Getting Started With CUDA C on an Nvidia Jetson: A Meaningful Algorithm
In this blog post, I demonstrate a use case and corresponding GPU implementation where meaningful performance gains are realized and observed. Specifically, I implement a "blurring" algorithm on a large 1000x1000 pixel image. I show that the GPU-based implementation is 1000x faster than the CPU-based implementation.
How to Implement Image Processing Algorithms in FPGA Hardware
Recognized for their parallelism and reconfigurability, FPGAs prove ideal for real-time processing in medical imaging and computer vision. The step-by-step approach starts with understanding FPGA basics, emphasizing their reconfigurable nature and parallel processing. It guides users in algorithm selection based on factors like processing speed, resource utilization, and adaptability, then highlights designing modular and scalable algorithms. The process includes simulation for verification, synthesis using tools like Xilinx Vivado and Intel Quartus Prime, interfacing with image sensors, and testing on real hardware. The conclusion underscores FPGA's advantages in image processing, presenting ongoing opportunities for innovation in diverse industries.
Return of the Delta-Sigma Modulators, Part 1: Modulation
Jason Sachs returns to delta-sigma modulators with a hands-on, code-first treatment that focuses on the DAC side of things. Part 1 walks through first- and second-order kernels, linearized analysis, spectra, and practical coefficient choices while illustrating results with Python simulations. Expect clear rules of thumb for A, R, and B, a derivation of noise shaping behavior, and a useful error bound for RC filtering.
A Second Look at Slew Rate Limiters
Picking the right slew rate can cut overshoot dramatically while keeping delay reasonable, Jason shows. He numerically analyzes a feedforward slew-rate-limited step into a normalized second-order system and proposes a simple empirical rule R = Δx/(2π α τ) with α ≈ 1. The post includes Python/Scipy code and a 3→5 V example that demonstrates about a 3× overshoot reduction and a ≈5τ peak delay.
Painting with Light to Measure Time
When Jason Sachs needed to verify a first-order sigma-delta LED dimming implementation but had no oscilloscope, he turned to long-exposure "light painting" to turn time into space on a photograph. By sweeping the camera across blinking LEDs he captured pulse trains, read the bit patterns from the light trail, and confirmed the result with a tiny Python accumulator model. The post shares practical tips on timing accuracy, exposure, and avoiding ambient-light artifacts.
Shibboleths: The Perils of Voiceless Sibilant Fricatives, Idiot Lights, and Other Binary-Outcome Tests
Binary tests look simple until you try to pick a threshold, because false positives, false negatives, and base rate all collide. Jason Sachs uses a deliberately absurd detective story, then walks through the math of expected value, medical screening tradeoffs, idiot lights, and even a triage-style three-way decision. The payoff is a practical way to think about when a pass/fail signal helps, and when raw data or a second test is worth the extra complexity.
Debugging DSP code.
Strange sinusoidal confidence scores from an HTM neural model revealed a familiar class of DSP bugs. Drawing from forum troubleshooting, the post maps common root causes: signed versus absolute value errors, wrong intermediate references, scaling mistakes, and sampling/stride problems in integer math. Embedded engineers will recognize the diagnostic clues and practical suspects to check first when DSP outputs vary with the input.
R1C1R2C2: The Two-Pole Passive RC Filter
Jason Sachs walks through the math and simulation for the common two-pole passive RC filter, turning repetitive algebra into a compact reference you can reuse. He derives the closed-form transfer function, extracts the natural frequency and damping ratio, and explains why the topology cannot be underdamped without inductors or active stages. The post finishes with a state-space simulation recipe and practical component guidance.
Linear Feedback Shift Registers for the Uninitiated, Part XIII: System Identification
Jason Sachs shows how the output of a linear feedback shift register can be used for active system identification, not just spread-spectrum testing. The article compares traditional sine-wave probing with LFSR-based PRBS methods, demonstrates a worked Ra-Rb-C example, and unpacks practical issues such as reflected pseudonoise, ADC quantization, sample counts, and noise-shaping tricks to improve estimates.
Linear Feedback Shift Registers for the Uninitiated, Part XII: Spread-Spectrum Fundamentals
Jason Sachs shows why LFSR-generated pseudonoise is a natural fit for direct-sequence spread spectrum, then walks through Fourier basics, spectral plots, and runnable Python examples. The article demonstrates how DSSS multiplies a UART bitstream with a chipping sequence to spread energy, how despreading concentrates the desired signal while scrambling narrowband interference, and how multiple transmitters can share bandwidth when using uncorrelated sequences.
VHDL tutorial
Gene Breniman presents a hands-on VHDL walkthrough for a programmable clock divider implemented on a Xilinx CoolRunner CPLD (XC2C32A). The example shows how to declare ports and internal signals, implement a clock-division process with reset and falling-edge detection, and create a simple addressable latch to select clock rates from a 40MHz master clock. It’s a compact, practical guide for embedded engineers learning VHDL and CPLD design.
Ten Little Algorithms, Part 2: The Single-Pole Low-Pass Filter
Jason Sachs shows how a single-pole IIR low-pass filter, implementable in one line y += alpha * (x - y), tames noise in embedded signals without floating point. The post explains how to compute alpha from tau and delta-t, practical tradeoffs like phase lag and oversampling, and fixed-point pitfalls including how many extra state bits you need to avoid quantization. Short, practical, and code-ready.
Return of the Delta-Sigma Modulators, Part 1: Modulation
Jason Sachs returns to delta-sigma modulators with a hands-on, code-first treatment that focuses on the DAC side of things. Part 1 walks through first- and second-order kernels, linearized analysis, spectra, and practical coefficient choices while illustrating results with Python simulations. Expect clear rules of thumb for A, R, and B, a derivation of noise shaping behavior, and a useful error bound for RC filtering.
Adventures in Signal Processing with Python
Jason Sachs shows how PyLab (numpy, scipy, matplotlib) can handle many signal-processing and visualization tasks engineers usually reach for MATLAB to do. He walks through practical examples including PWM ripple, two pole RC filters, and symbolic math with SymPy, and shares real-world installation tips and trade-offs. The post closes with pointers to IPython and pandas to speed interactive analysis and data handling.
Ten Little Algorithms, Part 3: Welford's Method (and Friends)
Jason Sachs takes a practical look at Welford's method, a numerically stable online algorithm for computing mean and sample variance without storing large batches. He demonstrates Python implementations, shows why the naive sum and sum-of-squares approach suffers catastrophic cancellation, and why Welford is a better fit for memory- and CPU-constrained embedded systems. Jason then turns Welford into simple filters for tracking time-varying noise and discusses heuristic fixes and tradeoffs.
How to Estimate Encoder Velocity Without Making Stupid Mistakes: Part II (Tracking Loops and PLLs)
Jason Sachs explains why simple differentiation of encoder counts often fails and how tracking loops and PLLs give more robust velocity estimates. Using a pendulum thought experiment and Python examples, he shows how a PI-based tracking loop reduces noise and eliminates steady-state ramp error, and why vector PLLs with quadrature mixing avoid cycle slips and atan2 unwrap pitfalls in noisy or analog sensing.
Ten Little Algorithms, Part 5: Quadratic Extremum Interpolation and Chandrupatla's Method
Today we will be drifting back into the topic of numerical methods, and look at an algorithm that takes in a series of discretely-sampled data points, and estimates the maximum value of the waveform they were sampled from.
R1C1R2C2: The Two-Pole Passive RC Filter
Jason Sachs walks through the math and simulation for the common two-pole passive RC filter, turning repetitive algebra into a compact reference you can reuse. He derives the closed-form transfer function, extracts the natural frequency and damping ratio, and explains why the topology cannot be underdamped without inductors or active stages. The post finishes with a state-space simulation recipe and practical component guidance.
Data Types for Control & DSP
Control engineers often default to double precision, but Tim Wescott shows that choice can waste CPU cycles on embedded targets. He separates numeric representation into floating point, integer, and fixed-point, then walks through the tradeoffs, including quantization, overflow, and performance. A concrete PID example highlights why integrator precision and ADC scaling should drive your choice of data type rather than habit.
OOKLONE: a cheap RF 433.92MHz OOK frame cloner
Fabien Le Mentec built a pocket device that listens to and clones 433.92MHz OOK frames, automating the tedious reverse engineering of cheap wireless outlets. The prototype uses a Moteino with an RFM69 to sample demodulated OOK data, stores pulse durations in SRAM, and replays frames; the code and hardware notes are available on GitHub along with limitations and next steps.
Adventures in Signal Processing with Python
Jason Sachs shows how PyLab (numpy, scipy, matplotlib) can handle many signal-processing and visualization tasks engineers usually reach for MATLAB to do. He walks through practical examples including PWM ripple, two pole RC filters, and symbolic math with SymPy, and shares real-world installation tips and trade-offs. The post closes with pointers to IPython and pandas to speed interactive analysis and data handling.
Ten Little Algorithms, Part 2: The Single-Pole Low-Pass Filter
Jason Sachs shows how a single-pole IIR low-pass filter, implementable in one line y += alpha * (x - y), tames noise in embedded signals without floating point. The post explains how to compute alpha from tau and delta-t, practical tradeoffs like phase lag and oversampling, and fixed-point pitfalls including how many extra state bits you need to avoid quantization. Short, practical, and code-ready.
How to Estimate Encoder Velocity Without Making Stupid Mistakes: Part I
Encoder velocity estimation is easy to get wrong, and Jason Sachs walks through the traps engineers fall into. He demolishes the common advice to time between encoder edges, shows how encoder quantization and state-width errors break that approach, and argues for fixed-rate sampling with sensible filtering for most control uses. Part II will cover more advanced estimators for higher performance needs.
VHDL tutorial
Gene Breniman presents a hands-on VHDL walkthrough for a programmable clock divider implemented on a Xilinx CoolRunner CPLD (XC2C32A). The example shows how to declare ports and internal signals, implement a clock-division process with reset and falling-edge detection, and create a simple addressable latch to select clock rates from a 40MHz master clock. It’s a compact, practical guide for embedded engineers learning VHDL and CPLD design.
Ten Little Algorithms, Part 3: Welford's Method (and Friends)
Jason Sachs takes a practical look at Welford's method, a numerically stable online algorithm for computing mean and sample variance without storing large batches. He demonstrates Python implementations, shows why the naive sum and sum-of-squares approach suffers catastrophic cancellation, and why Welford is a better fit for memory- and CPU-constrained embedded systems. Jason then turns Welford into simple filters for tracking time-varying noise and discusses heuristic fixes and tradeoffs.
How to Estimate Encoder Velocity Without Making Stupid Mistakes: Part II (Tracking Loops and PLLs)
Jason Sachs explains why simple differentiation of encoder counts often fails and how tracking loops and PLLs give more robust velocity estimates. Using a pendulum thought experiment and Python examples, he shows how a PI-based tracking loop reduces noise and eliminates steady-state ramp error, and why vector PLLs with quadrature mixing avoid cycle slips and atan2 unwrap pitfalls in noisy or analog sensing.
R1C1R2C2: The Two-Pole Passive RC Filter
Jason Sachs walks through the math and simulation for the common two-pole passive RC filter, turning repetitive algebra into a compact reference you can reuse. He derives the closed-form transfer function, extracts the natural frequency and damping ratio, and explains why the topology cannot be underdamped without inductors or active stages. The post finishes with a state-space simulation recipe and practical component guidance.
Slew Rate Limiters: Nonlinear and Proud of It!
Slew-rate limits are a small nonlinear detail that often decides whether a controller behaves nicely or wrecks hardware. Jason Sachs walks through why slew limits appear in electronics and actuators, then shows two practical digital ways to impose limits: constraining input increments and constraining input around the output. He compares performance on underdamped second-order systems, gives closed-form intuition for overshoot, and demonstrates simulations with scipy and ODE solvers.
Ten Little Algorithms, Part 5: Quadratic Extremum Interpolation and Chandrupatla's Method
Today we will be drifting back into the topic of numerical methods, and look at an algorithm that takes in a series of discretely-sampled data points, and estimates the maximum value of the waveform they were sampled from.
Return of the Delta-Sigma Modulators, Part 1: Modulation
Jason Sachs returns to delta-sigma modulators with a hands-on, code-first treatment that focuses on the DAC side of things. Part 1 walks through first- and second-order kernels, linearized analysis, spectra, and practical coefficient choices while illustrating results with Python simulations. Expect clear rules of thumb for A, R, and B, a derivation of noise shaping behavior, and a useful error bound for RC filtering.
