Why Your Pool Perimeter Lights Are Flickering—and Why Blaming the LEDs Is a Mistake
You’ve seen it: low-voltage perimeter lights along a 30-foot gunite pool edge, installed by a reputable builder, dimming toward the far end. Or worse—intermittent flicker in damp weather, even after swapping drivers and checking connections. The service call log reads “LED failure.” But I’ve pulled apart six such systems this year. Not one had a defective fixture. Every single issue traced back to the transformer—and specifically, to using 12V DC where 12V AC was required by code and physics.
Let’s cut through the confusion. This isn’t about “AC vs. DC preference.” It’s about NEC 680.23(B)(2), voltage drop under load in wet locations, and how LED drivers actually behave—not how datasheets pretend they behave.
The Code Isn’t Suggestive—It’s Structural
NEC 680.23(B)(2) states: “Low-voltage lighting systems installed within 5 ft of the inside wall of a pool… shall be supplied by a transformer that is listed for swimming pool use and rated not more than 15 volts.” That “listed for swimming pool use” clause points directly to UL 1838. And UL 1838 doesn’t list DC power supplies for underwater-adjacent applications—not unless they’re part of a fully isolated, double-encapsulated Class 2 system with ground-fault monitoring. Which most 12V DC “pool transformers” are not.
I’ve reviewed 17 UL 1838-certified transformers in the past 18 months. Sixteen are AC output. One is AC/DC switchable—but only when configured for AC output does it carry the full UL 1838 pool listing. The DC mode drops to UL 1310 (general-purpose Class 2), voiding compliance for pool-perimeter use.
This isn’t bureaucratic nitpicking. UL 1838 requires specific dielectric withstand testing at 1,500 VAC for 60 seconds between primary and secondary windings. DC supplies—even high-quality ones—don’t undergo that test because their isolation architecture (typically opto-coupled or transformer-isolated switching topologies) isn’t designed for sustained AC stress across wet-location creepage distances. In humid, chlorinated air, that difference becomes a leakage path—not immediately dangerous, but enough to destabilize constant-current LED drivers.
RMS Voltage Isn’t a Detail—It’s the Operating Condition
Here’s where installers misread spec sheets: “12V output” on a DC supply means 12.0 VDC ±5%. On an AC transformer, “12V” means 12 VRMS. That’s not the same as peak voltage. A 12 VRMS sine wave peaks at ~17 V. Most constant-current LED drivers used in perimeter lighting (e.g., Mean Well LCM-40, Inventronics EUD-40) require a minimum input voltage of 10.5–11.5 V to sustain regulation. They’ll accept up to 24 V input—but only if it’s DC.
So why does AC work better here? Because the driver’s internal rectification stage sees the full sine wave. Even with 20% voltage drop on a long feeder run (say, from 12.0 VRMS down to 9.6 VRMS), the peak remains ~13.6 V—still above the driver’s dropout threshold. A 12V DC supply dropping to 9.6 VDC collapses entirely below regulation. The driver shuts off, recovers, cycles—hence the flicker you see at the far end of a 100-ft circuit.
I tested this side-by-side on a 42-ft perimeter loop feeding twelve 3W LED bollards (36W total load). With a 12V DC switching supply (rated 5A), voltage at the last fixture measured 9.2 VDC at dusk—when ambient temperature dropped and conductor resistance increased. Driver dropout occurred at 10.1 VDC. With a toroidal 12V AC transformer (same VA rating), voltage at the same point was 9.8 VRMS—but drivers remained stable. Why? Peak voltage stayed at ~13.9 V. Regulation held.
Impedance Matching Matters More Than Wattage Ratings
Most installers size transformers on wattage alone: “I need 40W, so I’ll use a 60W transformer.” That’s insufficient—and dangerous near water. UL 1838 requires transformers to limit short-circuit current to ≤210% of rated output current for ≥3 seconds during fault conditions. Achieving that demands controlled secondary impedance.
AC transformers achieve this inherently via winding resistance and magnetic leakage reactance. A typical pool-rated 12V AC toroidal transformer (e.g., 100VA) has a short-circuit current of ~12.5A—just under the 13.2A ceiling for a 6.3A rated output. Its impedance is ~8–10%—enough to self-limit fault current without tripping upstream breakers prematurely.
DC switching supplies? Their short-circuit response depends on control-loop speed and MOSFET current limiting. Many budget units simply fold back to zero current in under 100ms—or latch off. That’s fine for landscape lighting. Near a pool, it creates an unmonitored open-circuit condition where a ground fault could go undetected until someone touches a wet fixture housing.
Worse: long feeder runs exacerbate the problem. A 100-ft 12 AWG copper run adds ~0.32Ω one-way. At 5A, that’s 1.6V drop—manageable on AC, catastrophic on marginal DC. But impedance mismatch also distorts waveform fidelity. Low-impedance DC supplies feeding high-impedance LED loads can induce ringing and overshoot on startup—especially with PWM-dimming fixtures. We measured 28V spikes on a 12V DC line during dimmer ramp-up. That’s enough to degrade electrolytic capacitors in drivers over time.
Harmonics Aren’t Just for Substations
Switching DC supplies generate harmonic currents—third, fifth, seventh—on the primary side. In dry locations, those harmonics dissipate harmlessly. In wet, conductive environments (chlorinated water, damp concrete decks, metal coping), they couple capacitively into grounding systems. We logged elevated 3rd-harmonic current (up to 1.8A RMS) on the equipment grounding conductor of a DC-powered pool light ring. That current found parallel paths: through rebar mesh, then up a wet ladder handrail.
No, it didn’t trip GFCIs—harmonics fall outside their detection band. But it did create measurable touch voltage (1.4VAC) on the ladder during pump operation. Not hazardous—but a red flag. UL 1838 explicitly prohibits power supplies whose harmonic distortion exceeds 30% THD on the input side for pool applications. Most off-the-shelf 12V DC bricks exceed 45% THD at partial load.
AC transformers? Negligible harmonics. A properly loaded toroidal unit runs at <5% THD—even under 30% overload. The iron core smooths everything. No high-frequency noise. No capacitive coupling into wet surfaces.
What Transformer Specs Actually Matter (and What to Ignore)
Forget “max wattage” as your primary spec. Focus on these four:
- UL 1838 listing status: Must state “For Swimming Pool Use” verbatim—not just “Class 2” or “Pool Rated.” Check the label, not the box.
- Secondary impedance: Should be 7–12% for 12V AC units. Anything lower risks excessive fault current; higher causes unacceptable voltage sag. Toroidal units typically list this. Laminated EI cores rarely do—avoid them for pools.
- Regulation rating: Look for ≤5% regulation from no-load to full-load. Anything above 8% means poor voltage stability over distance. Most quality pool transformers specify this at 25°C ambient—verify it’s tested at 40°C, since pool equipment rooms get hot.
- Terminal configuration: Screw-type terminals rated for wet-location wire (THWN-2 or XHHW-2), not push-in connectors. I’ve seen three corrosion-related failures in two years from moisture ingress at spring-clamp terminals.
Also: avoid multi-tap secondaries labeled “12/13/14/15V.” Those taps exist for landscape lighting compensation—not pool safety. UL 1838 requires fixed 12.0V ±0.3V output. Variable taps violate listing.
A Real-World Example: The 32-Foot Vinyl-Liner Pool
Client: Municipal aquatic center. Perimeter: 32 ft × 16 ft rectangle. Fixtures: sixteen 2.5W warm-white LEDs, spaced 3 ft apart, mounted in stainless steel deck rings. Feeder: 10 AWG UF-B direct-buried, 85-ft total circuit length (looped).
Initial install: 12V DC, 75W switching supply. Result: last six fixtures cycled on/off at dusk. Voltage at far end: 8.9 VDC. Driver dropout threshold: 10.2 VDC.
Fix: UL 1838-listed 12V AC toroidal transformer, 160VA (13.3A rated), 9% impedance, 4.2% regulation. Re-ran feeders with 8 AWG THWN-2 in PVC conduit (per NEC 680.10). Final voltage at far fixture: 10.1 VRMS (14.3 V peak). All drivers regulate cleanly. Zero flicker. Touch voltage on deck rings: <0.15 VAC, verified with Fluke 1587 FC.
This works because AC voltage drop is less destructive to LED driver operation—and because UL 1838 compliance ensures fault behavior aligns with wet-location risk models. It falls flat because DC “convenience” ignores how real-world conductors, drivers, and humidity interact.
If your perimeter lighting feels like a troubleshooting treadmill—check the transformer first. Not the bulbs. Not the wiring. The transformer. And ask for the UL 1838 label photo before it ships.