Embedded Systems · Cluster
Power Management
Sleep states, duty cycling, and energy budgets.
- 01 From Duty Cycle to Average Current and Battery Life How charge accounting turns an active and sleep schedule into an average current, and where the resulting battery-life estimate stops being trustworthy.
- 02 What a Sleep Mode Actually Turns Off The clock, regulator, and bias changes behind the named low-power states on a microcontroller, and why the deepest state is rarely the cheapest one to use.
- 03 Wake-Up Latency and Its Energy Cost Why returning from a low-power state costs microseconds and microjoules, how that cost sets a floor on the usable duty cycle, and where the transition sequence loses time.
- 04 Tickless Idle and Timer Reprogramming How a kernel suppresses the periodic tick while the system is idle, which timer it programs instead, and what the reprogramming costs in latency and in lost timekeeping.
- 05 Peripheral Clock Gating and Autonomous Transfer How gating a peripheral clock saves dynamic power, why a gated peripheral still costs leakage, and how DMA lets a conversion happen with the core asleep.
- 06 Voltage and Frequency Scaling with a Performance Constraint Why lowering frequency alone rarely saves what it appears to, how voltage follows frequency, and where the quadratic voltage term stops paying for itself.
- 07 Duty-Cycled Radios and Preamble Sampling Why a receiver that listens continuously dominates a radio budget, how preamble sampling trades latency for a short listen window, and where the trade breaks.
- 08 Sizing Energy Storage for an Intermittently Powered Device How a harvesting source, a buffer capacitor, and a load with a fixed energy per operation determine whether a device can run at all, and what happens at the boundary.
- 09 Retention Versus Reinitialisation in Deep Sleep The trade between keeping state alive across a sleep and rebuilding it on wake, expressed as a sleep-current penalty against a per-wake energy cost.
- 10 Attributing Energy to Code with a State Machine Model How to assign charge to firmware paths by tracking state transitions, why instruction counts are a poor proxy, and where the attribution model is structurally wrong.