A tingling enclosure is usually not an electric shock. It is a leakage-current path that the mechanical design cut in half. One of our charger projects came back from a client during the southern China plum-rain season: users reported a sharp prickling sensation when gripping the enclosure. Every whole-unit metric passed. Withstand voltage passed, insulation resistance was normal, leakage current was within limits. But the user's hand does not lie. It took two days of teardown to find the problem, and it lived in the safety clearance between a mounting screw and a Y-capacitor. This is not a rare failure, but it took a long time to localise, so it is worth writing down.
The direct answer: the root cause is that the Y-capacitor couples high-frequency interference onto the metal enclosure, and there is no effective ground bleed path, so a displacement current forms when a hand touches it. The common structural break points are a screw mounting hole sitting directly over high-voltage copper on the PCB, a Y-capacitor area with no insulating pad, and a missing creepage relief slot that IEC 62368 would require.
Pitfall One: Every Whole-Unit Test Reads "Pass"
We retested the returned charger in the lab. The hipot tester put 3750 V across it for a minute with no breakdown. The insulation resistance meter read above 50 MΩ from enclosure to line and neutral. The earth resistance read within specification on a milliohm meter.
Every indicator green, yet the surface genuinely tingles.
This is the first pitfall: whole-unit electrical testing masks local structural problems. The fixture clamps the enclosure earth point and the line/neutral terminals, so it measures the impedance of "the entire enclosure" against "the entire input". But the spot a hand grips is one region of the enclosure, not the whole thing. If the leakage path is interrupted locally, a whole-unit measurement simply cannot see it.
Structural checkpoints for grounding design: (1) metal-enclosure mounting screws must not sit directly above high-voltage copper on the PCB, keep at least 3 mm of clearance; (2) add an insulating pad around the Y-capacitor, rated for at least 130 °C and unable to shift under screw torque; (3) size the creepage distance between the mains input area and the metal enclosure against IEC 62368 Table 5-4, which asks for 2.5 mm minimum at basic insulation below 400 V peak, and 5 mm for reinforced insulation.
Feeling Along the Enclosure by Hand: the Tingle Is Only on One Side
Instruments could not find it, so we located it by feel. We had an engineer press a palm against the enclosure and work downward in zones. The result was unambiguous: the lower half of the enclosure on the plug-facing side tingled clearly, while the battery-compartment side was almost imperceptible.
Opening the enclosure showed why. The PCB sits crosswise in the shell, and at the end nearest the tingling zone, a mounting screw hole sits directly over the switching supply's Y-capacitor. The Y-capacitor (a Y2 220 pF part) is soldered between primary and secondary ground, and one of its legs sits about 1.2 mm from the metal enclosure. When the screw was torqued down, the insulating pad was compressed and crept, and the pad edge rolled into the gap between the capacitor leg and the enclosure. The actual clearance between leg and enclosure was squeezed to under 0.5 mm.
Why does under 0.5 mm cause tingling? A Y-capacitor exists to shunt high-frequency common-mode interference to ground, and it inherently carries a microamp-level AC leakage current. That current is tolerated by design, and it is fine as long as the bleed path to ground is unobstructed. But when the capacitor leg sits too close to the metal enclosure, leakage couples straight through the air gap onto the shell. A two-prong plug adapter is not earthed by design, so charge accumulates on the metal surface with nowhere to go. When a hand touches it, the contact area is large and the resistance low, so the current takes the path through the hand.
Checking the standards: IEC 62368 requires 2.5 mm of creepage between a metal enclosure and hazardous live parts for basic insulation. Our internal control generally keeps 3 mm or more, but the variable of a pad shifting during pilot assembly is not something routine inspection covers.
Root-Cause Criteria: These Three Signs Together and the Tingling Is Almost Certain
Looking back across this class of problem, we now have a set of criteria for the structural design stage. There are three signs.
Sign one: the Y-capacitor is close to the metal enclosure with no rigid isolation. The Y-capacitor is the main source of leakage current. If there is metal enclosure within 20 mm of it, the legs need insulation, or a relief slot must be cut to lengthen the creepage path. Our relief slots are at least 2 mm wide and at least 1 mm deep, in V-0 flame-retardant PC.
Sign two: the mounting screw hole falls inside the projection of high-voltage copper. If a screw hole lands directly beneath the switching transformer, bridge rectifier or Y-capacitor traces, the downward force of tightening changes the contact pressure between pad and PCB, compressing a clearance that was correctly sized on the drawing. Nothing shows up on the structural drawing, but production yield deserves a question mark.
Sign three: humidity visibly affects the tingling. Clients complained strongly in the plum-rain season and we received essentially no such complaints in winter. Above 70 percent humidity the breakdown voltage of an air gap drops, so a clearance that suffices when dry will leak when damp.
Hit all three at once and you can call it without disassembling anything: the ground bleed path is broken.
When we fixed that charger, we added a cross-shaped insulating standoff to the Y-capacitor legs, keyed onto an enclosure rib, and ran assembly validation on the screw torque specification and pad slippage. The relief slot was cut into the enclosure tool up front, sized to the 5 mm creepage of reinforced insulation. The revised pilot units then went through 48 hours of accelerated ageing at 85 percent humidity in an environmental chamber. Enclosure-to-earth voltage dropped from tens of volts to below 1 V, and the surface felt completely inert to the hand.
After this case we added a line to our internal structural design standard: on any plastic-shell product with a switching supply, no metal screw hole is permitted within 10 mm of a Y-capacitor. Where that is genuinely unavoidable, an isolating shim plus assembly validation is mandatory. A clearance annotation on a drawing is not sufficient.
FAQ: Enclosure Tingling and Grounding Design
Q: If an enclosure tingles, is it definitely a leakage fault?
A: Usually not. It is displacement current coupled to the enclosure through the Y-capacitor with no ground bleed path. The magnitude is typically microamps to milliamps, below the dangerous-shock threshold but perceptible. To tell them apart: the tingling is continuous and broad (across the whole palm), there is no stabbing pain or muscle contraction, and the symptom eases when you break contact between your feet and the ground by wearing insulating shoes. That pattern points to capacitively coupled displacement current rather than a real shock.
Q: Can I just remove the Y-capacitor to stop the tingling?
A: We would not advise it. The Y-capacitor is required for switching-supply EMC design; removing it pushes radiated and conducted interference over the limits and the product fails certification. The correct fix is to guarantee adequate insulation distance between the Y-capacitor bleed path and the metal enclosure: add a rigid insulating shim or pot it with insulating compound, and route the bleed trace straight back to primary ground rather than using the metal enclosure as the return path. Choose PC, PET or FR-4 rated to match the product's safety requirements, at least 0.5 mm thick.
Q: How do we prevent this at the structural design stage?
A: Three control points. No metal screw holes within 10 mm of the switching supply. At least 3 mm between Y-capacitor legs and the metal enclosure, via a relief slot. And anti-slip insulating pads on every mounting screw, with assembly validation. Calculate the relief slot creepage against IEC 62368 Table 5-4 using the working-voltage peak. For a 220 V AC input at pollution degree 2, that is 2.5 mm for basic insulation and 5 mm for reinforced. We also recommend an 85 percent humidity / 40 °C environmental chamber test during development. Screening for damp-environment tingling is far more effective than running electrical performance tests on the whole unit alone.
Written by the Hezi Industrial Design team in Dongguan. We do product appearance design, structural design and CMF for hardware teams shipping at volume.
Top comments (0)