Electronics Design AULDO overheating and shutting down on a 2S Li-ion powered board
Asked by grumpy_otter7 ·
Got a board running off a 2S Li-ion pack (7.4V nominal, up to 8.4V fresh off the charger) feeding an LDO down to 3.3V for the MCU and a couple of sensors, pulling maybe 300 mA combined when everything's active. The regulator (a basic SOT-223 LDO) gets uncomfortably hot within a couple of minutes and the board browns out — I'm assuming thermal shutdown.
Board space is tight (it's a handheld enclosure) so I picked the LDO originally to keep the BOM and layout simple. Is there an easy fix here, or do I need to redesign the power stage?
3 Replies
Do the math and the thermal shutdown stops being a mystery. At 8.4V in, 3.3V out,
300 mA load: (8.4 − 3.3) × 0.3 = 1.53 W dissipated as heat, in a SOT-223. That
package's junction-to-ambient thermal resistance, even with a reasonable copper
pour, is going to land you well over 100°C rise at that dissipation. It's not
malfunctioning — it's doing exactly what a linear regulator does when you ask it
to drop 5 volts at non-trivial current.
Simplest real fix: replace it with a buck converter. At ~90% efficiency you'd be dissipating roughly 0.15 W instead of 1.53 W for the same output — that's the whole argument for switching over linear once the drop and current both get past "tiny." There are buck modules in SOT-23/small-QFN packages now that aren't meaningfully bigger than the LDO you're replacing, so "no room for it" usually isn't actually true once you go looking.
Yeah, the math tracks with what I'm seeing on the thermal camera. Before I redo the layout for a buck — is there any way to claw back some margin and keep the LDO, given how tight this enclosure already is? Even buying time until the next revision would help.
A couple of things buy you margin, but neither changes the physics:
- More copper pour under and around the SOT-223 tab, with thermal vias down to an internal/bottom layer plane, can meaningfully lower junction-to-ambient resistance. Won't fix 1.5 W in a tiny enclosure, but it's worth doing regardless.
- If you genuinely can't fit a buck this revision, pre-regulate with a small, inexpensive buck down to ~4–4.2V first, then run your existing LDO from that into 3.3V. Drop becomes under 1V instead of 5V, dissipation drops below 300 mW, and you keep the LDO's clean, low-noise output for the sensors if that mattered to you in the first place.
For a handheld product shipping in volume, though, I'd just do the buck properly next rev rather than carry a band-aid forward. When you do, PCB layout for switching regulators is worth reading before you start placing components — the switching-node copper area and input cap placement are where most first-spin buck designs create their own problems.
Related Discussions
TPS63020 output drooping 200 mV and ringing as battery drains through 3.3 V — mode transition issue?
Building a battery-powered sensor logger — single Li-ion cell (3.7V nominal, 2.8–4.2V range during use). Using a TPS63020 to regulate down t
MAX17048 reporting 0% SoC at power-on — reading jumps to ~70% after several seconds
Adding a MAX17048 to an ESP32 project — single 18650 cell, standard SOT-23 breakout board, I2C on pins 21/22. Address 0x36, confirmed with a
TP4056 CHRG LED stays on for hours — is the charge ever going to terminate?
Built a small Li-ion charger using one of those cheap blue TP4056 modules from AliExpress. Battery is a single-cell 2000 mAh flat LiPo (nomi