At STG, we practice an integrative approach involving Simulation and
Statistics in the complete product development cycle. It is important to
understand the role of simulation and statistics in every phase of the
product development cycle. Below is a summary of how simulation and
statistics can be used in each stage:

  1. Concept Phase: During this phase, use simulation tools to
    verify circuit operation. One must start small using idle
    component models and build the system in stages. Each stage
    work should be saved. It is important to understand the theory
    and state of the art behind the circuit you are about to simulate.
    Without proper theoretical foundation, you will not be able to
    obtain useful information from simulation. Also, for majority of
    the engineers, a process methodology or steps to design must
    include simulation. Simulation is most effective when the circuit
    behavior is not well understood and we can construct several
    what-if scenarios or use simulation to build a repertoire of
    questions to be answered about the design problem at hand.
    Simulation effectiveness improves with experience and time. An
    engineering department must be dedicated to it. As with any
    other skill, to yield simulation as a potent competitive weapon,
    one must spend significant time and resources to hone it. A
    frivolous relationship or experimental tinkering with simulation
    tools will not yield any fruitful results.
  2. Design Phase: During design phase, as you begin to transform
    your work into schematics, one must pay careful attention to
    component selection and component models can be
    incorporated (especially in Spice based tools) one at a time. Do
    not be too ambitious to incorporate a host of models at one time.
    Also realize that incorporating each component model is never
    required. One must be quite prudent in incorporating essential
    component models. Just remember Pareto's principle - 20% or
    less determine 80% or more of the outcome. This must always
    be kept in mind.
  3. Prototype Phase: During this phase as prototypes are built,
    one must pay careful attention to collecting data during incoming
    inspection (mechanical variables) and testing (electrical
    variables). Here, we always recommend to use the  suppliers
    who would also build production units. It is important to do so to
    understand supplier capabilities and process variations. This
    can tremendously expedite the whole product development
    cycle. During this phase, at STG, we strongly recommend using
    statistics to understand variable distributions and correlation
    between various variables. It is important to start forming
    hypothesis on what could be troublesome variables which are
    going to effect the performance. Those variations must be
    incorporated into Simulations to re-characterize the system and
    understand overall performance variations.
  4. First Article Phase:  During this phase as First Articles are
    built, one again must pay careful attention to collecting data
    during incoming inspection (mechanical variables) and testing
    (electrical variables). During this phase, we use statistics to
    understand variable distributions and correlation between
    various variables. These correlations may change from the
    prototype stage. It is important to start forming new hypothesis
    on what could be troublesome variables which are going to
    effect the system performance. Those variations must be
    incorporated into Simulations to re-characterize the system and
    understand overall performance variations.
  5. Pre-Production/Production Phase: During this phase as Pre-
    Production or Production units are built, one again must pay
    careful attention to collecting data during incoming inspection
    (mechanical variables), in-process inspection (mechanical and
    electrical variables) and final testing (electrical variables).
    During this phase, we use statistics to understand variable
    distributions and correlation between various variables. These
    correlations may change from the earlier phases. It is important
    to start forming fresh hypothesis on what could be troublesome
    variables which are going to effect the system performance.
    Those variations must be incorporated into Simulations to re-
    characterize the system and understand overall performance
    variations. This is the process of continuous improvement and
    STG's integrated approach, if followed rigorously, yields not only
    superior products but also strong infrastructure capabilities.

There is always an "Edison approach" to design. With this approach,
you will need to spend countless hours and follow rigorous and
scientific method of design of experiments as well as truthful collection
of data. "Edison approach" is simply too expensive and unaffordable in
today's world. Nevertheless, with enough money and time, such
approach is always possible.

Spice is a great tool for Power Electronic circuits barring its annoying
and most irritating convergence problem. These convergence
problems are a great waste of time and a source of frustration.
However, there has been a steady rise in the tools and techniques in
the Spice arena, especially for the Power Electronics and Motor
Control, that Spice simulation expertise can be wielded effectively by a
company in launching its products through short product development
cycles. By no means, we are claiming that Spice expertise in Power
Electronics alone is sufficient to cut the time from concept to
production.  However, it is an important tool to have in the bag.

For Power Electronics Circuits, STG recommends the following
approach to using Spice during
the Concept Phase only:

  • Before you build your own circuit model, search to see if similar
    circuits are available in the public domain. A great engineer
    always builds her/his work on what is already available. Do not
    reinvent the wheel.
  • Always start with the most idle circuit model of components.
    Simplicity is the key. Subcircuits are great in ensuring that small
    parts of circuits can be made to work first. It is always prudent to
    use simple, well-tested models.
  • Now, transient analysis using idle components will get you only
    so far. Small Signal modeling is an important step in being able
    to overcome convergence problems as well as understand the
    circuit behavior fully. This technique requires state space
    averaging and is most vital in simulating Power Electronic
    circuits. STG specializes in small signal modeling of the Power
    Electronic circuits and if you run into hot waters using this
    technique, we are happy to help.
  • Control loop compensation is not a  simple matter for most of the
    Power Electronic circuits. However, the compensator can be
    easily designed using the small signal modeling techniques.
  • Once the above steps are followed and we have a working
    simulation, we can begin to run various what-if scenarios to
    understand the circuit behavior in time domain as well as
    frequency domain.

Below are some great references for those interested in pursuing
Spice simulation further:
Phone: Toll Free  (877) 592-4848 Fax: Toll Free  (888) 453-0997
Copyright © 2009 - 2010 STG All rights reserved.
An Integrated Approach to Simulation of Power Electronics and Motor Control Systems using PSpice
Power Electronics Consulting
Strategic Technology Group LLC
Design, Development and Manufacturing of Brushless
Permanent Magnet Motors, Motor Control & Motor Drives
PSpice
Simulations available for
Download
A Course on Advanced Power
Electronics and Motor Control
Simulation Techniques using
PSpice
Eight DVDs, One CD and One
Workbook covering over 25
Example Simulations in PSpice
Falco eMotors Product Catlog
$225 USD
2-3 Weeks for Delivery
eNews Letter Sign Up
Email:

Power Electronics

Motor Control Systems

Motor Drives