Check power polarity (the #1 cause)
Reversed polarity kills more strips than any other mistake. On 12V strips the + and โ pads are usually marked, but at the cut end or on a cheap connector they're easy to get backwards. If power was applied backwards even briefly, the strip's onboard resistors can blow โ and then the strip is genuinely dead and needs replacing.
How to test polarity safely
Before connecting anything, verify with a multimeter in DC voltage mode: measure across the strip's two input pads (or the connector's two wires). You should see the supply voltage with the correct sign โ around 12V for a 12V strip, 5V for a WS2812B. If you read the right voltage but the strip is dead, the strip itself is blown and swapping polarity won't bring it back.
Also check the supply isn't the problem: measure it with no load first, then with the strip connected. A supply that reads 12V unloaded but collapses under load is failing โ that's Fix 7 territory.
Verify the data signal (addressable strips)
For WS2812B-style addressable strips, "has power but no light" is almost always a missing or weak data signal. The strip is powered and ready, but nothing is telling the first LED what to do โ so the whole chain stays dark.
Common ground, then data
- Common ground first: strip GND, power supply GND, and ESP8266 GND must all be tied together. No common ground, no data reference โ the strip stays dark or glitches.
- Check the data pin: on an ESP8266, the sketch's data pin must match the physical wire. D1 (GPIO5) is the safest default; a mismatch gives you nothing.
- Signal strength: WS2812B expects a 5V data signal but the ESP8266 outputs 3.3V. Short runs usually work; past ~1m, add a 74HCT245 level shifter or a 5V pull-up on the data line.
- First LED in the chain: if the very first pixel is dead or its pads are bridged, everything after it is dark. Bypass the first LED with a jumper to test.
Common ground is the invisible wire
ESP8266 Power supply (5V)
โโโโโโโโโโโโโโ โโโโโโโโโโโโโโ
โ D1 โโโโโโโโโผโโโบ DIN โโโผโโโบ strip โ
โ GND โโโโโโโโผโโโโ โ โ
โโโโโโโโโโโโโโ โ โ 5V + โโโบ strip +
โโโโโโโโผโ GND โโโบ strip โ
โโโโโโโโโโโโโโ
All three GNDs tied together = data works
Missing any one GND = dark or glitching strip
Voltage drop & power injection
This is voltage drop along the strip's copper traces โ not broken LEDs. Each metre of strip acts like a small resistor, so the far end gets less voltage and the LEDs there run dimmer and shift colour (whites go pink, reds go orange). The fix is power injection: feed the supply voltage into the strip again at the midpoint or far end, every ~5m for dense strips.
You don't need to cut the strip to inject โ solder onto the copper pads at a cut line, or use a connector designed for it. Match the injection to the strip's needs: a 5m / 60-LED-per-metre WS2812B strip can pull around 18A at full white, which is why one end alone can't feed it.
If the strip is under ~2โ3m and running at moderate brightness, it usually runs fine off one end โ injection is only needed when the tail visibly suffers.
Bad solder / bad connector (after cutting)
Cutting a strip is where most connection faults happen. The copper pads at a cut line are small, and a cold solder joint, a bridge between adjacent pads, or a half-seated connector after re-joining will kill the segment โ or the whole strip if it's an addressable chain (data flows through every pixel).
- Reflow the pads with fresh solder โ a dull, lumpy joint is a cold joint.
- Check for bridges: a blob of solder across two pads shorts the data line to power or ground.
- Reseat connectors fully. The JST-style connectors on cheap strips are the weakest link; a connector that looks plugged in but isn't latched is a classic intermittent fault.
- After re-joining, test the segment in isolation before reconnecting the whole strip.
Wrong GPIO pin on ESP8266
Not every GPIO on an ESP8266 is safe to drive. Some pins are involved in booting the chip, and hanging an output on them can glitch at startup โ or stop the board from booting at all. If your strip worked in one sketch and not another, or fails intermittently, check which pin the sketch is actually driving.
Safe GPIO choices
| NodeMCU label | ESP8266 GPIO | Best use |
|---|---|---|
| D1 | GPIO5 | โ Safe โ relay or data line, my default |
| D2 | GPIO4 | โ Safe โ relay or data line |
| D5 | GPIO14 | โ Safe โ relay or data line |
| D6 | GPIO12 | โ Safe โ relay or data line |
| D4 | GPIO2 | โ ๏ธ Boot quirk โ needs a pull-up at boot; fine for short strips on known boards, avoid for relays |
| D3 / D8 | GPIO0 / GPIO15 | โ Avoid โ flash and boot mode pins |
If the strip worked on the bench and died at the venue, suspect the GPIO before the strip. Same with "worked in Arduino, broke after I added WiFi" โ some libraries reconfigure pins at boot.
Relay not triggering (the active-low trap)
If your strip is switched through a relay module, the classic fault isn't the strip at all โ it's the trigger logic. Most cheap relay modules are active-low: the input is pulled HIGH internally, and the relay switches on when you pull the input to GND. Sending HIGH does nothing visible, and it's the #1 "my relay does nothing" cause.
Flash the blink sketch with your relay pin swapped in. If the relay clicks when the pin goes LOW, code the logic to match โ write digitalWrite(pin, LOW) to turn the light on. This single test fixes more "dead strip" reports than any other check.
Blown fuse or PSU
If the whole strip went dark at once โ not a gradual dimming โ check the fuse and the power supply before suspecting the strip. A fuse rated just above the strip's current draw is the correct first thing to blow; replace it with the same rating, never a bigger one (that defeats the protection).
- Measure the supply output with no load. It should read cleanly at its rated voltage.
- Measure again with the strip connected. If it sags hard, the PSU is undersized or failing.
- Check for a warm or bulging PSU โ a failing supply can kill a strip on the way out.
If your ESP8266 build keeps failing, the Kachang Sia pack's troubleshooting guide covers the exact wiring mistakes that cause dead strips โ with the fixes diagrammed.
๐ See the ESP8266 Event Lighting Pack ($19)Which fix should you start with?
Don't work through all 7 in order โ start from your symptom. Two paths cover almost every case:
No light at all โ start here
- Power polarity โ Fix 1. Wrong polarity kills strips; verify the supply output and the strip's input before anything else.
- Data signal โ Fix 2 (addressable). Common ground, then data pin, then level shifter.
- Fuse / PSU โ Fix 7. A dead supply explains a dead strip instantly.
- Relay trigger โ Fix 6. If it's on a relay, test LOW vs HIGH before touching the strip.
Half-lit or dim tail โ start here
- Voltage drop โ Fix 3. Inject power every ~5m; don't buy a new strip.
- Bad joint / connector โ Fix 4. Check the pads where the strip was cut or joined.
- GPIO pin โ Fix 5. Confirm the sketch's pin is a safe one and matches the wire.
The 30-second version
Strip completely dark? โโ Was power ever connected backwards? โโโบ Fix 1 (may be blown) โโ Is it addressable (WS2812B)? โโโโโโโโโโบ Fix 2 (data signal) โโ Is everything dead at once? โโโโโโโโโโโบ Fix 7 (fuse / PSU) Strip half-lit or dim at the tail? โโ Dimmer further from the supply? โโโโโโโบ Fix 3 (inject power) โโ Died after cutting / joining? โโโโโโโโโบ Fix 4 (solder / connector) โโ Controlled by an ESP8266? โโโโโโโโโโโโโบ Fix 5 (GPIO), Fix 6 (relay)
The order matters: power โ data โ connections โ controller. Most people replace the strip first; the strip is usually the last thing to blame.
Quick answers when a strip dies.
Why is only half my LED strip working?
Half-lit or a dim tail is a power problem, not broken LEDs. The strip is dropping voltage along its length โ usually because it's too long for the supply or the wire gauge. Fix it with power injection every ~5m or thicker supply wire, not a new strip.
Why did my LED strip stop working after I cut it?
Cutting is where most connection faults happen. Check the solder pads at the cut: a cold joint, a bridge between pads, or a half-seated connector after re-joining will kill the segment or the whole strip. Reflow the pads and reseat any connectors.
Why is the far end of my LED strip dim?
That's voltage drop along the strip's copper traces. The fix is power injection: feed 5V (or 12V) into the strip again at the midpoint or far end, every ~5m. You don't need to cut the strip โ solder or clip onto the pads at a cut line.
My RGB strip is stuck on one colour โ why?
A stuck colour on an analog RGB strip usually means one or two of the R/G/B channels aren't driven. Check the connections at the controller end โ a loose signal wire or wrong GPIO pin for one channel is the common cause. On addressable strips, a stuck colour is more likely a data-line issue: common ground, signal quality, or the first LED in the chain.