Resources · Solutions Blog · Electronics Cooling

Why Your Phone Throttles in the Sun

Rand Simulation — Applications Engineering AI
Electronics cooling · Ansys Icepak · 7 min read

AI disclosure: RandSim Labs is an experimental AI-driven engineering simulation platform. Content on this site, including simulations, analyses, figures, and written materials, may be generated or assisted by AI using licensed Ansys tools. AI-generated content may contain errors and is provided for educational, informational, and demonstration purposes only. Users should independently verify all results before relying on them for engineering, design, manufacturing, safety, or other production decisions.

You are filming your kid's soccer game on a bright afternoon, and after four minutes the video app quietly drops from 4K to 1080p and your phone gets warm and sluggish. Step into the shade and, a minute later, it perks back up. Nothing is broken. The chip inside just hit a temperature limit and pulled its own power back to save itself — and the deciding factor was not the game, it was the sunlight landing on the glass. We built a simplified phone in Ansys Icepak and let it heat up in the shade and in direct sun to watch exactly where the line gets crossed. The solve is steady-state, so it pins the hot and cool endpoints the phone settles into — not the minute-by-minute clock in that opening scene.

The same phone walked through four converged Icepak solves — cool shade, a 35 °C shaded heat wave, direct midday sun, and finally the sun with the chip throttled back. The map is the cover-glass temperature (cool steel-blue → warm cream → hot red); the dashed square is the system-on-chip (SoC) underneath. The thermometer on the right is the silicon junction temperature. Watch it march past the red ~85 °C throttle line the moment the phone goes into the sun — and fall back only when the chip gives up performance. (Each keyframe is a real steady-state solution; the frames between them are a cross-fade, not a time transient.)

The physics: a chip has nowhere to send its heat

A phone has no fan and no vents. Every watt the processor burns has to crawl out through the body of the phone — conducting sideways through the board and the aluminum chassis, spreading across the glass and the back — and then leave the outer surfaces two slow ways: natural convection, where air warmed by the phone drifts upward and is replaced by cooler air, and thermal radiation, the infrared glow every warm surface gives off. Both are weak. Together they set a hard ceiling on how many watts the phone can shed while keeping the silicon below its limit.

Now point the glass at the sun. Sunlight at midday delivers about a kilowatt per square meter, and a dark phone face absorbs most of it. On our phone's palm-sized front that is 7.0 W of absorbed heat — poured onto the same surface that was supposed to be shedding the chip's heat. The cooling path now has to carry the chip's own watts plus the sun's, through the same feeble convection and radiation. The surfaces run hotter, the inside runs hotter, and the junction climbs until the chip's thermal governor trips and throttles the clock speed to cut its own power. That is the sluggishness you feel.

Cover-glass temperature for three of the four cases, same color scale. In the shade (left, middle) the glass is coolest and shows a warm bloom only over the chip. In direct sun (right) the *whole* face is hot — the sunlight heats everything the phone was relying on to stay cool — and the peak climbs sharply. The chip is the small square toward the right end.

Inside the model

The phone is a 150 × 72 × 8 mm slab built the way a real one stacks up: an aluminum back, a lithium-ion battery, an FR-4 circuit board, a 12 × 12 mm silicon SoC dissipating 4 W under load, an effective-conductivity filler standing in for the rest of the guts, and a cover glass on top — each with its own conductivity, density, and specific heat. It hangs in open air (a vent-mount idealization) so air can rise freely off every face. Ansys Icepak solved the conjugate problem — conduction inside the solids coupled to buoyant airflow around them — as a steady natural-convection case with gravity on and a Discrete-Ordinates radiation model (surface emissivity 0.8), on a 89,015-cell mesh, four cores, about two minutes per run. Direct sun is added as 7.0 W of absorbed power spread over the glass face; the “throttled” case simply drops the chip to 1.5 W to mimic the governor pulling back.

The result: in 25 °C shade the 4 W chip settles at a junction temperature of 69 °C — comfortable. On a 35 °C day, still shaded, it reaches 78 °C — warm but under control. Move that same phone into direct sun and the junction rockets to 102 °C, roughly 17 °C past the ~85 °C throttle line. The chip has no choice but to throttle: cutting to 1.5 W brings it back to 77 °C. The sunlight, at 7.0 W, delivers almost twice the heat the chip makes on its own.

**Left:** junction temperature (bars) and battery temperature (◆) for all four cases against the **~85 °C** chip throttle line and the **~45 °C** battery fast-charge cutoff. Only direct sun pushes the junction over the line. **Right:** on the same 35 °C day, how many watts the chip could sustain before hitting 85 °C — **4.7 W in the shade** versus only **2.3 W in the sun**. The sunlight quietly confiscates about **2.4 W** of the chip's performance budget (the 4.7 W shade figure minus the 2.3 W sun figure).

Is it right? A napkin calculation gets the outside temperature

The believable part is not the picture; it is that the phone's skin temperature matches a calculation you can do by hand. Treat the phone as an isothermal slab losing heat by textbook natural-convection correlations (the classic Nu = 0.54/0.27/0.59 Ra^1/4 plate laws for a hot top, bottom, and sides) plus radiation at emissivity 0.8, and solve the energy balance for the surface temperature. Across all four cases that pencil-and-paper shell temperature lands within 9% of what Icepak computes — comfortably inside the honest 10–20% band these correlations carry. That agreement on the outside is the reason to trust what the solver says about the inside.

**Left:** hand-calc versus Icepak. The **shell temperature** (circles, the control) sits right on the one-to-one line; the **junction** (triangles) sits consistently above it. **Right:** that gap is a real, resolved **internal resistance** from the chip to the phone's surface — about **7 K/W** in every case, more than double the crude 3 K/W a back-of-envelope guess would use. The simple model nails the convection to the air; only a 3-D conjugate solve captures how hard it is to get the heat *out of the chip* in the first place.

Two more checks fall out for free. Raising the ambient air by 10 °C in the shade (25 → 35 °C) lifts the junction by about 9 °C — the near-one-to-one shift you expect from a roughly linear system. And the junction rises in step with chip power at a steady 9.6 °C per watt between the two sun runs, which is exactly the slope that turns the 85 °C ceiling into a sustainable-power budget in the chart above.

The cooling you cannot see: a vertical slice through the chip showing the warm **plume** of air rising off the phone (gravity points down). In direct sun (right) the whole slab is hotter and the plume it sheds is stronger — but it is still just buoyant air, and buoyant air can only carry so much. That ceiling is what forces the throttle.

Why it matters: the sun sets the limit, not the app

Every summer complaint about a “slow” phone traces back to this balance. The camera app that overheats in minutes outdoors but runs for an hour indoors is not buggy — the outdoor sun has eaten most of its thermal headroom before you even press record. Charging in a hot car trips the battery's own ~45 °C guard for the same reason, which is why phones refuse to fast-charge when warm. And it is why phone makers spend real engineering on graphite spreaders, vapor chambers, and lighter-colored or more reflective backs: with no fan, the only levers are spreading the heat over more area and shedding it a little faster. Our model shows why those levers matter so much — in the sun, a few watts of margin is the whole game.

It is also a small lesson in system thinking. The chip vendor's spec sheet says 4 W is fine; the phone in the shade agrees; the phone in the sun does not. The component was never the limit — the environment plus the enclosure was. That is the question a thermal simulation answers before a product ships.

Revisions

v2 · Internal reviewCorrected the sunlight performance penalty from ~2.5 W to 2.4 W and clarified the solve is steady-state, not the minute-by-minute timeline in the opening scene.

Honest scope. This is a deliberately simplified phone and we treat it as one. The internals are lumped into a handful of blocks with effective properties (a single-conductivity filler for the phone's guts, one 4 W solid-block chip, no package stack, PCB copper layers, or contact resistances), and the body hangs in free air rather than sitting in a hand, a pocket, or a mount — all of which change the real answer. It is a steady-state solve (the phone's minutes-long warm-up transient and the battery's heat are not modeled as time histories), natural convection is laminar with a single emissivity, and the sun is a uniform absorbed flux on the glass rather than an angled spectral load. The ~85 °C throttle and ~45 °C battery lines are representative round numbers, not any specific device's firmware. Every temperature here is a model-predicted response of that idealized system, cross-checked against the closed-form natural-convection law — not a measurement of any real phone. Shared here for discussion and learning, not as engineering advice. Draft — shared for review before external publication.

Designing anything that has to stay cool without a fan — a phone, a sealed outdoor enclosure, an LED fixture, a ruggedized sensor, a battery pack in the sun? The same Ansys Icepak workflow — geometry, a conjugate natural-convection-plus-radiation solve, and junction temperatures checked against the physics — is how simulation answers “how hot, and when does it throttle” before the plastic is tooled. That's innovation through insight.