy 45 µF. Do not use 40 µF or 50 µF. |
| Voltage Rating | Must be EQUAL TO or GREATER THAN the original. | If original is 370V AC, you can use 370V or 440V. Never use 250V. |
| Physical Shape | Must fit the existing mounting bracket and spade terminals. | Check if the bracket requires Round or Oval case. Check terminal width (usually 1/4'). |
| Frequency/Type | Run vs. Start vs. DC Electrolytic. | Motor Run caps are continuous duty (film/foil). Never swap a Start cap (electrolytic) for a Run cap. |
The Concrete Pick: HVAC Motor Run Replacement
If you are replacing a standard 45 µF HVAC compressor/fan run capacitor that tested low (e.g., reading 32 µF on your meter), stop searching through generic assortments. You need a component with a proven track record for thermal stability and exact tolerance.
Buy this exact part: AmRad 45µF 440V Round Run Capacitor (Part# 45C440).
- Why this pick: AmRad uses heavy-duty film/foil construction rated for 440V (exceeding the standard 370V requirement, providing a larger safety margin against grid voltage spikes). It features a 1/4' spade terminal, a round case that fits 95% of standard HVAC disconnect brackets, and is filled with non-toxic dielectric fluid rather than flammable oils. It typically retails for $14 to $18 on HVAC supply sites, making it the definitive bench-and-jobsite standard for this exact failure mode.
For deeper theory on how dielectrics behave under AC stress, refer to the capacitor fundamentals chapter on All About Circuits, and for official safety protocols on testing, consult the Fluke guide on capacitance measurement.
