Printed Circuit Design Tools

A nice article on a quick way to estimate PCB trace impedance - intro to Ohms Per Square

If you need higher fidelity, Saturn PCB Design's Toolkit is a one-stop-shop for most calculations that you'll need to size traces, vias, controlled impedance lines etc. properly

Bipolar Junction Transistor (BJT)

General Info:
BJTs are three-terminal solid-state current amplifying devices.

Device part numbers follow JEDEC convention as 2NXXXX (from JESD370B) where 2 designates the number of junctions (electrodes - 1) and N indicates packaged (versus bare die) and the four digits indicate the particular device.

Like MOSFETs, BJTs come in 2 varieties, NPN (analogue to NMOS) and PNP (analogue to PMOS), so named for the physical silicon doping zones/order used to manufacture the devices. 

Common cheap/available general purpose parts:
NPN: 2N3904
PNP: 2N3906

Semiconductor Make-up:
"Cartoon" Diagram showing doping regions

Simplified silicon cross-section

Schematic Symbols:
NPN: 
   
PNP: 

B: Base
C: Collector
E: Emitter

Collector/Emitter is in reference to electron emissions/collection. Remember positive current convention is "hole" flow, opposite to electron flow.

It can be helpful to remember that the arrow in the symbol is indicating the intrinsic diode

TO-92 Pinout:
Usually the same between models (also between NPN/PNP) but check your datasheet

Large Signal Model:
BJTs are current amplifying devices.
As seen in the large signal model below (NPN style - for PNP just flip everything), two electrodes (Base-Emitter) effectively act as a diode, while the other electrode (Collector) sinks current proportional to the Base-Emitter multiplied by a device-specific parameter Beta ($\beta$) (generally has a value ~200), also referred to as $h_{FE}$ the forward current transfer ratio. 

$\beta$ is specified on the datasheet and is lower in PNP devices than NPN.
The emitter current is the sum of the base-emitter diode current 

This predictable behavior makes BJTs convenient to use, understand, and analyze.

When analyzing a circuit with a BJT in forward-active mode, the two rules of thumb to remember are:
1) The device will try to maintain $V_{BE}$ at the PN junction diode forward potential (~0.7V)
2) The device will try to maintain $i_C=\beta *i_B$

In many cases the base current contribution to the emitter current can be ignored as a simplifying assumption (high $\beta$ means $I_E \approx I_C \gg I_B$ overwhelms the base contribution)

A BJT transistor with forward-biased $V_{BE}$ can only drive $V_{CE}$ so low, a condition referred to as saturation (not to be confused with the MOSFET saturation condition).

From a switching speed perspective, driving a BJT into saturation (low $V_{CE}$) can make "releasing" the switch (interrupting conduction) take a significantly longer amount of time - this is referred to as BJT "storage time" as charge carriers that build up in the base in saturation must be removed to allow the depletion region to form and cut off conduction.

This effect can be prevented by avoiding driving the transistor into saturation.
A clever circuit to "automatically" prevent this effect is the "baker clamp."
A simple implementation is to use a Shottky diode with low forward voltage drop with anode on the base and cathode on the collector - this holds $V_C$ within ~0.3V of $V_{B}$, which means $V_{CE}>0.4V$ which is sufficiently higher than typical saturation voltages $V_{CE(sat)} \approx 0.2V$.

Cuisinart DCC 1200 Teardown and Power Module Schematic

Tore apart and analyzed a broken cuisinart coffee maker. Pictures after the break.

Connector Terminology

Connector gender (Male/Female) for electrical connectors is determined by convention for the given connector, but as expected, the one with the pointier bits compared to it's mating partner is usually designated the male gender.

Jacks, Sockets, Plugs, etc. however, can get a little more complicated.
The terms Jack, Plug, and Socket interpreted in accordance with ASME Y14.44 Reference Designations for Electrical and Electronics Parts and Equipment (plus my own opinion on some others):

• Jack: A jack is the fixed (or less-movable) connector in a pair, regardless of whether the connector is male/female gender. Reference designator "J" is used in schematic diagrams.
• Plug: A plug is the moving component of a connector pair, regardless of gender. Reference designator "P" is used. If there are two moving parts such as cables that plug into each other, BOTH connectors can be designated "plugs". Said another way, the plug is the thing you hold in your hand, the jack is the thing you plug into.
• Socket: Sockets are generally fixed connectors that are used to mount & connect a device or printed circuit directly. Reference designator "X" is used, usually followed by the reference designator of the device it is intended to accept (see the ASME standard for rules for special cases). The distinction between a socket and a jack is that usually jacks are for mating with connectors whereas sockets accept a device package or structure directly (think light-bulb socket which mates directly with the device).
• Receptacle: Generic term for the electrical contact zone of a female gendered connector - the receiver of a male electrical contact "pin". When describing the number of electrical contacts on a connector, I usually use "N-position pin/receptacle connector."  Sometimes "socket" is used to refer to the individual receptacles.
• Outlet: Usually used where you have a point where power is being "let out." It is a good idea for power outlets, whether they be jacks or plugs, to be female gendered (having a recessed receptacle) to prevent inadvertent contact with electrically energized surfaces.

Some examples after the break

Surface Mount Packages Explained

PART 1: Discreet Devices

Surface-mount technology (SMT) discreet devices such as resistors, capacitors, etc. commonly use a four-digit package code such as 0402, 0603, 1206 to indicate their size on datasheets & for ordering. 

The two pairs of digits may represent the English units OR MAY represent the rounded unit in tens of mil (milli-inch) for a metric component!

e.g. 0603 may be describing

[06]     [03]
.06"   x .03" 
length   width

when the actual dimensions are 1.6mm x 0.8mm which is .0630" x .0315"

A metric version specification of the same would be 1608M where the pairs of digits represent the number of tenths of a mm. In the US, English style (milli-inches) units is far more prevalent and usually assumed. Always refer to a manufacturer's datasheet for actual device dimensions & tolerances.

Topline has a handly conversion page if you want to know the actual dimensions that correspond to the rounded English-unit description code. 

PART 2: ICs

Have you seen the terms SOIC, TSOP, SOT23, PLCC, QFP and other alphabet soup in reference to surface mount integrated circuit packages and have trouble keeping them straight? Topline to the rescue again - their SMT nomenclature guide is a nice reference. It describes most surface mount package types for ICs.

LTspice Parameter Stepping Example

LTspice is an excellent SPICE circuit simulation and modeling tool provided free by Linear Technology.

One of the nice features that LTspice provides is the ability to drop SPICE simulation commands right onto your schematic as text - the image after the text illustrates two extremely useful commands: .MEASURE and .STEP

.MEASURE allows you to shortcut the process of using the waveform viewer to painstakingly trace along a waveform for the value you want. The command lets you evaluate a user-customizable expression - key when you want to evaluate the same parameter for a variety of simulation conditions - and will even prepare a table of results if you automate the conditions with .STEP. To see the results (or error messages for malformed commands ☺ ) in LTspice, after simulation open the "SPICE Error Log" (or keyboard shortcut CTRL-L).

LTspice Wiki (unofficial) is a great LTspice resource and includes documentation for the .MEASURE syntax copied from the in-program help (see link for more details):  

Syntax: .MEAS[SURE] [AC|DC|OP|TRAN|TF|NOISE] <name>

 + [<FIND|DERIV|PARAM> <expr>]

 + [WHEN <expr> | AT=<expr>]]

 + [TD=<val1>] [<RISE|FALL|CROSS>=[<count1>|LAST]]

.STEP allows you to automate multiple simulations across a parameter. You can define a generic parameter using .PARAM <name> which can be used as a component value by de-referencing it using {name} and then calculate multiple simulation results/waveforms by sweeping or stepping the parameter. Other things you can "step" include temperature (to look at the temperature effect on your circuit) via parameter temp, model parameters such as transistor width or on-resistance, and even entire part models/subcircuits!

The syntax for .STEP is a little confusing so check out the LT Wiki copy of the in-program help for some examples - 

Syntax: .STEP [OCT|DEC] [<model>|PARAM] <parameter>|<source instance> [LIST] <<start value> <stop value> <steps>>|<list of values>

SPICE commands are activated in LTspice by placing them using the .op tool onto the schematic (Keyboard shortcut 'S').

Below is an example schematic with commands and resulting table of values in the SPICE Error Log. Two different parameters are stepped - temperature and load resistance - and the peak-to-peak difference across the temperature range as well as the average operating point across the temperature range is displayed in the error log. The step numbers correspond to the order specified in the command.

Coupled Noise Suppression and Brief Overview of Inductive/Capacitive Coupling

I came across a nice illustration (misspellings "chock" aside, haha) of common and differential mode interference and interference suppression. The document concentrates on AC Power Supplies but would be applicable to any signal line.

As a reminder, mutual-inductance indicators ("dots") demonstrate how the common-mode choke suppresses high frequency common-mode currents by generating an opposing/canceling current for each entering line (imagine two equal current pulses both entering in the same direction, each one generating a corresponding pulse through mutual inductance that suppresses the other line's pulse in a kind of cross-cancellation)

via Fresheneesz and Creative Commons Attribution-ShareAlike 3.0

High frequency differential noise is shunted across lines by the "X"-style capacitors and high frequency common-mode noise is shunted to chassis/ground by the "Y"-style capacitors (Class X and Y are based on safety capacitors types but I'm using them here to describe their connection configuration).

This document illustrates the difference well.

A discussion on Capacitive (Electric-Field) and Inductive (Magnetic Field) coupling (causes of interference) follows the break. 

Smith Chart and MATLAB Code to Generate

To help me grasp the Smith chart (see Antenna-Theory.com's explanation) I coded up the generation of one in MATLAB. See chart below for illustration of the chart with some explanation and some example moves. 

Code is after the break - it hasn't been cleaned up so apologies for the sparse comments. I used a convenient MATLAB library by Adam Danz for labeling points (labelpoints). 

Note that where I say "Constant Reflectance Coefficient" it's technically only the magnitude of the complex value that's constant along the line.

US Frequency Allocations by NTIA

A fun graphic to see how the radio frequency has been divided in the United States.

A few relevant frequencies (from Wikipedia):
• 4g (cell) - 700 MHz | 750 MHz | 800 MHz | 850 MHz | 1700 MHz | 2100 MHz | 1900 MHz | 2500 MHz
• 3g (cell) - 850 MHz | 1700 MHz | 2100 MHz | 1900 MHz
• 802.11a (WiFi) - 5 GHz | 3.7 GHz
• 802.11b/g (WiFi) - 2.4 GHz
• 802.11n (WiFi) - 2.4 GHz | 5 GHz
• 802.11ac (WiFi) - 5 GHz

Durometer

In engineering, there are a variety of means of quantifying hardness (squishiness) of materials. It's not a straightforward problem because there are many possible dynamics that can come into play with real world materials. Imagine trying to quantify the 'squishiness' of a rubber band versus a pillow versus memory foam - there are many time-dependent factors that greatly complicate things.

A common metric for 'soft' materials is ASTM D2240 Shore Durometer which is a family of tests that gauge the penetration distance of a so-called indentor when a force is applied. The indentors have various specifications for different materials/tests and range from a relatively large diameter (flat-ish) to quite small (sharp-ish).

from http://pitstopusa.com/n-2850-how-to-maximize-street-tire-grip-on-dirt.html via google images

Common values for some common tests/materials are shown below

from http://www.smooth-on.com/Documents-Duromete/c1351_1370/index.html via google images

For materials that are not so soft (like metals) a similar method to shore durometer is the Rockwell scale where the plastic (permanent) deformation is measured when pressing an indentor into the test material.

By Djhé (Own work) [GFDL (http://www.gnu.org/copyleft/fdl.html) or CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons

An alternative for hard objects is the Vickers hardness test which is based on the surface area of indentation of a standard diamond pyramid indentor.

All the above metrics are experimentally determined but provide standard comparative controls for materials.

ASSOCIATED LINKS

Durometer

Rockwell Scale

Vickers Hardness

Antenna Basics

For a great basic overview on antenna design and basic theory, see Peter Bevelacqua's antenna-theory.com

Digital Signal Processing

For a basic introduction to Fourier Transforms, the theory behind signal decomposing signals into a combination of various single-frequency sinusoidal, I found thefouriertransform.com to be a good overview.

For an excellent, free, practical, and reasonably comprehensive overview of a lot of digital signal topics, see Steve Smith's online textbook (downloadable for personal use).

Topics include
* Sampling theory/fundamentals
* Analog to digital and digital to analog conversion
* Convolution and Fourier Transforms
* Digital Filters
* Various Applications examples
* Further advanced mathematics (e.g. Laplace transform, z-transform)

For most algorithm discussions (FFT etc.), he includes code (sadly it's BASIC) to illustrate, a nice touch.

ASSOCIATED LINKS
Fourier Transform
Discreet Fourier Transform
Fast Fourier Transform

MATLAB Quick Reference and Function/Class Templates

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% MATLAB QUICK REFERENCE AND FUNCTION/CLASS TEMPLATES %

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% INDEXING & MATRIX REFERENCING %

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% NOTE: Matlab uses indexing that starts at 1, not 0

a';     % Transpose vector/matrix 'a'

a(1);       % Return first element of vector 'a'

a(1,:);     % Return first row of matrix 'a' (all columns)

a(:,1);     % Return first column of matrix 'a' (all rows)

a(i:j:end); % Extract every jth value starting at i until the last value

% Note: 'end' keyword can be used to refer to last index without using size() method

a(1:5, 10:20)   % Return matrix rows 1 through 5 and columns 10 through 20 of 'a'

[r,c]=size(a);  % Find number of rows & columns in matrix 'a'

length(a);      % Length of vector 'a'

size(a,d);      % Find length of 'd'th dimension, e.g. 1=rows,2=columns,3=z

find(a > 1);    % Returns list of indexes in 'a' whose value is >1. Can use any logical e.g. a=1

[r,c]=find(a);  % Returns list of all indexes with nonzero values, in this case row/col index pairs from a 2-D matrix

zeros(r,c); % Creates a r-rows by c-columns matrix of zeros

ones(r,c);  % Creates a r-rows by c-columns matrix of ones

rand(r,c);  % Creates a r-rows by c-columns matrix of pseudo-random numbers from the 0-1 scale (uniform dist)

%%%%%%%%%

% Spans %

%%%%%%%%%

linspace(a,b,n);    % Creates a row vector that spans 'a' to 'b' in 'n' number of linearly spaced segments

logspace(a,b,n);    % Creates a row vector that spans 10^a to 10^b in 'n' points

x=a:n:b;            % Creates a row vector from 'a' to 'b' spaced by value 'n'

repmat(A,r,c);      % Tile matrix A 'r' times in the rows direction, 'c' times in the cols direction

flip(A,<dim>);      % Flip order of a vector/matrix, optionally along a specified dimension

horzcat(A,B);       % Horizontally concatenate vectors/matricies A, B, C, D, etc.

vertcat(A,B);       % Vertically concatenante vectors/matricies A,B,C,D, etc.

%%%%%%%%%%%

% FIGURES %

%%%%%%%%%%%

close all;  % Close all open figures

figure;     % Create new figure

figure(1);  % Select given figure and make it active, or create one with given handle

subplot(numRows,numCols,selectActivePlotNum); % Create framework for multiple plots within one figure and/or select one

plot(x,y,<LineSpec>);   % X values, corresponding Y values, LineSpec optional

% LineSpec is a string composed of concatenating the following specifiers e.g. '--xc'

% Note: LineSpec with Marker but no Style will plot only markers without a line

% Styles: '-'Solid | '--'Dashed | ':'Dotted | '-.'Dash-Dot

% Markers: 'o' '+' '*' '.' 'x'  's'square 'd'diamond '^'triangle 'v'triangle & others...

% Colors: 'y'yellow | 'm'magenta | 'c'cyan | 'r'red | 'g'green | 'b'blue | 'w'white | 'k'black

stem(x,y);  % Plot discrete sequence data

hist(data,<nbins>); % Plot histogram of 10 equally spaced bins, or bins as specified

histfit(data,<nbins>,<distToFit>);  % Plot histogram with best-fit distribution

normplot(data); % Normal probability plot i.e. normal quantile plot

gcf;    % 'Get Current Figure' Returns handle/reference to active figure

gca;    % 'Get Current Axis' Returns handle/reference to active axis

% Note: for one-time plots it can be easier to format using Show Plot Tools in menubar

% Note: TeX markup can be used in strings

% Note: Formatting/plot commands are applied to the currently active figure

title({'Main Title Here';'\fontsize{8}Smaller Subtitle Here'});

xlabel('X Axis Units');

ylabel('Y Axis Units');

xlim([xmin xmax]);  % Manually set X limits for plot

ylim([ymin ymax]);  % Manually set Y limits for plot

saveas(gcf,'figure.png');   % Save figure as image

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Loops and Logical Control %

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

for i=1:n:10    % for(i=1;i<=10;i+=n)

    % Do Things

end

if <logical statement>

    % Do Things

elseif <logical statement>

    % Do Things

else

    % Do Things

end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Curve Fitting & Regression %

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

fitobject = fit(x,y,fitType);   % REFER TO MATLAB HELP/DOCUMENTATION for details - Curve Fitting Toolbox

plot(fitobject,x,y,{'fit','residuals','deriv1'}); % Plot fit along with input data, as well as residuals vs. fit and 1st derivative

%%%%%%%%%%%%%

% FUNCTIONS %

%%%%%%%%%%%%%

% Note: Main function names and corresponding filenames should match.

% Note: Any additional functions in an m-file beyond the first/main function

% are 'local functions' with calling scope limited to the main function

function [returnA, returnB] = functionName(inputA, inputB, varargin)

    %functionName(inputA, inputB, <variable number of arguments>)

    %   Detailed explanation here

    n=nargin; % Number of input arguments

end

%%%%%%%%%%%

% CLASSES %

%%%%%%%%%%%

classdef exampleClass < handle      % Inheriting from handle is necessary to be able to refer to class instance

    %EXAMPLECLASS Summary of this class goes here

    %   Detailed explanation goes here

   

    properties

        Data;

    end

   

    methods

        function obj=exampleClass(input)    % Constructor

            obj.Data=input;

        end

        function []=methodOne(obj,inputArg)

            % This function would be called by exampleClass.methodOne(inputArg)

            % The first argument is a reference/handle to the class instance       

            % This function does not return a value

            obj.Data=inputArg;  % This type of statement will change the class instance's variable value

        end

    end

end

%%%%%%%%%

% OTHER %

%%%%%%%%%

% NOTE: any statement without a trailing semi-colon will display the result in the command window

% General suggestions

% • Use the workspace variable explorer when creating complex matrixes to visually validate your code is creating what you expect regarding dimensions etc.

% • Make liberal use of the debugger and the capability to perform commands in the Command Window and see workspace variables DURING debugging to explore what your code is doing

% Keyboard Shortcuts:

% • Ctrl-R comments line(s)

% • Ctrl-T un-comments line(s)

% • Tab to indent line(s) (select multiple if desired)

% • Shift-Tab to de-indent line(s) (select multiple if desired)

clear;      % Delete/destroy workspace variables

clc;        % Clear command window

isempty(A); % Check whether variable A is an empty array

fprintf('String %.1d \n',value);    % Print strings and values to cons

-----------------------------------------------

ASSOCIATED LINKS

MATLAB