It's tempting to think you just touch the multimeter leads to a capacitor and watch the reading. That's what I thought in my first year (2017). I was on a Schneider Electric APC UPS unit that kept tripping the output breaker — a typical failing capacitor symptom. I touched the probes, saw a number that looked okay in the ballpark, and signed off on the repair. Two days later, the same UPS went down. Hard. The replacement logic board alone cost me $320. And the embarrassment of explaining to the client why they had a second outage.
The problem wasn't the equipment. The problem was I didn't know which test to run, and I didn't understand that a capacitor can pass one test and fail another. Here's what I've learned since then, documented in this field guide to testing capacitors with a multimeter. It's broken down by the situation you're likely in, because there is no universal "test" that catches everything.
Why "just test the capacitance" advice ignores the real problem
The standard advice you'll find online is simple: set your multimeter to capacitance mode (if it has one), discharge the capacitor, connect the probes, and read the value. If it's within 10% of the rated value, it's good. If not, replace it.
That advice is correct — and dangerously incomplete. It assumes you have a multimeter with a capacitance mode (many cheaper ones don't), and it assumes the capacitor has failed in the one way that shows up as a capacitance shift. In my experience, that's maybe 40% of failures. The other 60%? They fail as high ESR (Equivalent Series Resistance), as a short, or as an open circuit — none of which a capacitance reading alone will catch. It's tempting to think you can just compare the reading to the label. But identical capacitance from a bulging cap can result in wildly different outcomes, like the $320 saga I started with.
Scenario A: You have a multimeter with a capacitance mode (and you're checking general health)
This is the most common scenario for anyone working on Schneider Electric drives, PLC power supplies, or industrial sensors. Most modern mid-range multimeters (Fluke 17B+, Klein MM700, etc.) have this. Here's the mistake that cost me a data center's HVAC controller board in 2022:
I tested a 470µF electrolytic capacitor on the board's power rail. Capacitance reading: 462µF. Perfect, right? I put the board back in. It died again within the same week. The problem was ESR. The capacitor had internal resistance that had doubled from its normal spec, causing ripple voltage that killed the downstream logic. The capacitance test didn't catch it. You need to pair a capacitance test with a visual inspection and, if possible, an ESR test.
The right approach for this scenario:
First, discharge the capacitor safely. Use a high-wattage resistor (like a 10kΩ, 5W) across the terminals for 10 seconds. Don't short the terminals with a screwdriver — that can damage the capacitor or worse. After discharge, set your meter to capacitance mode (often marked with a "-|(-" symbol). Connect the probes. Black to negative (usually marked with a stripe on electrolytics), red to positive. Wait for the reading to stabilize — it takes a couple of seconds for larger caps.
What most people don't realize is that if you just measure the capacitance:
- A shift of more than 15% from the rated value means replace it. Done.
- A reading within 10% means it might be good — but you need to also check for bulging at the top or leaking electrolyte. If it looks physically perfect and capacitance checks out, you're probably okay for non-critical circuits (like a simple decoupling cap).
- A zero or OL (open line) means it's an open circuit. Trash it.
For critical stuff — UPS power supplies, drive DC bus capacitors — a clean capacitance reading is not enough. You need the next scenario.
Scenario B: You only have a multimeter with resistance mode (and no capacitance mode)
Here's something vendors won't tell you: a simple resistance test, while less precise than a dedicated capacitance test, can catch failures that the capacitance test would miss. It's also the only option if you have a basic $20 multimeter, which is what many field techs carry for quick checks.
I've been there. In June 2023, I was troubleshooting a Schneider Electric circuit breaker trip in a small manufacturing plant. The client had a simple clamp meter with continuity and resistance modes — no capacitance. The motor start capacitor looked normal. No bulging. No leaking.
I set the meter to 200kΩ resistance mode. After discharging the cap (critical step — skipping it here can blow your meter's fuse), I touched the probes across the terminals. On a good electrolytic capacitor, you'll see the resistance start low (say 10-20kΩ) and then climb slowly toward infinity as the capacitor charges up. That's the charging curve. The fact that you see the reading change tells you the cap is actually charging, which means the internal structure is intact.
In my case, the reading shot to OL immediately — no charging curve. That meant an open circuit. The cap was shot. Replaced it ($4 part), and the motor started fine. The whole diagnosis took 2 minutes.
What to look for with resistance mode:
- Low reading that climbs: The capacitor is charging. Likely good (but not confirmed for ESR).
- Stays at 0 (or near 0): Short circuit. Replace it.
- OL immediately with no transition: Open circuit. Replace it.
- Reads a stable, moderate value (like 10kΩ) and stays there: The cap is leaky. Replace it; it won't hold a charge.
This method isn't perfect — it won't catch all ESR failures — but it catches the two most common failure modes: opens and shorts. In my experience, that's about 70% of bad capacitors in industrial equipment.
Scenario C: You need to confirm a capacitor is dead or alive (quick field check)
Sometimes you're in a hurry, and you just need to know: is this capacitor the problem or not? The "charge and hold" test is your friend. I use this on data center UPS fan motor caps and sensor power supply caps when I need a decision in under 30 seconds.
First steps as always: discharge, discharge, discharge. Then, set your multimeter to a moderate resistance setting (200kΩ or 2MΩ). Connect the probes. Watch for the charging curve. Once the reading stabilizes (which means the cap is fully charged), quickly disconnect the probes and switch your meter to DC voltage mode. Remeasure the cap terminals. If you see any voltage — even 0.5V to 5V, depending on the cap size and how fast you switched modes — it means the cap held a charge. That's a strong indicator it's functioning.
The numbers said the capacitor was functional — it held 1.2V for about 15 seconds. My gut said it looked suspicious (a very slight, almost invisible bump on the top). I went with my gut and replaced it anyway. Turns out that bump was an early sign of internal failure. The next week, the sensor started giving erratic readings. I dodged a field failure by 7 days.
How to know which scenario applies to you
Here's my quick decision flow for anyone who's saved this article to reference later:
Do you have a meter with a capacitance mode?
→ Yes, and the capacitor is in a non-critical circuit: Use Scenario A. A good capacitance reading + a visual check = probably good.
→ Yes, but it's critical: Use Scenario A, then also check ESR if you have a dedicated tester. If not, at least pair it with the visual check and maybe a resistance quick test from Scenario B for redundancy.
Do you only have a basic resistance/continuity meter?
→ Use Scenario B for a "pass/fail" determination. It catches the big failures. If the resistance reading shows a good charging curve and the cap looks physically healthy, it's likely fine for most non-critical applications.
Do you need a quick yes/no in the field where you can't get to your bench?
→ Use Scenario C. The charge-and-hold test is fast and surprisingly reliable for catching completely dead caps. But don't use it as the sole check for something like a Schneider Electric drive's DC bus capacitor — those need a proper ESR test.
One more thing: when in doubt, replace it. I know, that sounds like a cop-out advice. But here's the reality from my $320 mistake: a $4 capacitor can save you a $320 logic board. Per FTC guidelines on substantiating claims, I'll be specific: on a Schneider Electric APC UPS with part number SURTD3000XLI, a replacement main bus capacitor (rated 470µF, 400V) costs around $12 from authorized distributors. A logic board replacement for the same unit costs $320. The math has never been about the capacitor cost — it's about the cost of being wrong.
And per USPS pricing effective January 2025, if you're sending capacitors for RMA or replacement: First-Class Mail large envelope (1 oz) is $1.50. An additional ounce for larger ones: $0.28. Just in case you're wondering about the economics of "should I test it or just replace it."
