WEBENCH® Tools/UCC28C42: Custom part numbers?

Hello Vishal,

Thank you for your interest in designing with the UCC28C42 PWM controller.

The Webench design tool is intended to produce a design that will work, based on the input and output information provided, but it is not sophisticated enough to produce an optimal design. Essentially, it solves a series of design equations for each component and tries to match the calculated parameters to a database of parts to select the closest fit. However, the Webench database is relatively limited compared to the inventory of a parts distributor, for example. In some cases, as in the case of the output diodes, there may not be a suitable part available from the database, so the Webench program provides only the parameters necessary so that the user may find an equivalent part that is readily available. I don’t know why the program uses 2 diodes in parallel, unless it did so to minimize losses.

Similarly, I think a 2N2222A was designated because of a limited set of small-signal transistors in the database. You are not required to use this exact part. An experienced designer knows that every one of the parts in this design have multiple alternatives: either for sources of parts with the exact values, or substitutes with similar values with perhaps some other advantage such as price or availability.
A surface-mount substitute for the 2N2222 can be an SMD version of the 2N3904 or 2N4400 devices to name a few. I don’t know the actual SMD part numbers off-hand, but an on-line search should turn up several vendors and searching their product line will turn up many possibilities. In this application, the transistor needs minimum ratings of 10V Vceo and a few mA of collector current.

As for designing an input EMI filter, one must first determine whether such a filter is even necessary based on conducted noise-level specifications that must be met. With a low-voltage DC input, any noise specs are likely to be different than those usually required for an off-line AC application.
Then the amount of noise generated by the converter must be measured to determine the amount and type (common mode and/or differential mode) noise that needs to be attenuated. Finally, a filter arrangement is selected and values calculated to achieve that level of attenuation. Her is a link to some training in EMI design: training.ti.com/.../understanding-emi-and-mitigating-noise-dcdc-converters input line filter&tisearch=Search-EN-Everything
This training may contain more information than you need, but it should give you a good overview of what is involved.

Regards,
Ulrich

Hello Vishal,

The dot notation on schematic diagrams of transformers is used to indicate the polarities of each winding with respect to each other. Using dots (other symbols may also be used, but dots are the most common), the voltage polarity of all of the windings can be established when the voltage of one winding is known. The voltages across all of the windings can be either positive or negative at any given time, but those voltages follow the polarity marks.

It doesn’t matter if the transformer is used in a flyback topology or a forward-mode topology, when a voltage is forced across a winding (such as when the main transistor is turned on and bulk voltage is applied to the primary winding), if the applied voltage is positive at the dotted end of the winding and negative at the non-dotted end, then all the other windings will be positive at their dotted ends. The opposite can also be true: if the applied voltage is negative at the dotted end of the winding and positive at the non-dotted end, then all the other windings will be negative at their dotted ends. It also does not matter if the dots indicate positive or negative voltage as long as they are all of the same polarity, but usually (although not always) a positive voltage is ascribed to the dot.

The dots are used to indicate winding polarity on paper, for convenient reference. In a physical transformer, the direction of winding construction determines the relative polarities of each winding. Usually, the paper schematic is drawn to indicate the desired circuit function and the physical transformer is constructed to match it. Clockwise and counter-clockwise winding directions with respect to the core are established to obtain the required polarities. They are based on the “right-hand rule” for flux direction with respect to current direction. This rule says that when you curl the fingers of your right hand into a fist and stick your thumb out, current flowing in the direction of your thumb (as in a wire) generates flux around the current in the direction of your fingers. The converse is also true: flux pointing in the direction of your thumb is established by a current flowing around the flux (as in a coil of wire) in the direction of your fingers. If you look at your right-hand thumb pointing at your nose, your fingers are wound in a counter-clockwise direction around your thumb.

To picture a winding structure more fully, I sometimes draw a vertical core on paper and then overlay each winding along the core. Start with a main winding, such as the primary (for example), and assign the start of the winding to some pin number and the finish of that winding to another pin number. Assign the direction of flux in the core as point up to the top of the sheet for a positive voltage applied to the winding. Lay your hand on the sheet with thumb pointing up in the direction of the flux in the core. Your fingers point in the direction of the wire winding necessary to generate that flux based on a current flowing in that direction. Beginning at the start pin, wind the wire counter-clockwise across the core, going down into the sheet on the right side, coming back under the core and coming up out of the sheet on the left side to complete one turn. Continue with all of the turns needed from the start pin to finish pin for that winding (or winding segment for split windings). All other windings wound in the same direction will have the same polarity as the first winding. Those wound in the opposite direction will have the opposite polarity.

Following this method with your transformer design, if we arbitrarily assign a dot at “start” pin 3 of W1, then a dot is also placed at “start” pin 2 of W4, since it is wound in the same direction as W1. Since W2 is wound in the other direction, the dot is placed at “finish” pin 4. Similarly, a dot goes to “finish” pin 10 of W3.

I believe it is good practice to VERIFY that your completed transformer matches your paper design by applying a high frequency sine-wave voltage from a signal generator to one of the windings and checking the other windings for voltages of the expected polarities, appropriately scaled by the expected turns-ratios. Only a 1V (or a few volts) peak is necessary, at a frequency higher than 10kHz (like 50~100kHz) to avoid loading down the signal generator due to low winding impedance, and to avoid saturating the core of very small transformers.

A treasure-trove of information on magnetics can be found in the TI/Unitrode Magnetics Design Handbook at www.ti.com/.../slup132.pdf , but ironically little is discussed about polarity and winding direction. Some mention of these are made in the reference articles R3, R4 and R6 near the end of the handbook. More reference sources on polarity fundamentals can be found by searching the web for “magnetic dot polarity” or other similar search string.

Regards,
Ulrich

Hello Vishal,

Your polarity table and the dot notation in the schematic coincide.  I see no problem with it.

When the primary FET U2 is on, there is +Vbulk at pin 3, and all of the dotted ends will have a negative voltage with respect to the other ends.  When U2 turns off and the core demagnetizes, all of the dotted ends will have positive voltage with respect to the other ends.  This is completely normal and proper for a flyback transformer. 

Good luck with your design.

Regards,
Ulrich