LDO Power Dissipation and Junction Temperature Calculator
Work out how much heat a linear regulator dissipates and how hot its junction gets, from the input and output voltage, the load and quiescent current, and the package's theta-JA. It also gives the margin to the maximum junction temperature, the hottest ambient the design survives, the most current the package can deliver, and the input voltage above which a buck converter is the better part. Every input lives in the URL, so a copied link reproduces the calculation.
- Power dissipation
- 510 mW
- load 510 mW + quiescent 500 µW
- Margin to Tj max
- -2.1 °C
- to 125 °C
- Max ambient
- 22.9 °C
- at this load
- Max load current
- 294 mA
- at 25 °C ambient
- Efficiency
- 66.0%
- 990 mW delivered
- Buck at 88%
- 135 mW
- loss for the same output
The junction would sit 2.1 °C above its 125 °C maximum. Most LDOs shut down somewhere around 150 to 170 °C, then restart as they cool, which looks like random resets. Use a bigger package or more copper, drop part of the voltage before the LDO, or use a buck.
A buck at about 88% efficiency would lose 135 mW instead of 510 mW, saving 375 mW. For a 3.3 V output a buck wins on efficiency whenever Vin is above 3.75 V. Here a buck, or a buck ahead of the LDO when the rail must be quiet, is the answer.
The link reproduces these exact inputs - paste it in a design review or a pull request.
Bring the rails, the loads and the enclosure. We will look at the whole power tree with you (regulators, copper, thermal paths and noise) before the board goes to fab. If you need the whole board done, see PCB design and layout services.
Package comparison at 25 °C ambient, 5 V to 3.3 V
Typical theta-JA on a JEDEC high-K board, the power each package can shed before the junction reaches 125 °C, and the load current that allows at your voltages. Change the inputs above and the table follows.
| Package | Theta-JA | Max dissipation | Max load | Tj at your load |
|---|---|---|---|---|
| SC-70-5 | 250 °C/W | 400 mW | 235 mA | 152.6 °C |
| SOT-23-5 | 200 °C/W | 500 mW | 294 mA | 127.1 °C |
| DFN 1x1 (X2SON) | 200 °C/W | 500 mW | 294 mA | 127.1 °C |
| DFN 2x2 (exposed pad) | 80 °C/W | 1.25 W | 735 mA | 65.8 °C |
| SOT-89 | 120 °C/W | 833 mW | 490 mA | 86.3 °C |
| SOT-223 | 65 °C/W | 1.54 W | 905 mA | 58.2 °C |
| DFN 3x3 (exposed pad) | 50 °C/W | 2 W | 1.18 A | 50.5 °C |
| DPAK (TO-252) | 45 °C/W | 2.22 W | 1.31 A | 48.0 °C |
| TO-220, no heatsink | 55 °C/W | 1.82 W | 1.07 A | 53.1 °C |
| D2PAK (TO-263) | 30 °C/W | 3.33 W | 1.96 A | 40.3 °C |
The equations
A linear regulator passes the load current straight through and drops the difference between input and output across its pass transistor, so it dissipates (Vin - Vout) × Iload. Its own circuitry draws the quiescent or ground current from the input, adding Vin × Iq. The junction then sits above the ambient by that power times the thermal resistance from junction to ambient, theta-JA: Tj = Ta + Pd × theta-JA. Rearranged, the same line gives the hottest ambient the design survives and the most current the package can deliver at your ambient.
Theta-JA belongs to the board, not just the package
The theta-JA in a datasheet is measured on a JEDEC 2s2p test board: four layers, two of them solid copper planes, in still air. Most of the heat leaves through the pins and the exposed pad into that copper. A compact product board with a small pour around the regulator spreads far less, so the real figure is often 30 to 50% worse, and an exposed-pad package with no thermal vias under the pad loses most of its advantage. Use the datasheet's value for the exact part, give the regulator copper on every layer it touches, and measure the case temperature on the first prototypes at full load and maximum input.
When a buck converter makes sense
An LDO can never be more efficient than Vout / Vin. A small synchronous buck runs at roughly 85 to 92% at a few hundred milliamps, so on efficiency alone the buck wins whenever the input is more than about 14% above the output. Efficiency is not the only criterion: an LDO is silent, needs no inductor, costs less and takes less board area, so it stays the right part for low currents, small voltage drops and noise sensitive rails such as RF and precision analog. The usual pattern when both matter is a buck from the high input to just above the target, followed by an LDO for the quiet rail.
Frequently asked questions
How much power does an LDO dissipate going from 5 V to 3.3 V at 300 mA?
The regulator drops the difference, so it dissipates (5 - 3.3) × 0.3 = 0.51 W, plus the input voltage times its quiescent current, another 0.5 mW at 0.1 mA. Only 66% of the input power reaches the load. In a SOT-23-5 at a typical 200 °C/W that is a 102 °C rise, a 127 °C junction at 25 °C ambient and past the usual 125 °C limit. The same load in a SOT-223 at about 65 °C/W runs at 58 °C.
How much current can a SOT-23-5 LDO supply?
As much as its package can shed as heat, which depends on the voltage it drops and the ambient temperature, not only on the current rating in the datasheet title. From 5 V to 3.3 V with a typical 200 °C/W, it holds about 294 mA at 25 °C ambient before the junction reaches 125 °C, and only about 220 mA at 50 °C ambient. A 500 mA LDO in a SOT-23-5 is a 500 mA part only when the input is close to the output.
When should I use a buck converter instead of an LDO?
An LDO's efficiency can never beat Vout / Vin, while a small buck runs at roughly 85 to 92%. With 88% as a planning figure, a buck wins on efficiency once Vin is above Vout / 0.88: 3.75 V for a 3.3 V output. From 12 V to 3.3 V at 500 mA an LDO burns 4.35 W at 27.5% efficiency, where a buck would lose about 0.23 W. From a 3.7 V Li-ion cell to 3.3 V at 150 mA the LDO is 89% efficient, and it is quieter, smaller and cheaper than the buck.
What theta-JA should I use for my regulator?
The one in your regulator's datasheet, for your exact package. Datasheets quote it on a JEDEC 2s2p test board with large copper planes, so it is a best case: a small two-layer board with little copper around the regulator can run 30 to 50% hotter. The presets on this page are typical values for common packages, for comparing options. Before a design is final, measure the regulator's case temperature on the real board at full load.
What junction temperature is too hot for an LDO?
Most LDOs specify operation up to a junction temperature of 125 °C, some to 150 °C, and shut down somewhere around 150 to 170 °C. A regulator that reaches thermal shutdown does not fail politely: it cuts the output, cools, restarts and repeats, which looks like random resets. Keep at least 20 °C of margin below the rated maximum at the hottest ambient the product will see.
If the board is the product
One regulator is a five-minute decision. A power tree that survives a sealed enclosure in the sun, a cold battery and a radio burst at the same time is not, and that is our day job: PCB design and layout explains how we work and what a board costs in engineering hours, and the battery life calculator covers the other half of the power budget. If a number here decides something real, a free 30-minute technical call is the fastest way to check it against someone who has shipped the board.