Author: gknow81

Another personal project inspired by electric motors. My friend gave me an alternator he pulled from a mid 2000s GM car in a junk yard (not a subtle jab) here in metro Detroit.

My vision was to turn this into a wound field electric motor. I planned on designing my own wireless power transfer system and making 3d printed parts to test my design. This is obviously not any type of groundbreaking innovation but rather was meant to be a learning project. It would force me to practice CAD, review power electronics, and test out some control theory. I began by taking apart the brush system and control board, then did DMM testing to make sure the windings were still conductive. I then studied how the control board worked, which was a really good fundamentals of electrical engineering review.

Shortly after this point I had the idea for my PCB eMotor and pivoted my time towards that project. I believe the PCB eMotor has some truly innovative aspects while incorporating many of the same learnings I would have gotten from this project. Please follow the PCB eMotor project in the dedicated respective posts.

I will explain my learnings from the control board study below, as there was valuable information in that for me and it became the bulk of my takeaway from looking into alternators.

The main source of content I used for looking into alternator controls and operation was this guide I found online from Exxotest. Here is the link

I had to begin with learning about BJT operating principles. There was very little exposure to this in my coursework as most of the focus was on newer switch types. In short, a BJT (Bipolar Junction Transistor) is a semiconductor that controls current flow. When the threshold voltage drop is met across the base and emitter terminals (typically around 0.7V), a small current flows, which modulates the larger current flow between the collector and emitter. This functionality is commonly used as a switch or amplifier and can be doped as an NPN or PNP transistor.

Conducting State:

In the conducting state, the rotor is excited and the alternator generates power to charge the battery.

Key aspects of conducting operation:

• When rotor is excited, transistor T2 is conducting. 

• The potential across the Zener diode remains below its breakdown threshold, thus preventing the activation of transistor T1 and allowing T2 to conduct. 

• With T1 open, the base-emitter potential on PNP transistor T2 reaches a sufficient magnitude to initiate conduction.  

• Though not specified, my assumption is that the base current of T2, and consequently its conduction, is regulated by the ground resistor (Rg). 

• The voltage divider formed by resistors Ra and Rp establishes the reference voltage applied to the Zener diode. 

• The output diode acts as a unidirectional conductor, preventing the rotor excitation current from flowing to ground.  

Non-Conducting State:

The non-conducting state prevents current flow to the rotor, preventing power generation and battery charging.

Key aspects of non-conducting operation:

• The fully charged battery voltage provides sufficient potential to overcome both T1’s Vbe drop and the Zener breakdown voltage, facilitating base current flow through the Zener diode. 

• When T1 saturates, its low collector-emitter voltage (Vce) reduces the base-emitter voltage (Vbe) of T2 below its activation threshold. This is because a saturated BJT exhibits a very low Vce, typically 0.1V to 0.2V, which is less than the typical 0.7 Vbe activation threshold. 

• With T1 active, current diverts to ground via the ground resistor (Rg). 

• This design opens T2 while minimizing T1 current draw, preventing rotor excitation in a high-charge state. 

LT Spice Simulations:

I am very much a visual learner. I created this LT Spice model so that I could visualize the inputs and outputs as well as exploratory analysis of circuit behavior. I believe the voltage drop between the output and input waveforms is due to the diode. If you take a look at the difference between the cursors, the maximum difference is about 0.9V. This is close to 0.7V, which is typical for a PN junction. The low-value resistor creates a high load, which likely influences the diode’s voltage drop and output voltage ripple. I tried adding a 1mF capacitor across the output and this removed all of the ripple in the output voltage.

I also created this circuit for visualization purposes. The breakdown voltage of this particular Zener diode is 8.2V. You can see that the voltage rolls off at this point and flattens out as current begins to flow through the reverse biased diode.

Conclusions:

This project served as an effective review of circuit fundamentals and offered a practical opportunity to study BJT switches. It was really neat to see an example of a control strategy using only analog components. In the future, I’m interested in looking into the complexities of the Rp resistor in the voltage divider circuit.

This short research topic came up through discussion with my coworkers. We all knew that copper had lower resistance than aluminum and that it was due to resistivity, but didn’t know the details of that. We realized that none of us could answer the question “Why does copper have a smaller skin depth than Aluminum?” To gain a complete understanding, we naturally turned to first-principles physics. 

 

The DC Case: 

Copper has a lower intrinsic resistivity than aluminum, which is a multiplier in the formula for resistance: 

R = ρ * (L / A) where: 

• R is resistance (Ohms) 

• ρ is resistivity (Ohm-meters) 

• L is length (meters) 

• A is cross-sectional area (square meters) 

Lower resistivity means the material is a better conductor.  

When an electric field is applied across a conductive material, it causes electrons to move in a specific direction relative to the field. Electrons of equal magnitude are the charge carrier in both copper and aluminum.  

 

Consider the equation for approximating conductivity (Drude Model), which is the inverse of resistivity: 

σ ≈ (n * q² * τ) / m   (m is the effective mass of the electron)  

τ is the mean free time, which is longer in copper due to its atomic structure resulting in less scattering of electrons compared to aluminum. 

 

Understandably, electron mobility is the main factor in the resistivity of materials. Copper’s electrons have the ability to move more freely, leading to higher conductivity or inversely lower resistivity.  

Room Temperature Resistivity values:  

Copper: ≈ 1.68 x 10⁻⁸ Ohm-meters 

Aluminum: ≈ 2.65 x 10⁻⁸ Ohm-meters 

Aluminum is roughly 1.58 times more resistive than copper.  

The AC Case (skin effect):  

Let’s look at the equation for skin depth: 

 δ = 1 / √(π * f * μ * σ) Where: 

• δ = skin depth (meters) 

• f = frequency  

• μ = magnetic permeability ≈ μ₀ = 4π x 10⁻⁷ H/m (for both copper and aluminum) 

• σ = electrical conductivity (S/m) 

As described in the DC section, copper has a higher conductivity than aluminum. Looking at the formula above we can see that larger conductivity value will result in lower skin depth. As a result, the skin depth of copper at any given frequency is going to be less than aluminum.  

The question remains, “what causes skin effect?”  

With AC current, the direction and magnitude of the flow of electrons is constantly changing. Due to Ampere’s law, the changing current creates a fluctuating magnetic field around the conductor. This changing magnetic field induces a voltage within the conductor (Faraday’s Law), which in turn drives eddy currents. Lenz’s law dictates that the induced eddy currents always oppose the change in magnetic field. The changing magnetic field is strongest in the center of the conductor. Consequently, the eddy currents in the center flow opposite to the main current, which create an opposing magnetic field that pushes the main current away to the edges of the conductor. This situation can be thought of as the center of the conductor having an inductive impedance. Current wants to flow in the path of least resistance, therefor redistributing to the surface of the conductor.  

Summary:

What we really wanted was to understand why skin effect happens. Working through the cause-and-effect using Maxwell’s equations helped us get there and provided the understanding we were looking for. It’s now clear that the skin effect is the direct reason for power losses in conductors at high frequencies – those “AC losses” that are so important to consider when designing electromagnetic systems.

Materials Used:

• http://hyperphysics.phy-astr.gsu.edu/hbase/Tables/rstiv.html

• https://maxwells-equations.com/

• https://solidstate.quantumtinkerer.tudelft.nl/3_drude_model/

• Google Gemini

LiDAR learning was my senior design project in my final semester at Purdue. I worked as an electrical engineer on a team of 4 with three computer engineering students. My responsibilities were to guide and lead the team as well as develop the power system for the project and design the PCB for our end product.

Please see below for more details on the project!

Project Overview

The primary goal of this project is to design a device that assists teachers in identifying potential threats outside their classrooms without compromising their safety. The device integrates a camera mounted on a servo motor, a LiDAR sensor, and a light sensor to provide a comprehensive monitoring system. The camera captures real-time video, while the LiDAR sensor detects the location of subjects outside the classroom. This data is processed by a microcontroller, which controls the servo motor to keep the subject centered on the monitor inside the classroom. Additionally, an LED light is activated in low-light conditions to ensure clear visibility.

Demo Video

Inspiration

The Oxford High School shooting during my Fall 2021 internship in Michigan took place nearby and deeply affected me. As an engineer, I immediately began thinking about how technology could help mitigate such tragedies. My initial concept involved a comprehensive defense system using strategically placed LiDAR sensors throughout a school to track individuals in real-time. This LiDAR tracking data would be the input for a neural network designed to identify potential threats and, based on its analysis, control the building’s doors. By analyzing movement patterns, the system could identify potential threats and quickly and safely contain them, limiting their mobility. This would allow students and staff in unaffected areas to evacuate safely while the threat was isolated.

Budget Constraints

Due to a limited budget of approximately $400, the project scope had to be significantly reduced. This constraint prevented the team from developing the multiple-device system and incorporating the machine learning capabilities of multiple scanners that my original vision entailed. Despite these limitations, we successfully created a functional safety device using a single 2D LiDAR scanner and a camera.

Key Features and Design Requirements

1. LiDAR Sensor: The RPLiDAR sensor scans the hallway and detects moving subjects. It provides accurate distance and angle measurements, which are used to control the camera’s movement.

2. Camera and Servo Motor: The camera pedestal rotates to follow the moving subject as indicated by the LiDAR mapping. This ensures continuous monitoring of potential threats.

3. Light Sensor and LED: A photoresistor detects the brightness of the environment and activates an LED fixture if the hallway is too dark. This feature ensures that the camera can capture clear video even in low-light conditions.

4. Power Supply: The device is powered by a standard wall outlet, making it easy to install and use in any classroom.

5. User-Friendly Interface: A TFT display inside the classroom receives continuous live video feed via WiFi connection. An on/off switch is located next to the display for easy control.

System Integration and Operation

The system is designed for easy mounting outside classrooms and is powered by standard wall power. The LiDAR sensor continuously scans the surrounding area, tracking the closest object and rotating the camera to the corresponding angle. During a lockdown, the integrated illumination system activates, ensuring clear visibility for the camera. The video feed is then wirelessly transmitted to a display inside the classroom, allowing teachers to monitor the situation without exposing themselves to potential threats.

System Block Diagram Explained

The 5 Subsystems

• RPLiDAR – Self-contained LiDAR, controlled by ESP32-S2

• ESP32-S2 – Main controller, processes LiDAR data, controls hardware

• ESP32 – Camera/controller interface, Bluetooth communication

• Camera Support – Servo-driven camera mount, light sensor, lighting

• Power System – AC/DC converter, 3.3V buck converter

My Responsibilities:

As the EE guy on the team, my main responsibility was to design the power system for the project. The +3.3V rail powers the ESP32-S2 and its USB interface components, while the +5V rail powers the remaining components and sensors. The external display is powered via a dedicated +5V DC cable.

Buck Converter

I designed a buck converter voltage regulator using TI’s LMZ10504. Prototyping on a copper clad board proved challenging. After resoldering two times, I was confident it wasn’t a soldering issue or design flaw. Throughout the process, I used an oscilloscope to verify inputs and outputs and double-checked my feedback loop design. Despite the resistors being within the datasheet’s specified range, my testing pointed to the feedback loop as the problem.

Confused, I consulted our lab mentor, a retired circuit designer with over 40 years of experience. He independently diagnosed the issue as the feedback resistors, specifically R1 and R2. While the datasheet values seemed appropriate, they were too high, preventing sufficient current from reaching the IC’s feedback pin for proper regulation. We lowered the resistor values, and the regulator then functioned as expected. This taught me a valuable lesson of the importance of experience; sometimes, theory and datasheets just aren’t enough.

PCB Design

During one of my internships, I gained experience reading schematics and PCB layouts, and sat in on a design review where experienced avionics designers critiqued a junior designer’s work. Those two hours were a crash course in PCB design.

For our senior design project, I volunteered to design the PCB, which involved consolidating subsystem schematics and tackling the puzzle of component placement. Our focus was on placing components according to physical interface locations, not size or efficiency. While I wouldn’t want to do it 100% of the time, I find PCB design, like coding, to be a rewarding exercise.

Whether through luck or thorough preparation, our PCB and components worked perfectly after assembly. Teammates 3D-printed camera and controller enclosures, and we had a laser-cut enclosure fabricated. We were quite happy and proud of how our final product turned out!