This file contains answers to commonly asked questions related to the following topics:

Program Features and Background

Program Databases: Cores, Materials, Wires

Designing Different Types of Magnetics

Losses and Thermal Effects

SPICE Related

Student and Demonstration Versions, Documentation

Example Button Scripts

Switch Mode Power Supply Wizards

Program Features and Background

Q. What makes Magnetics Designer so unique?

Magnetics Designer has a number of features that set it apart from all programs, past and present, that have tried to perform magnetics design. Some of the more outstanding features include:

Q. What is the maximum number of windings I can have?

The winding limit has been increased in the Windows version from to 32. It was 12 in the DOS version.

Q. Where was Magnetics Designer created? Is it related to any other program on the market?

The Magnetics Designer software was originally developed by Analytic Artistry, the leader in the field of nonlinear magnetics design tools. In 1995, Intusoft acquired the rights to the “Transformer and Inductor Spreadsheet,” originally a DOS based program. Intusoft has ported the Analytic Artistry product to Windows, added a number of key enhancements, and changed the name to Magnetics Designer. Intusoft is now the sole owner and developer of the Magnetics Designer software. The DOS version is no longer sold, but it is supported. Power supply designers with decades of real-world experience have validated the software at both companies. Product satisfaction is guaranteed through the many years of successful operation and the hundreds of current users in the installed base.

Q. Does Magnetics Designer account for wire margins required for bobbin use?

Magnetics Designer has an automargin feature that automatically makes room for wires to exit from the bobbin wound (or pot core) designs. The available room is reduced by the required wire thickness. For windings with an odd number of layers, room to exit on both ends of the bobbin is provided. The thickest previous margin is used for all subsequent windings. For example, use of heavy wire on the first layer will penalize other layers.

Q. Can inductors and transformers be designed at the same time or do I need separate programs?

Inductors and transformers are independent and can both be designed in the same session. Separate programs or separate runs are not required.

Q. Does Magnetics Designer account for winding pitch?

Winding pitch is a user-controlled variable and may be set to any real value from 0 to n where n is the width of a wire.

Q. What is the core design algorithm used by Magnetics Designer?

Magnetics Designer uses sophisticated design algorithms that go far beyond previous versions. An expert system automatically tries logical manipulations of turns, wire size, number of strands and core geometries to meet electrical requirements and temperature rise/window fill constraints. the algorithm is fully explained in the on-line help screens of the demonstration version.

Q. Can I automatically split and move windings?

Magnetics Designer allows you to clone (or split) or move windings at will. When a winding is split, Magnetics Designer is smart enough to evenly distribute the currents between the new windings. Proximity fields are re-calculated based on the new interleaved configuration. Windings may be split in parallel or in series with the press of a single button. Windings can be interleaved in any desired fashion also with the click of a button.

Q. How can I add a shielding screen between the primary and secondary in order to reduce capacitance?

To specify an electrostatic shield, you can use a 1 turn foil winding with the appropriate thickness of wrapper insulation and end margins to satisfy the voltage breakdown requirements between the windings and the shield. You must initially specify a voltage and current; then after the initial design has been made you can edit the turns and wire size to achieve the desired results.

Q. Does Magnetics Designer support half turns?

Cores with center legs have a physical capability for ½ turn increments. Half turn windings only link flux in ½ of the core which results in a very high leakage inductance. Magnetics Designer does not support half turns because of the poor power conversion performance resulting from the added leakage inductance. Future versions will address the problem of multiple magnetic paths that arise in “integrated” magnetic structures

Q. What are the dimensional limitations of Magnetics Designer?

There are no real practical limitations as far as the maximum air gap length, core weight, copper weight, or core dimensions are concerned. There are some reality checks/limitations for different fields such as 1000 pounds for core weight and 1 meter for window length, etc.

Q. Are there limitations on the range of dimensions used to calculate leakage inductance?

Again, there are no practical limits. Short of removing the core, there should be no dimensional-related problems with calculating the leakage inductance. The only limits are those associated with estimating the current distribution.

Q. Does Magnetics Designer use an algorithm to correct for fringing in long air gaps?

It is convenient to set the magnetizing inductance of a transformer by inserting an air gap. When this is done, flux is no longer confined to the core. Flux between the core halves will begin to fringe as the gap is increased so that the effective gap area will be larger than the physical cross section. The fringing flux creates several problems:

  1. For laminated cores, some flux passes normal to the lamination resulting in increased eddy current losses.
  2. Flux will intercept windings transversely, causing eddy current losses in the winding.
  3. As gap length increases, fringing increases its effective area.

Magnetics Designer accounts for gap fringing when calculating the effective gap area used in permeability computations. The resulting area multiplier is also made available to the user in the Kgap function. This function calculates the ratio of effective area to core area according the following:

Kgap(lg) = (1 + lg / sqrt(Ac)*log (2*(lwindow)/(lg+1e-6))) ;

where lg = gap length in cm

lwindow = window length in cm

Ac = core cross section in cm^2

An additional small correction is made using tabular data derived from FEA analysis; this correction can be viewed when performing the field analysis when “Apply using fields” is pressed.

Q. What reports does Magnetics Designer produce?

Magnetics Designer produces two reports, a winding sheet suitable for manufacturing purposes, and a complete performance report providing all operating and functional characteristics of the designed magnetic device. The reports use a text format and may be printed or copied to other Windows applications.

Q. Can the Magnetics Designer be used to simulate electric motors, i.e. DC permanent magnet or switched reluctance motors?

Magnetics Designer was not designed with this in mind, although you may find a way. We recommend using IsSpice4, which is much better suited to simulating motors and electro-mechanical systems. ICAP/4Windows currently contains motor models and its library of motor types is always expanding.

Q. Is the regulation calculated by efficiency or by the vector result of the leakage impedance, primary and secondary resistance and load power factor, as might be required for voltage transformer design.

Magnetics Designer does not calculate regulation but you can set up the User Data buttons to make a calculation of Pcopper/Output power. The power losses take into account the resistance and if you the square wave option, the added losses due to harmonics in the waveform. Generally speaking, it is best to leave this calculation for SPICE to perform, especially in light of the second order effects related to the load. This is true more so for high frequency designs.

Q. I have been playing with the demo and have several questions. The winding fill comes out to be 240.9%. How is this controlled?

The fill is calculated from the obvious things, wire size, turns, wire insulation, and wrappers, etc. You should check out the Core Selection Algorithm detailed in the Demos help. When we say we are "Calculating the build parameters," that's when the fill is being checked. We do try to meet the window fill constraint (Kwindow%), but that is not the overriding factor so sometimes we exceed it without moving you up to the next biggest core. The algorithm is good but its not perfect. sometimes you can even go back 1 smaller core and find a configuration that fits and meets specs by locking the core geometry. All in all the Core Selection Algorithm is very powerful. Setting Kwindow% to something less than 100% (say 70, 80, or 90%) can sometimes "shake-up" the design and cause the program to find a configuration that fits the core you have selected. In the demo, however, you really don't have a lot of geometries which exacerbates the problem.

Q. If I set the Max current Density to a low number after first seeing the currently used current densities per winding, (ex 300-400) and then doing a new apply, the current densities used STILL exceed the max specified. Should there be an error message in the core trials dialog?

The problem is that low current density requires larger wire, but if the largest core is being used, the problem has no solution because the larger wire won't fit. Therefore the program constraint is really the intersection of all constraints. MD doesn't tell the user which constraint controls the outcome when there is more than one. For this particular case, reducing the current density to where the core hits maximum is an unreasonable thing to do... Constraining densities to around 1000 A/cm^2 is about all anybody would do, the copper won't run out of electrons, it just gets hot and MD tells the user how hot it gets.

Q. I noticed that if I locked the core geometry, making changes to the desired Trise did not affect wire gauge or the design in any appreciably sensible way. I would have expected that the User Data window would have been updated with new Trise figures to reflect the maximum target I set in the optional target entry box. It only seemed to change if I allowed the core design to change. Have I missed something?

There is generally but one design per core that gives the lowest loss. The algorithm in Magnetics Designer proceeds from the smallest to the largest capacity core and picks the smallest core that meets your requirements. That is why changing the temperature specification doesn't change the design when you lock the core down.

Q. How high a frequency transformer can Magnetics Designer design.

There is no inherent limitations in the program but there are practical limitations. Magnetics Designer calculates the skin and proximity effects in the windings but does not calculate the eddy current effects or the gap loss. The change in permeability with gap is calculated but the gap loss is problematic and usually requires a finite element analysis for accurate results. Therefore, Magnetics Designer can design magnetics up to the frequency where eddy current losses and gap loss become a factor.

Q. Why are you required to input both voltage and current for primary and secondary (s) etc.? How can you specify an input power less than output power? Surely it would be appropriate to input voltage, output voltage, and power or VA ( or output current and have VA change appropriately )?

Current and voltages are circuit properties. We don't know which windings are connected to the circuit at each instant of time so we can't compute resulting voltages and currents at the other windings. The currents and voltages are used to calculate core and copper losses. There is a User Button Script to address this concern under certain circumstances.

Q. In power supplies it is often required to specify Regulation as a desirable goal to achieve. e.g. AC secondary volts may vary say 5% from unloaded to full load. This seems to be missing - is it something that cannot be identified and quantified with out the load type (resistive, etc.) or can this be entered as a user button.?

The actual voltage drops depend on circuit topology. For example, if you loaded a winding using a 1/2 wave rectifier, connected to a capacitor, the resulting current waveform would have a different shape than for a resistive load. The regulation and terminal voltages would be different for the same input power. We have given a user script capability to get some data during the design and we have provided a SPICE model output so you can get more detailed information. Most people that claim to do regulation deal only with a sinusoidal system with resistive load.

TOP

Program Databases: Cores, Materials, Wires

Q. Can I add my own wire type and sizes?

Yes, you can add a variety of alternate wire types by specifying the proper parameters in an external text file. For magnet wire you can add a series of gauges (non-uniform spacing allowed, for example only even gauges). Requirements are the bare copper area and the total wire area. For foil you need a list of thicknesses. The program will calculate the common mode performance, but not the transmission line effects.

Tables of heavy formvar, small formvar, foil, square, double square, and PCB trace wire are provided.

Q. How are the core, material and wire databases configured? Can I add my own cores and materials?

The core and material database is in the form of an ASCII text file, however, an Excel spreadsheet template is provided to make editing and augmenting of the database easier. You can add your own core geometries and materials although a great variety is already included. Intusoft’s technical support staff will also enter your core/material data at no charge if proper information is provided.

Q. Does Magnetics Designer handle different types of units?

Yes, unit conversion is available for magnetic (gauss/tesla), physical(in, m, cm), and temperature (F/C) related parameters. You can mix metric and English units, using inches for insulation and metric for physical parameters, such as core cross section.

Q. What specs do I need for the PCB trace wire used with planar magnetics?

Thickness of copper, Minimum trace width, Minimum trace spacing, and the PWB Thickness. The dielectric constraint is a default parameter which you can change.

Q. How is Litz wire handled?

The actual stranding configuration of the Litz wire is not accounted for. It is assumed that you will be using a Litz wire configuration such that the AC resistance effects at the operating frequency are minimized. It turns out that when this is done, adding 10% to the DC resistance comes very close to solving the problem for all of the stranding variations (twisting of the internal wire).

For non-toroidal designs, the field-based estimate assumes a Litz strand equal to tLitz diameter. TLitz defaults to .0025 inches or #42 wire. The field simulation will give accurate results for the tLitz size. If tLitz < sd (skin depth) the approximation that Rac = Rdc is pretty good. You can change tLitz in the calculator or by making your own tLitz button.

Q. Can the winding lengths be different for primary and secondaries?

Yes.

Q. How are skin depth or eddy current losses treated?

Eddy current losses in the winding stack are calculated using the equations developed by Bennett and Larson. These equations require an estimate of the magnetic field in the winding stack. The primary method of estimation is heuristic and is based on the ampere-turns of the various windings. A field simulation method is available that takes more time to run but it is more accurate. If a winding has zero or reduced current, the losses are reported in the spreadsheet Proximity Loss row.

Q. Can the wire sizes and dimensions be metric?

Yes. See the options menu for unit changes.

Q. When using a lamination how do I know what stack was used?

In the full program, the EI lam core set is much more extensive providing much higher granularity in the lamination stack. When the design is complete the size of the stack is shown in the core screen. For all types of cores, the program tries to pick the core for you from one for the geometries in the list. We still don't increment the geometry size in 1 lam thickness steps, but we do have a lot of geometries. The algorithm probably should be modified for the EI lam exception and we are thinking about this the next upgrade, but all the core geometry parameters are listed in the core screen.

TOP

Designing Different Types of Magnetics

Q. How can I design a Balun?

Baluns are generally wound using pairs of wires that behave as transmission lines. You can make a wire table that reflects the geometry of wire pairs and magnetics designer will solve the common mode part of the problem. IsSpice can then be used to make a hybrid model, accounting for the transmission line behavior.

Q. Does Magnetics Designer handle Magamp design?

You can choose the “Use Edt for Bac” check box and enter the volt-second blocking capacity for the mag-amp.

Q. Does Magnetics Designer handle flying inductors or swinging chokes (nonsymmetrical windings)?

No, variable geometry magnetic paths are not included in the design and analysis algorithms but it does handle sector (split-bobbin) designs.

Q. Can Magnetics Designer handle Multi-phase (3 phase) transformers?

Multiphase designs can be broken into 3 equivalent single phase designs. You will have to make a special core table to account for the appropriate window and core area. This process will be automated in a future release.

Q. Can I design printed wiring (planar) magnetics? How does Magnetics Designer handle this?

Yes. There is a set of planar cores and 2 printed wiring board files that come with the program. You can add more cores and “wires” to fit your specific requirements. The planar “wire” table includes the geometrical description of the material, including copper thickness, separation between turns, minimum width and substrate thickness.

Q. Can I design an auto-transformer? How?

An auto transformer is a special case of a two winding transformer, where most of the current in the primary is canceled by the secondary. Simply enter the resultant primary current and Magnetics Designer will do the rest.

Q. What kinds of magnetics can’t Magnetics Designer handle at this time?

  1. Variable reluctance magnetic path’s; for example, swinging inductors and resonant voltage regulating transformers.
  2. Motors.

Q. How does Magnetics Designer handle a transformer that operates in a burst mode of operation?

Magnetics Designer assumes that the temperature is constant during circuit operation. It makes this assumption in order to be able to calculate the temperature rise of the core and copper. Magnetics Designer requires that you enter the average voltage, DC and AC RMS currents for each winding. The formulas for calculating these quantities for different wave shapes are available in the on-line help. They will work well and generate the proper current specifications so long as the temperature remains constant. However, when the duty cycle becomes very small the temperature may actually vary over each period. In cases where the temperature varies, the current specifications using a small duty cycle will yield the proper copper losses but not the proper core losses. Changing the average voltage specification should not be done as this will result in an incorrect calculation of flux density and output voltage. Therefore, you should enter electrical specifications that are applicable when the transformer is on and adjust the Kconv (convection coefficient) by the duty cycle. The Kconv variable is made available to the user and can be changed. It will affect how Magnetics Designer calculates the temperature rise and thus the core selection process.

Another case comes about when the system operates in a burst mode. For example, the power supply is on for a some time and then off for some time. Again, as for the previous case, the temperature is not at a steady state level for the entire period and adjustment of the Kconv variable should be made.

A Burst mode is defined as when a device operates intermittently, that is it runs at some frequency and waveform that can be described by Magnetics Designer in terms of a steady state, but for some time period, the device is off. In this mode if we assume that temperature is constant, then the average power is found by multiplying steady state power by the burst mode duty ratio. Then temperature rise, Tr = Kbm * Kconv * P. The user can then lump the product Kbm * Kconv into a new Kconv. This is works because Magnetics Designer always uses the Kconv term to get the average temperature. The Temperature, and winding resistance is iterated to find the steady state winding resistance. Failure to account for the temperature properly will adversely impact both core selection and operating point calculations.

TOP

Losses and Thermal Effects

Q. I have my device mounted to an isothermal cold plate. Can Magnetics Designer account for that?

Yes. Magnetics Designer includes 3 thermal models and provides access to various thermal coefficients. A complete write-up on this topic is included in the on-line help in the demonstration version.

Q. What wire related losses does Magnetics Designer account for?

AC resistance is calculated for either Sine or Pulse waves. The IsSpice model includes predictions for skin and proximity effects and is matched at DC, the operating frequency, and the expected resonance frequency. Loss calculations for pulse waves extend up to the 100th harmonic.

Q. What core related losses does Magnetics Designer account for?

Core losses are estimated using the formula Pcore = Kp Bn Fm Vol
where the coefficients, Kp, n and m are a best fit to manufacture’s measured data.

Q. What is the impact or radial/axial current crowding on AC resistance of stacked foil windings in axially/radially wound transformers?

The orientation of the winding inside the transformer tends to not effect the magnetic (H) field. We use idealized fields to calculate leakage inductance. So basically, there isn't any difference in a PC foil vs. a wound foil. Fringing does effect this, but so far its an intractable problem because there isn’t enough geometrical data present and the complexities of the interaction of the windings, gap, and core.

Q. What is the impact of gap loss on AC resistance of inductor windings?

Air Gaps result in flux fringing which cause eddy currents in laminations and in the wire. Eddy current losses in the windings are available using the “Apply using fields” button. There is no built-in mechanism to calculate eddy current losses in the lamination; however the documentation cites an example for calculating gap loss in laminations. This can be included, along with many other calculations, using the user defined equations feature of Magnetics Designer.

Q. Does Magnetics Designer handle eddy current losses?

At this time eddy current losses are not included in IsSpice model, but they are included in core loss equations since these equations are based on measured data.

Q. Is the temperature rise of the transformer a controlling parameter or a result. Can a design be thermally based?

Temperature rise is a key factor in determining many things including the selected core size. It is a controlling parameter, but it is also an output parameter.

TOP

SPICE Related

Q. Magnetics Designer Outputs a SPICE Model. What SPICE programs will that work on?

The SPICE netlist produced by Magnetics Designer uses SPICE 2G syntax with the exception that some nodes use names instead of numbers for easy reading and capacitors can have negative values. This format can be used by virtually ALL SPICE programs available today including IsSpice4. The models work best with Gear Integration which is included with advanced SPICE programs like IsSpice4.

Q. What effects are included in the SPICE Model produced by Magnetics Designer?

The SPICE model includes an ideal transformer, linear core model, DC and AC resistance that varies with frequency, magnetizing and leakage inductance, and winding capacitance. The ICAP/4Windows system includes a nonlinear saturable core which can be added to the model produced by Magnetics Designer.

Q. How does Magnetics Designer integrate with ICAP/4?

Magnetics Designer stores SPICE models in a form compatible with ICAP/4’s model libraries (ASCII text). Extensions for ICAP/4’s schematic database are included in the model listing (*SYM, *SRC). A special dialog in Magnetics Designer allows you to configure and save a schematic symbol (SpiceNet) of your transformer or inductor.

Q. What other SMPS design products does Intusoft have?

Intusoft offers a full range of analysis tools including analog and mixed signal simulation capability and the industries most extensive array of SPICE models for power devices. Intusoft has models for power semiconductors (IGBTs, power MOSFETS, SCRs, Triacs, BJTs), PWM ICs (nonlinear and state space), PFC ICs, saturable cores, and much more.

TOP

Student and Demonstration Versions, Documentation

Q. What are the demonstration limitation? Can I copy it?

The demonstration version is the same as the production version with the exception that the core database is greatly reduced. The demonstration database includes cores with realistic geometrical characteristics but not exactly matched to any specific part number. The material characteristics are exactly matched, however. The production version also allows you to add and modify the core database. The demonstration version does not. You can copy the demo version so long as it is copied in its entirety. You can not change or add to the demonstration core database, however you can set the core and material parameter directly inside the program. The reports, symbol, and SPICE netlist outputs are also enabled.

Q. Where can I get a demo of Magnetics Designer?

A demo version of Magnetics Designer is available on Intusoft’s web site http://www.intusoft.com.

Q. Is there a student version of Magnetics Designer?

The demo version can also be used as a student version. Since the program is fully functional it can be used to teach transformer and inductor design concepts and techniques.

Q. Is there any program documentation available?

Yes. The demo includes the FULL manual set including all the technical information and program usage details in the on-line help.

No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without written permission of Intusoft. Intusoft, the Intusoft logo, Magnetics Designer, IsSpice, IsSpice4, SpiceNet, IntuScope, and IsEd are trademarks of Intusoft. All company/product names are trademarks/registered trademarks of their respective owners. Windows, Windows 95, and Windows NT are trademarks of Microsoft Corporation.

TOP


(C) copyright Intusoft; all rights reserved.