The following information provides helpful tips on the installation and use of the PI Expert power supply design software.
I. General
II. Setup and Installation
III. Power Supply Design
IV. Optimization
I. General
What's new in PI Expert Online?
 Access though your web browser – no software installation required
 Ability to receive updates and new features as they are added
 Full PI XLs integration – running a second program is no longer required
II. Which browsers are compatible with PI Expert Online?
The following table shows PI Expert Online browser compatibility.
Browser

Version*

Supported

Limited Support**


Internet Explorer  11  X  
10  X  
9  X  
Google Chrome  33  X  
32  X  
31  X  
30  X  
Firefox  28  X  
27  X  
26  X  
25  X  
Opera  20  X  
Safari  5.17  X 
*PI Expert Online will aim to support new versions as they become available.
** PI Expert requires HTML5. Older version browsers may not be
fully compatible.
III. Power Supply Design
 What is "Optimize with Fixed Key Parameters"?
PI Expert allows the user to constrain the optimization engine to optimize with certain parameters being fixed to user selected values. This feature is useful in cases where for example the user wants to use the power of the engine but subtly make certain decisions along the way. For example, you can request the engine to go through the optimization routine but lock the VOR and/or the KP to a specific value. The engine will go through several iterations and present the top solutions with the imposed restrictions. Note that in certain cases you can impose too many constraints on the engine, which can cause the optimization routine to render no optimized results.
 Has the file extension for PIXls designs changed?
Yes. The file extension for all PIXls spreadsheets is now a single extension filename.pixls When you convert your old designs this change will automatically take place.
 Why does PI Expert and PI Xls use multiple parallel wires in transformer constructions?
PI Expert and PI Xls both use multifilar windings. This means that rather than use one thick wire it uses two or three (or more) parallel stands of thinner wire. This minimizes skin effect losses, fills the bobbin width, lowers leakage inductance, and thus provides a more optimized design.
 In PI Expert and PI Xls, what is "maximum terminations per pin?"
PI Expert and PIXls both use multifilar windings in the transformer design. This means that rather than use one thick wire it uses two or three (or more) parallel stands of thinner wire. As more and more parallel wires are used, it becomes difficult to terminate all these wires onto a single pin especially if the wires are thick and the pin is thin as is the case with small bobbins. In this case it is better to split the wires between 2 or 3 different pins for the terminations. The "maximum terminations per pin" specifies how many wires can be terminated onto a single pin and the default is set to 4 wires per pin. For thinner wires this number may be as high as 6 and for thicker wires this number may be as low as 2.
 The terminology used within PI Expert is unfamiliar to me. Where are the definitions for these terms?
There is a Help button located on the toolbar and popup window within PI Expert or you can press F1 function key at any time. Selecting the Help button will launch the Help Utilities consistent with that form or window. You will find answers to common questions, terminology and guidelines to software usage within the individual Help sections. For further details on general power supply terminology we recommend that the user read the applicable Application Notes for the selected devices. (See Help for details).
 How can I specify a design with a negative DC output?
PI Expert allows one negative output when the total number of outputs is 2 or more (the negative output cannot be the main output). To work around a design where a single negative output is required simply enter the negative DC output as a positive value. Make sure you remember which output is negative when you determine the transformer pinout and resulting PCB layout.
 What determines Transformer maximum power capability (PMAX)?
Transformer power capability is determined using the Area Product method (Ae x Aw). Efficiency, device switching frequency and transformer margin are factored into this calculation.
 What is "Margin on Left" and "Margin on right"?
PI Expert and PIXls provide more flexibility in specifying the margin on which side of the bobbin (left and right side for horizontal bobbin and top and bottom side for vertical bobbin). This is particularly useful in specifying asymmetrical margins. For PIXls the parameter M represents half the total margin that is required. Thus if M is specified as 3.0 mm, the software will assume 3.0 mm on the left (or top) and 3.0 mm on the right (or bottom) of the bobbin.
 I noticed some inconsistencies with the gapped core effective inductance calculation (ALG). Is this a bug?
PI Expert uses noninteger primary turns within the ALG calculation. In most cases, since primary turns are large, the resulting error is minimal. Please make the necessary changes to the ALG value when communicating with your magnetic vendor.
 What is the basis for the default efficiency estimates?
Default efficiency is based on the efficiency curves presented in AN21 (TOPSwitchII), AN26 (TOPSwitchFX) and AN29 (TOPSwitchGX). TinySwitchII efficiency is estimated and based on the evaluation of actual power supplies. These curves estimate power supply efficiency given AC input range and output voltage. For output voltage between 5 V and 12 V, efficiency is estimated using linear interpolation. PI Expert estimates efficiency deviation for outputs outside the 5 V to 12 V range.
 Can PI Expert help me with package choice and thermal evaluation?
PI expert estimates the heatsink size that will be required for the PI device as well as the secondary side diodes. This size is a function of the type of heatsink used namely copper on PCB, Aluminum sheet metal or Aluminum extruded heatsink.
For copper on PCB heatsinking it is assumed that a square area is used for heatsinking. The shape of the area has a large effect on the effective thermal resistance of the heatsink and therefore the usefulness of the heatsink.
For Al sheet metal type heatsinks a rectangular area of height 20 mm is assumed. The calculations are based on material which is assumed to be Aluminium Alloy (3003 or 5052) with a thickness of 1.6 mm.
For external extruded heatsinks, the software derates the datasheet thermal resistance by 20% and then makes recommendations on required size.
PI Expert can be used in conjunction with AN21, AN26 and AN29 to gain insight into package dissipation. In general, Ypackages and Epackages with suitable heatsinking should be considered when device dissipation exceeds approximately 1.5 W (open frame) or 1 W (adapter/enclosure); assuming a 50 °C ambient temperature.
Thermal design can be further hindered by extreme operating temperatures, poor layout, high altitude, inefficient transformer design and/or airflow limitations. When used near maximum current capability, Power Integrations recommends a maximum operating junction (die) temperature of 110 °C for all TOPSwitch, TinySwitch,LinkSwitch and PeakSwitch products. This generally provides adequate design margin to the minimum device thermal shutdown, taking into account device and unit to unit variations.
IV. Optimization
 Optimization is not solving with the input Capacitor, TOPSwitch/TinySwitchII and/or Transformer that I wish to use. How can I work around this problem?
If the output power is beyond the capability of the device in question, the Optimization tool will choose the device combination which best suits your requirements. This choice is dependent on Optimization Parameters section within the Active Design on the main menu. If Optimization chooses a device, which is smaller than the preferred device, the user may force the preferred component choice(s). Simply select the desired device from the PI device selection dialog and then from the solutions filter dialog make sure that the core start entry and the core end entry is the same and is the core that you want the software to design with. Then click on the optimize button and allow the software to present solutions with the additional constraints of device and transformer size.
Transformer power processing capabilities are derived using the areaproduct method (Ae x Aw). Power supply efficiency and switching frequency are factored into this calculation.
In all cases, Optimization will choose the optimum input capacitor. The user cannot override this component choice. The user should pursue manual design if an alternate input capacitor is desired. Alternately you can enter a DC input voltage corresponding to the input capacitor average DC voltage and ask the software to redesign with this specification.
 Why do I get the same efficiency estimate for Cost and Efficiency Optimization?
PI Expert assumes a current waveform parameter consistent with AN21, AN26 and AN29 (varies according to power requirements, AC line voltage and PI Deviceused). Therefore the efficiency estimate does not take account of the change of optimization goal. In practice a prototype using a transformer design optimized for efficiency should have a higher efficiency than that optimized for cost.
 What is Cost Optimization?
During the cost optimization process a number of designs using different combinations of cores, secondary turns, secondary output stacking configuration and output diodes are scored based upon criteria internal to the software. A list of the top scoring designs is maintained throughout the optimization process. Upon successful completion, the list of top designs is displayed. You can then choose the design which you believe to be the most appropriate for your application.
Cost optimization is based on the following key concepts:
 The smallest PI device that is capable of delivering the power is chosen first.
This step is not simply a check to ensure that the device's rated power, taken from the data sheet, is greater than the power specified. This first step in optimization takes into the account a number of key operating parameters, which include the maximum duty cycle (DMAX), the peak primary current (IP), the reflected output voltage (VOR), and the peak to ripple primary current ratio (KP).  The smallest transformer core that is capable of delivering the power is then chosen.
Similar to the selection of the proper PI device, the selection of the transformer core is made using a number of key operating parameters. Some of these parameters include flux density (BM and BP), gap length (LG), primary layers (L) and the physical dimensions of the windings within the bobbin window fit factor (FF)
 What is Efficiency Optimization?
Efficiency optimization is available for the TOPSwitch and DPASwitch family of devices only. There are two key factors that are crucial to understanding the efficiency optimization method:
The PI device capable of delivering the specified power with margin to the device's current limit is chosen.
The PI Expert efficiency optimization routine will first find the smallest PI device capable of delivering the power specified. The software will then reduce the peak to ripple primary current ratio (KP) in an attempt to make the design more continuous. This decreases the peak primary current (IP) which will decrease the peak and RMS currents in both primary and secondary windings. This reduction in current magnitudes increases the overall supply efficiency due to reduced conduction losses.
The core capable of delivering the specified power efficiently is then chosen.
This step in the optimization routine ensures that the core selected can not only deliver the power but can also do so without significant losses. The losses accounted for are core losses as indicated by flux density (BM), copper losses as indicated by primary layers (L) and primary RMS currents (KP) in addition to leakage inductance indicated by gap length (LG).
As in cost optimization a number of designs are then examined and passed on to the optimization process. PI Expert then retrieves a list of the best scoring designs and then displays it.
 How does Optimization work?
Optimization generates multiple design solutions and compares to an expert design rules database, compiled by the Power Integrations engineering staff. The Optimization engine seeks the minimum design solution that meets, or exceeds, the limits specified in this database.
Power Integrations recommends that all design solutions be constructed and performance verified to your specific requirements. This includes independent verification of safety, thermal and system reliability.
 Cost Optimization provides a result, which generates the warning "Core size may be too small for this power level (Po)". Is this a serious problem?
This message may appear following successful Cost Optimization. Cost Optimization will consider transformers with power ratings (PMAX) within 90% of specified output power. This warning states that the design, running under continuous rated power, is a approaching the maximum capacity for that core/bobbin size. In these cases, further evaluation of power supply thermal performance is recommended.
The user may want to consider a solution with a larger core. To generate these results, you must set the desired core in the core start and core end drop down boxes on the "Solutions Filter" dialog during the optimization run.