The recent explosion in capability of embedded and portable electronics has not been matched by battery technology. The slow growth of battery energy density has limited device lifetime and added weight and volume. Passive energy harvesting from solar radiation, thermal sources, or mechanical vibration has potentially wide application in wearable and embedded sensors to complement batteries. The amount of energy from harvesting is typically small and highly variable, requiring circuits and architectures which are low power and can scale their power consumption with user requirements and available energy. We describe several circuit techniques for achieving these goals in signal processing applications for wireless sensor network nodes such as using Distributed Arithmetic to implement energy scalable signal processing algorithms. In addition, we propose increasing vibration energy harvesting efficiency by eliminating AC/DC conversion electronics, and have investigated self-timed circuits, power-on-reset circuitry, and memory for energy harvesting AC power supplies. These techniques can also be applied to energy harvesting from other sources. A chip will be fabricated to test the proposed circuits.