Your addressable LED strip turns red, amber, or dim at the far end because of voltage drop. A 5V power supply connected at one end loses voltage as it travels along the strip’s thin flexible copper traces. When the supply drops below roughly 3.2V, the blue and green LED diodes inside each pixel shut off, leaving only the red diodes illuminated because red requires a lower forward voltage to operate.

The Anatomy of the "Red Shift"

Every individual pixel on a WS2812B (often sold under brand names like NeoPixel) or SK6812 strip contains three tiny light-emitting diodes: one red, one green, and one blue, alongside a tiny integrated circuit (IC) that interprets digital data. To produce pure white light, the microchip turns all three colours on at full brightness.

However, different colours require different amounts of voltage to illuminate—known as their forward voltage (V_f):

• Red LED: ~1.8V to 2.2V
• Green LED: ~3.0V to 3.2V
• Blue LED: ~3.0V to 3.4V
• Internal Driver IC: ~3.3V to 3.5V for stable logic

When you feed 5V in at the start of a strip, all three channels receive sufficient voltage. But as current travels down the strip, the voltage gradually drops. Once the available rail falls below ~3.0V, the blue diode turns off entirely. Soon after, the green diode dies out. The red diode, which needs only ~2.0V, continues to draw current happily. A strip told to show pure white transitions from bright white to pale yellow, then orange, and finally dull, muddy red at the tail end.

Why the Voltage Drops: The Copper Bottleneck

Flexible LED strips are printed on paper-thin flexible printed circuits (FPCs). While copper is a very good conductor, it is not a superconductor; thin copper traces have measurable electrical resistance.

The arithmetic follows Ohm’s Law: Voltage Drop (V) = Current (I) × Resistance (R).

A typical 1-metre strip with 60 WS2812B LEDs draws roughly 50 to 60 milliamps (mA) per pixel at full white (20mA per colour channel, plus ~1mA for the internal chip). At maximum brightness:

60 LEDs × 0.055A = 3.3 Amperes

Pushing more than 3 Amperes through a trace that is often just 0.5mm wide and 35 microns thick causes significant resistance. If the trace resistance over a 2-metre strip is 0.5 ohms, pulling 3A down that line will drop 1.5V before the power reaches the final LED. Your 5.0V supply is now 3.5V at the tail end. Add another metre or two, and the tail end drops well below 3.0V.

Fix 1: Software Brightness Limits (Zero Cost)

Before buying new hardware or running extra wires, look at your code. Human perception of brightness is logarithmic, not linear. An LED running at 30% electrical power does not look one-third as bright to human eyes; it looks roughly 60% to 70% as bright.

If you are using the FastLED library in Arduino or C++, add this to your setup() routine:

FastLED.setBrightness(64); // Scales brightness to 25% of maximum

In Adafruit NeoPixel:

strip.setBrightness(64);

Dropping the brightness cap from 255 (100%) to 64 (25%) divides the total current draw by four. By slashing current (I), you slash the voltage drop (V = I × R) by the same factor. On strips under two metres, this single line of code often eliminates the red shift entirely without requiring hardware modifications.

Fix 2: Power Injection

If your project demands full brightness or runs longer than 2 metres (120+ LEDs), software adjustments are not enough. You must implement power injection.

Power injection means running thicker secondary power wires in parallel with the strip and splicing them directly into the strip's 5V and GND copper pads at regular intervals. Instead of forcing 5 or 10 amps to travel entirely through the strip's paper-thin copper foil, current travels through low-resistance external wire and enters the strip where it is needed.

To inject power at both ends:

1. Connect the 5V and GND wires from your power supply to the input end of your strip as usual, alongside your microcontroller's data line.
2. Run a separate pair of 18 or 20 AWG wires directly from the same 5V power supply to the 5V and GND pads at the far end of the strip.
3. Never connect the data line at the end—data travels in one direction only (indicated by arrows on the strip PCB). Power, however, is a shared rail and can enter from both directions.

Do You Need an Expensive Bench Power Supply?

No. Makers and school labs often assume that addressable LEDs require bulky lab bench equipment. In reality, a standard, regulated 5V power source works perfectly, provided you supply enough current:

• USB Phone Chargers: A standard 5V 2.4A or 3A USB wall adapter can power up to 100 LEDs comfortably at reasonable brightness. Cut an old USB cable open: the red wire is +5V and the black wire is GND.
• 5V Mean Well / Enclosed Switching Supplies: For long classroom installations, permanent signage, or high-density matrix panels, a dedicated enclosed 5V switching power supply (e.g. 5V 10A or 20A) provides reliable, clean DC power at low cost. If you need sourcing for reliable supplies and hardware for school labs, browse our power and prototyping components at Sheen Robotics Store.

A Critical Rule: Protect Your Microcontroller

Never power more than 8 to 10 addressable LEDs directly from the 5V or 3.3V pin of an Arduino, ESP32, Raspberry Pi Pico, or BBC micro:bit. The onboard voltage regulators on microcontroller development boards are tiny linear regulators designed to deliver 150mA to 500mA total.

Forcing a dev board to supply 2A to an LED strip will overheat the onboard regulator, cause brownout resets (the board randomly restarts when LEDs flash white), or permanently burn out the board's traces.

Instead, wire your 5V power supply directly to the LED strip's power rails. Then run power to the microcontroller separately, or use that same external 5V supply to feed the board's VIN or 5V pin. Always ensure the microcontroller’s GND and the LED strip’s GND are connected together; without a common ground reference, the high-speed digital data signal will become corrupted, resulting in erratic flickering.