PixelPower

LED power supply calculator

A power supply is sized from one figure: every pixel at full white, at the top of its estimated range, plus the controller, with a margin so it never runs at its nameplate limit. PixelPower does that arithmetic in the open and rounds to a size you can buy — or checks the one on your shelf.

You enter

I’m building of at ,

Change anything — the answer updates as you type.

What you need

Calculated

91.5W

Estimated maximum power
i
Every channel of every pixel at full duty — the only figure a supply may be sized from. Based on the selected LED type and its preset figures; actual consumption depends on brightness, effects and conditions.

300 LEDs · 18.3 A at full white

could be 72.9 – 91.8 W

Recommended

5 V · 25A

Power supply
i
Sized from the top of the estimated range plus the controller, divided by 0.8 so the supply never runs at its thermal limit, then rounded up to a size that is actually sold.

The nearest size sold, with 20 % kept in reserve

Recommended

6

Power injection
i
Power weakens along the strip’s own copper. Feeding it in more than one place keeps the far end from dimming — this is the fewest evenly spaced feeds that stay inside your drop threshold.

Evenly spaced feed points, both ends included

Works with 6 feed points. The strip’s own copper carries about 4 A, so one feed can’t supply 18.3 A. Fed as recommended, the estimate stays within limits. 2 other things to check below.

Needs attention

  • Each run draws up to 18.3 A, past the 4 A its own power trace is good for. Roughly 0.07 mm² of copper on a 10 mm strip — about the ampacity of 28 AWG. Push much past this into one end and the trace, or the connector feeding it, is what fails. Far below what the supply can deliver into a fault.

    Feed it at several points — the injection plan says how many — and fuse each feed just above its own worst-case current. The supply's own limit is far above what this copper survives.

  • "Strip 1" has no fuse planned. Around 25 A would suit it.

    Fuse every branch close to the supply. A supply’s own over-current threshold sits far above what LED wiring survives, so it is not protecting this cable.

  • The cable on "Strip 1" carries roughly 10 A, below the 18.36 A worst case.

    4 mm² or thicker, or run a second pair in parallel. Ampacity also depends on bundling, ambient temperature and routing.

How it’s wired

Estimated voltage along the strip at full white · Calculated
Supply 5 V · 25 A0 m1 m2 m3 m4 m5 m
Feed point — as recommended Strip colour: within limits · colour shift likely · power won’t reachWorst point: 4.56 V of 5 V

More, when you want it

Detailsranges, standby draw, running cost, why this supply

What it draws

Estimated maximum current

18.3 A

could be 14.6 – 18.4 A

Estimated maximum power

91.5 W

could be 72.9 – 91.8 W

Worst case: every channel of every pixel at full duty — the only figure a supply may be sized from. Medium confidence. Measuring one run at full white with a clamp meter replaces the estimate with a fact — the field is under Advanced settings.

Pixels
300 over 5 m
At typical content
6.60 A · 33.0 W

Average content draw is an estimate of what the installation usually does, not of what it can do. Use it for the electricity bill, never for a supply.

With every LED off
0.300 A · 1.50 W

This is the floor, not the load.

At brightness 255 would repeat the worst case above, so it is not shown.

Power supply

Recommended

5 V / 25 A (125 W)

Sized from 18.4 A — the top of the range for what is on this supply, not the headline figure — plus 0.15 A for the controller, with 20 % kept in reserve so the supply never runs at its thermal limit.

Show the arithmetic
Worst case, top of range
18.4 A
Controller
+ 0.15 A
Required continuously
18.5 A
Divided by 0.8 derating
23.1 A
Nearest size sold
25 A

Running cost

7.3 kWh / month· 2.2 €

Drawn from the wall at 35 % average content, 6 h a day. 5.7 kWh of it is the LEDs; 0.27 kWh is the ICs idling, 0.14 kWh the controller, and 1.3 kWh is lost in the supply.

Wiring6 feed points · no supply configured · fuses and cable

Where the power goes

One row per protected branch · Calculated
Power supply 5 V 25 A? 2 m of 1.5 mm² 18.3 A · 91.5 W Strip 1300 LEDs · 6 feed points

Fuse and cable are coloured by whether they suit the branch: suits it · check it · not protecting it. A rating with a ? is suggested — nothing is fitted yet. The taps that carry power into each strip are drawn above, not here.

Estimated voltage along the run

Evenly spaced, this run wants 6 feed points. If your strip sits at the thin end of its copper range and draws at the top of its current range, it would take 7 — unlikely to be both at once, but you buy the cable only once.

Supply
at the far end of the feed cable
5.00 V
Feed at 0 m
carries 18 A
4.16 V
At 5.0 m
unfed tail, 5.0 m
beyond the model

Beyond what this model can estimate. The model predicts a drop of 566 % of the feed voltage. Past the configured 25 % it is no longer describing a working installation, so no figure is given. The strip copper is the limiting element here: add injection points or feed from both ends. Verify against the actual strip and cable specifications.

Injection points1 on a 5.00 m run
Compare layouts what another feed would buy
No injection — fed at one end as planned
beyond the model
2 evenly spaced feeds
beyond the model
4 evenly spaced feeds
3.96 V(21 % drop)
6 evenly spaced feeds
4.56 V(8.8 % drop)
8 evenly spaced feeds
4.74 V(5.2 % drop)
Cable to each feed point 2 m of 1.5 mm²

The cable running to each injection point drops voltage too, and unlike the strip it carries the full current of that feed along its whole length. This is the tap after the strip’s power arrives — the cable from the supply itself is set per branch under Power distribution. Set the length to 0 to leave it out.

Thresholds where the bands sit

These are judgement calls, not physics. Around 10 % of the feed voltage, colour shift and dimming toward the far end become visible, and the rising ground shifts the data reference with it. Past 25 % the linear model is no longer describing a working installation, so no figure is reported at all. An architectural install holding a white tolerates far less than a party effect.

Power distribution

Total load91.5 Wacross 1 branch
Current18.3 Aworst case, every LED at full white
PixelPower recommends5 V · 25 Anone configured yet
Utilization—state a capacity to compare against

Power supply 5 V

  • Strip 1 18.3 A · 91.5 W · 4.16 V at the load

    no fuse planned — around 25 A would suit it. Cable: thinner than this branch needs (4 mm² or thicker).

Planning estimates, not an electrical design. A branch here is the cable from a supply to a load and the fuse protecting that cable — if you split that feed into several injection points, each tap needs its own protection too. Fuse and cable sizes are ordering guidance: real ampacity depends on insulation, bundling, ambient temperature and routing. PixelPower does not check any electrical code. Verify against the parts’ own datasheets and your local regulations.

Power supplies

No supply configured. PixelPower still recommends one — add a supply to check a capacity you already own, or to split the installation across several.

Layoutdraw the shape, several strips

1 strip has no shape, so it is not drawn — still counted in every total. Give it a shape below to place it.

5 m · 300 pixels. Drag a shape to move it. The picture is to scale; LEDs are drawn at their real spacing until they are too close together to tell apart.

ControllerGPIO, colour order, export for WLED

PixelPower can write the LED section of a WLED configuration from this project: the pixel count, one output per strip, the LED type, and a power budget from the supply you configured. Everything else in WLED is left exactly as you set it.

Outputs

  • Strip 1 300 LEDs from index 0 · WS281x · 111 fps

  • Worth knowing

    WLED's brightness limiter caps brightness to stay inside this budget. It is not a fuse and not a protective device.

    A config reset, a firmware update, a wrong LED count or a crash removes it. Size the supply and the fusing for full white regardless — PixelPower does.

Preview the configuration 300 LEDs · 20.0 A budget

80 % of the 25.0 A PixelPower recommends, since no supply capacity was stated.

{
  "hw": {
    "led": {
      "total": 300,
      "maxpwr": 20000,
      "ledma": 61,
      "ins": [
        {
          "start": 0,
          "len": 300,
          "pin": [
            16
          ],
          "order": 0,
          "type": 22,
          "rev": false,
          "skip": 0
        }
      ]
    }
  }
}

Sending writes the LED section of the controller’s configuration and applies it immediately. Nothing is sent unless you press the button. The power budget above is WLED’s brightness limiter — it caps brightness, it is not a fuse, and it does nothing about a short.

Parts10 lines · CSV and printable report

PixelPower states what each part has to do. It does not know what is on your shelf, so no product or supplier appears here unless you type one in.

  • LED strip Strip 1 — WS2812B (5 V RGB), 60 LED/m 5 m · You entered

    300 pixels at 5 V.

  • Power supply 5 V 25 A power supply 1 piece · Recommended

    18.51 A required ÷ 0.8 continuous derating = 23.137 A, rounded up to a 25 A supply.

  • Controller ESP32 LED controller 1 piece · You entered

    300 pixels across 1 output.

  • Cable 1.5 mm² copper 28 m · You entered

    Branch and injection runs, counted out and back.

  • Fuse and holder 25 A fuse with holder 1 piece · Recommended

    Smallest standard size above the worst case with the usual 1.25× margin. Nothing is planned for "Strip 1" yet.

  • Fuse and holder 5 A fuse with holder 6 pieces · Recommended

    One per injection point, sized for the busiest tap — an interior feed carries more than an end feed.

  • Connector Injection connector or solder joint 6 pieces · Recommended

    One per injection point, per run. Contacts add resistance — four at 5–30 mΩ are worth about a metre of strip copper.

  • Data line 1000 µF electrolytic capacitor 1 piece · Recommended

    Across the rail at the first injection point.

  • Data line 300–500 Ω resistor 1 piece · Recommended

    In series with the data pin, at the controller end.

  • Data line Logic level shifter, 3.3 V to 5 V 1 piece · Recommended

    A 5 V strip wants 3.5 V on its data line; an ESP32 GPIO gives 3.3 V.

Open the printable report

Ordering guidance, not a wiring specification. A fuse has to sit above the load it feeds and below what its cable and connectors can carry, and real ampacity depends on insulation, bundling and how a run is routed. Check against the parts’ own datasheets and your local regulations.

Advanced settingsthe strip itself, how you drive it, your own figures

The strip itself

How you drive it

Where the power figures come from

Assumptions4 figures PixelPower had to guess at

Worth knowing

  • A 5 V strip wants at least 3.5 V on its data line, while an ESP32 GPIO delivers 3.3 V. It usually works, marginally, and fails intermittently.

    Use a level shifter, a 300–500 Ω resistor in series with the data pin, and a 1000 µF capacitor at the first injection point. Common the grounds.

This result rests on 4 assumptions show

These are the figures PixelPower had to guess at, and they are why the range above is as wide as it is. Replace any of them with your strip’s datasheet — or better, a clamp-meter reading at full white — and the uncertainty it contributes disappears.

  • 0.600 (0.450 – 0.900) Preset · Low confidence

    A 10 mm strip with roughly 1 oz copper on both the +V and GND traces. The default, and about what most branded reels measure.

  • 4.00 (2.00 – 6.00) Preset · Low confidence

    Roughly 0.07 mm² of copper on a 10 mm strip — about the ampacity of 28 AWG. Push much past this into one end and the trace, or the connector feeding it, is what fails. Far below what the supply can deliver into a fault.

  • 0.350 (0.100 – 0.550) Assumed · Low confidence

    Typical animated content averages roughly 35 % of full white. Sparkle effects run near 10 %, slow rainbows near 55 %. Strongly content-dependent.

  • 150 (80.0 – 250) Assumed · Low confidence

    An ESP32-class controller with Wi-Fi active, 80–250 mA depending on the board and transmit duty.

91.5 W needs 5 V · 25 A

Reference figures, 5 m at full white

Computed by the same engine as the calculator above, so they agree with it. Ranges reflect the spread between vendors. Verify against the actual strip and supply specifications.

StripLED/mPixelsCurrentPowerSupply
WS2812B (5 V RGB)30150 9.15 A (7.29–9.18)45.8 W5 V · 15 A
WS2812B (5 V RGB)60300 18.3 A (14.6–18.4)91.5 W5 V · 25 A
WS2812B (5 V RGB)144720 43.9 A (35.0–44.1)220 W5 V · 60 A
SK6812 RGBW (5 V)30150 12.2 A (9.69–12.2)60.8 W5 V · 20 A
SK6812 RGBW (5 V)60300 24.3 A (19.4–24.4)122 W5 V · 40 A
SK6812 RGBW (5 V)144720 58.3 A (46.5–58.5)292 W5 V · 74 A
WS2815 (12 V RGB)30150 3.45 A (2.70–4.50)41.4 W12 V · 8.5 A
WS2815 (12 V RGB)60300 6.90 A (5.40–9.00)82.8 W12 V · 12.5 A
WS2815 (12 V RGB)144720 16.6 A (13.0–21.6)199 W12 V · 30 A

The sizing chain

Required continuous current = the worst-case current at the top of its estimated range + the controller. Recommended capacity = that ÷ 0.8, rounded up to the next size sold: 5 V supplies come in 5, 10, 15, 20, 25, 30, 40, 50, 60 and 70 A; 12 V in 2, 3, 5, 8.5, 10, 12.5, 15, 17, 20, 25 and 30 A. The list is dense in the middle on purpose — rounding 11.4 A up to 17 A because 12.5 A was missing would be bad advice, not caution.

Why 0.8: nameplate ratings assume roughly 25 °C in free air, electrolytic capacitor life halves for every 10 °C, and a fan failing at full load is a fire path. It is a heuristic, not a substitute for the model’s own derating curve, and an enclosed or hot location needs more.

Checking a supply you already have

Enter its capacity and PixelPower reports utilisation and headroom against the worst case, with a verdict: comfortable above 20 % headroom, tight between 10 and 20 %, not enough below 10 %, overloaded past 100 %. The supply’s own overcurrent protection is set far above what a strip’s copper or a JST connector survives, which is why each branch also gets a fuse recommendation.

Several supplies are sized one group at a time — the load on a supply is the sum of what hangs off it, not of the installation.

Questions people ask

What size power supply do I need for 300 WS2812B LEDs?

PixelPower estimates 18.3 A at full white on 5 V and recommends a 25 A supply: 18.4 A at the top of the range plus 0.15 A for the controller, divided by 0.8, is 23.2 A, rounded up to the next size sold. Verify against the actual strip and supply specifications before you buy or build.

Should I divide watts by the supply’s efficiency when sizing it?

No. A supply’s nameplate is its DC output; efficiency is what it takes from the wall to deliver that. Dividing by efficiency oversizes the supply by around 15 % for the wrong reason. Efficiency matters for the running-cost estimate, where PixelPower applies it.

Can I use one big supply for several strips?

Yes, and PixelPower sizes it from the sum of what is on it. Each strip still needs its own fused branch, because the supply’s protection will not trip before a strip trace or a connector fails. The Wiring section plans the branches, fuses and cable.

Is a 60 A 5 V supply dangerous for a small strip?

The supply will happily deliver 60 A into a fault on a strip trace that is only good for a few amps. That is why every branch from the supply to a load should carry its own fuse sized for that branch — PixelPower recommends a value per branch and lists it in the parts list.

Take the plan further

The answer above is the start. The full planner draws the layout, plans the branches, fuses and cable from each supply, writes a WLED configuration from the plan and produces a parts list and a printable report — all from the numbers you just typed, and nothing to sign up for.

Open the full plannerWS2815 power, explainedPower injection, done properly