WS2815 power, explained

What a WS2815 pixel draws

A WS2812B pixel drives each die straight from the 5 V rail through a 20 mA constant-current sink, so full white is 3 × 20 mA plus about 1 mA for the IC: roughly 61 mA. A WS2815 pixel sits on 12 V and drives its three dies through per-pixel regulators; the current the strip draws per pixel at full white is commonly measured at 18–30 mA, with 23 mA a fair central figure. That is a third of the current at 2.4× the voltage — about 0.28 W per pixel, comparable to the 5 V part.

Where does the power go? The dies still need only their forward voltages, around 2 V for red and 3 V for blue and green. The regulator drops the rest of the 12 V as heat. So a WS2815 strip runs warmer than its light output suggests, and its idle power — the IC’s own supply current at 12 V — is 2.4× a 5 V strip’s: a thousand idle pixels are about 12 W, not 5.

PixelPower carries the 18–30 mA range rather than the central figure, and marks the preset as an assumption with low confidence. A clamp-meter reading of your own strip at full white replaces it with a fact.

Per run, at full white

Estimated current and power for one 5 m run at full white, from the same engine the calculator uses. Ranges reflect vendor spread.

What supply it needs

The recommendation takes the top of the estimated range, adds the controller, divides by 0.8 so the supply is never at its nameplate limit continuously, and rounds up to a size that is sold. For 300 WS2815 pixels that is about 9 A at the top of the range plus 0.15 A, over 0.8, or 11.4 A — a 12.5 A 12 V supply. The same 300 pixels of WS2812B want a 25 A 5 V supply.

Do not divide by the supply’s efficiency when sizing it. The nameplate is DC output; efficiency is what the wall pays. It belongs in the running-cost figure, and that is the only place PixelPower uses it.

Why 12 V needs fewer feeds

Voltage drop along a strip fed at one end is r·L·I / 2 — the strip’s round-trip resistance per metre, times its length, times its total current, halved because the current falls along the run. Cut the current to a third and the drop in volts falls to a third; put it on a 12 V rail instead of 5 V and the same volts are a smaller fraction of what the pixels have to work with. Both together mean a WS2815 run reaches something like six or seven times further from a single feed before colour shift than the same strip in WS2812B would.

It is not unlimited. The regulators hold current constant as the rail sags, right down to roughly 9–11 V, and then the pixels simply stop — there is no gradual dimming to warn you. PixelPower treats the region past 25 % drop as beyond what the model can describe and reports no figure for it.

The same run, fed in more places

Estimated drop at the worst point of a 5 m WS2815 run at 60 LEDs/m, standard copper, for evenly spaced feeds including both ends. Drop falls with the square of the feed count.

What to check before you buy

Measure one run at full white if you can; it replaces the widest assumption in the plan. Find out how much copper the strip has — 5 mm strips and half-ounce copper roughly double the drop. Fuse each branch from the supply, because a 12 V 20 A supply will deliver 20 A into a fault on a trace good for four. And verify every figure against the manufacturer’s specifications: these are planning estimates, not a certification.