Strip Flickering
Whether it is a power fault or a data fault, read off the strip itself, and the fix for each cause
Why an LED strip flickers
An LED strip flickers for one of two reasons: a power fault, where every lit LED dims or blinks in step because the supply, the wiring or a joint cannot hold the voltage under the current being drawn, or a data fault, where individual pixels flash random colours because the signal reaching them is corrupted - and the strip itself shows you which one you have, because power flicker moves everything together and data glitches strike pixels independently.
That split is the whole diagnosis. Power faults and data faults share no causes and no fixes, so the first question is never "which of eight things is wrong" but which half of the problem you are in - and the strip answers it before you pick up a meter.
What you are looking at
The pattern names the fault, and one of these is not a fault at all.
Everything dims or blinks at once, worst on white or bright scenes
Power. The flicker tracks the load: bright frames draw the most current, so a supply or a cable at its limit gives way exactly then. A strip that runs effects happily and stumbles on white is the textbook case.
The strip lights, cuts out, and lights again in a slow loop
Power. The supply's overcurrent protection is tripping and restarting: the strip asks for more than the rating, the supply shuts down, the load disappears, the supply recovers, and the cycle repeats about once a second. An undersized supply or a short, not a bad strip.
Only the far end flickers or shifts colour
Power, but specifically voltage drop. The tail of a long run sits at the bottom of the voltage budget, so a sag the first LEDs ride out pushes the last ones below their working voltage first. More feeds fix it; a bigger supply alone does not.
It changes when you press a connector or flex the strip
A joint. A clip-on connector or a dry solder joint makes and breaks under movement, and on a low-voltage run every joint carries the full current of everything behind it. Flicker that follows your hand is that joint, found.
Random pixels flash the wrong colours, anywhere on the strip
Data. The pixels are being told to do this - a corrupted signal is still a signal. Position-random, content-random glitching means interference, a missing common ground, or a data line the controller is driving at the edge of what the first LED accepts.
Fine up to one point, glitching everywhere after it
Data, with the culprit located. Each pixel re-times the signal before passing it on, so one failing LED corrupts everything downstream while everything upstream stays clean. Cut out the pixel at the boundary and rejoin.
Flickers on camera, fine to the eye
Not a fault. The strip dims by switching thousands of times a second, and the camera's shutter beats against that frequency. Nothing on the wall is wrong - see the section on flicker that is not a fault.
Two faults, not one
They look different, fail differently, and nothing that fixes one touches the other.
Power flicker
Every lit LED moves together, because they all sit behind the same failing volt. It tracks the load - worst on white, worst at full brightness, worst at the far end of a long run - and it is fixed with supply sizing, cable, feeds and joints.
Data glitches
Individual pixels do their own thing - wrong colours, random flashes, garbage after one point - because a corrupted signal is still a signal and the pixels obey it. It ignores the load entirely, and it is fixed with grounding, short data leads and level shifting.
The draw your supply was sized for
A 5 m run of each strip, running RGB effects and at full white. On a WS2812B, effects draw about 35% of full white - which is what a wall meter shows you, and what the supply then gets bought for.
| Strip | On effects | Full white | Supply needed | Full-white current |
|---|---|---|---|---|
| WS2812B 60 LED/m | 21.1 W | 60.6 W | 72.7 W | 12.12 A |
| WS2815 60 LED/m | 49.5 W | 47.1 W | 56.5 W | 3.93 A |
| WS2811 FCOB 24V 720 LED/m | 40.4 W | 105 W | 126 W | 4.38 A |
The supply column is full white plus the 20% margin, which is what the PSU calculator recommends for your own strip and length. The current column is what every joint, connector and centimetre of cable carries at the peak - and it is the column that separates these rails, because watts for watts a 12V run carries less than half the amps of a 5V one, which is why the rail you build on decides how forgiving the whole install is. It is also the column that condemns most clip connectors: what a joint actually carries sets these figures against the 3 A a clip is good for.
WS2815 60 LED/m is the exception, and the row worth reading twice: a 12V single-addressable chip regulates every LED down from 12V and burns off the difference as heat, so it draws as much for one colour as for white. There is no surge to be caught out by - the wall meter was already showing you the worst case. The trade is that it cannot be sized down for effects either, so if you have been given a supply chosen from an effects figure, it was too small from the start. The WS2815 power reference has the measured figures and what to size against instead.
Power flicker, cause by cause
In the order worth checking them. Everything here moves the whole strip, or the whole tail of it, together.
The supply was sized for the average, not the peak
Runs effects for weeks, flickers or restarts on white flashes and bright scenes. On a WS2812B a wall meter during effects reads about 35% of what full white asks, which is how the undersized supply got bought in the first place.
Size from full white plus a 20% margin - the table above has the figures, and the PSU calculator computes them for your own strip and length. Until the bigger supply arrives, a current limit in the controller holds the peak below the rating by dimming instead of browning out.
The supply is sagging, not switching off
The whole strip dims in step with the content and white looks warm. The supply is delivering its current but below its voltage - common at the top of a cheap supply's rating, and on adjustable supplies trimmed low.
Measure across the supply terminals while the strip is lit at full white. A reading sagging under load is the supply's problem; trim it to nominal if it is adjustable, replace it if it is not holding its rating.
Voltage drop is doing it, but only at the tail
The far end flickers or shifts colour while the first LEDs stay solid. The end of the run is already at the bottom of the budget, so it crosses the line first when anything sags.
This is a feeding problem, not a supply problem: the voltage drop guide measures where the loss is, and the power injection guide places the extra feeds that fix it.
A joint that cannot carry the current
Intermittent, and it moves when the strip does. Clip-on connectors are the usual culprits on high-current runs: their contact area is small, and on 5V the current is highest exactly where the voltage margin is smallest.
Solder the high-current joints, or use connectors rated for the full-white current in the table above. The wiggle test in the isolation steps finds the joint; the fix is making it permanent.
The controller is browning out with the strip
The strip goes dark and the effect restarts from the beginning - the controller rebooted. It usually shares the same 5V that just sagged under the white frame.
Feed the controller from a stable source rather than the far end of the load: its own regulator or supply, and a capacitor across the strip's input terminals to ride out the frame-to-frame spikes.
Two of these hand off: the voltage drop guide measures where a run is losing its volts, and the power injection guide places the feeds that fix a flickering tail.
Data glitches, cause by cause
In the order worth checking them. Everything here strikes pixels independently and ignores brightness.
No common ground between supplies
Random colours, typically arriving with a second supply added for power injection. The data signal is a voltage measured against ground, and two supplies with separate grounds give it two different ideas of what zero is.
Tie every supply's negative together, always. The power injection guide covers the wiring; this is its most common mistake, seen from the symptom side.
A long lead between controller and first LED
Glitching that gets worse with distance, brightness of the wiring run nearby, or a controller that works on the bench and not on the wall. The data signal is fast and unterminated, and every centimetre of lead degrades its edges.
Keep the controller within about a metre of the first LED. If it has to live further away, send the signal differentially or re-time it with a spare pixel mounted at the controller end of the lead.
A 3.3V controller driving a 5V strip
Works with some strips and not others, or stops working when the supply voltage rises. A 5V pixel wants its data high level near 3.5V, and a 3.3V controller sits just under that line - marginal by specification, so it holds until temperature or a new reel tips it.
A level shifter between controller and strip - a one-chip part designed for exactly this - or the sacrificial-pixel trick: one LED fed slightly below 5V re-transmits the signal at full strength for the rest of the run.
Interference coupled into the data line
Glitches when a motor, relay or appliance switches nearby, or a data wire run bundled with the power wires for the same strip.
Route data away from power, keep it short, and pair it with a ground wire - twisted, if it is more than a jumper. A small resistor in series at the controller end damps the ringing that long lines add.
One failing pixel corrupting everything after it
A clean strip up to one point and garbage after it, with the boundary in the same place every time. Each pixel regenerates the signal for the next, so the first bad one poisons the rest of the run.
Cut the failing pixel out at the cut marks either side and rejoin. This is also why a mid-strip glitch after physical damage sits exactly where the strip was bent or crushed.
Which strips carry which signal - and which have a backup data line that rides out a failed pixel - is the strip types guide. The wiring that prevents the first cause outright - every supply negative on one common ground - is laid out in the multiple power supplies guide.
Finding yours
Five checks in order, each one splitting the remaining causes roughly in half.
- 1
Set full white at full brightness
This is maximum current, so every power-side cause shows itself now or not at all. A strip that only ever misbehaves on bright content has already told you which half of this page you are in.
- 2
Set a solid colour at low brightness
Now the current is close to nothing and the data stream is still running. Glitching that survives this step is a data fault - there is no load left to blame.
- 3
Measure the supply terminals during the flicker
Meter on DC volts, across the supply's output, strip lit. A reading that dips in step with the flicker means the supply is losing the fight - undersized, sagging or tripping. A steady reading moves the problem downstream.
- 4
Measure the far end of the run
Steady at the supply but low or dipping at the last pair of pads means the run is losing the voltage on the way - the voltage drop guide splits that loss into the wire's share and the strip's share, and the injection guide fixes the second.
- 5
Flex every joint while the strip is lit
Press each connector and solder joint in turn. Flicker that follows your hand is that joint, and clip-on connectors on high-current runs are the first place to press.
When nothing is broken
Three flickers with no fault behind them, so nothing on this page fixes them.
Banding or strobing on camera, invisible to the eye
PWM dimming. The strip switches its LEDs on and off thousands of times a second and shows you the average; a camera shutter samples faster than an eye integrates, and beats against the switching frequency. Filming works better at full brightness - where the LEDs are simply on - or with the shutter speed matched to the strip's PWM rate.
Shimmer or visible steps at very low brightness
Resolution, not electronics. Each colour channel has 256 levels, and at 5% brightness an effect is working with a dozen of them, so smooth fades become visible stairs. Raising the brightness floor or using a strip with more bits per channel changes it; no wiring will.
An analog strip flickering behind a wall dimmer
A mains dimmer chopping the input of a driver that expects clean mains. The driver's output follows its mangled input at twice the mains frequency. Dim on the DC side instead - a PWM dimmer between driver and strip - or use a driver rated for the dimmer type on the wall.
Questions
Why are my LED strip lights flickering?
An LED strip flickers for one of two reasons: a power fault, where every lit LED dims or blinks in step because the supply, the wiring or a joint cannot hold the voltage under the current being drawn, or a data fault, where individual pixels flash random colours because the signal reaching them is corrupted - and the strip itself shows you which one you have, because power flicker moves everything together and data glitches strike pixels independently.
Why does my strip only flicker on white or at full brightness?
Because on most chips white is the peak load. An RGB effect draws about 35% of full white on a WS2812B, so a supply sized from what the strip usually draws is running well past its comfort zone the moment every channel turns on. A 5m WS2812B reel draws about 21W on effects and 61W at full white - a supply bought for the first number folds at the second, and it folds exactly on the bright frames. If your strip is a WS2815 or a GS8208, this is not your fault: those chips regulate each LED down from 12V and burn off the difference, so they draw as much for a solid colour as for white and never show this surge. A strip that flickers on white but not on colour is telling you it is one of the chips that does surge.
Can a power supply that is too small cause flickering?
Yes, and it is the most common cause. Just over its rating, a supply sags and the strip dims and warms in step with the content; further over, its protection trips and restarts, which reads as the strip blinking on and off in a slow loop. Size the supply from full white plus a 20% margin, and set the controller's current limit as the software guard under it.
Why do random pixels flash the wrong colours?
The data signal is being corrupted, and corrupted data is still data - the pixels obediently display it. The usual causes in order: two supplies without their grounds tied together, a long lead between controller and first LED, a 3.3V controller driving a 5V strip at the edge of its specification, and interference from a data wire bundled with power wires.
Why do my LED lights flicker on camera but not in person?
PWM dimming. The strip switches thousands of times a second and your eye sees the average, but a camera shutter samples fast enough to catch the switching, and the beat between the two frequencies is the banding on the footage. Nothing is wrong with the strip - film at full brightness, or match the shutter speed to the PWM rate.
Why does only the far end of the strip flicker?
Voltage drop. The tail of a long run already sits at the bottom of the voltage budget, so it is the first place a small sag becomes visible - past a 10% loss the tail is visibly wrong even in steady state. The fix is feeding the run in more places, not a bigger supply: the power injection guide places the feeds, and the voltage drop guide measures what the run is losing and where.
Why did the glitching start when I added a second power supply?
The grounds are not tied together. A data signal is a voltage measured against ground; two supplies with separate negatives give the controller and the far half of the strip different references, and the difference reads as noise. Connect every supply's negative together and the signal has one ground again.
Do LED strips flicker when dimmed?
Two different things get called this. An addressable strip at very low brightness shows steps and shimmer because each channel only has 256 levels - a resolution limit, not a fault. An analog strip behind a mains wall dimmer genuinely flickers, because the dimmer chops the input of a driver that expects clean mains; dim on the DC side instead, or fit a driver rated for that dimmer.
Most flicker is a supply sized for the average. Get the full-white figure for your own strip and length.
Open PSU Calculator