SMPS Push-Pull Transformer Calculator
Another Transformer Calculator
Many times I've had to go through old documentation to figure out how to wind a transformer. I decided to put the calculations in a web-based calculator. It is a simple calculator - it just calculates the number of turns. You have to figure out how big the wire needs to be and if it will fit on the bobbin.
The two things you need to know about the transformer are the Bmax, which you can generally guess at without too much trouble, and the cross-sectional area in cm2. Bmax is the maximum flux density you want in the core. 1500G with a 3622-77 pot core at 25kHz will produce 0.68W in core losses. Lower the Bmax as the frequency increases. Don't use type 77 over 100kHz. For type 77 it looks like you can't saturate the core if you keep Bmax below 3000, but then those core losses will get you. One-half to one-third of that is more appropriate. Your turns count will increase as you decrease the Bmax. So will your winding losses. Ae is the cross-sectional area, and that is always in the datasheet. If it is given in mm2 divide by 100 to get cm2.
Pay attention to the switching frequency. If you are using something like an LM3524D, the frequency it runs is twice the actual transformer frequency. If it has a clock frequency of 50kHz, the transformer is only running at 25kHz. If you don't take that into account you will design a transformer that is very much too small for the frequency.
When I get a fractional turn on the primary, I round up. It is safer than rounding down. Higher turns on the primary means more inductance and less current at the same frequency. If you round down, you may increase B to the point that the core saturates or overheats. You'll also have to decide if you need to round up any secondaries, since rounding the primary changes the number of turns required on the secondary.
The schematic above shows the transformer for a push-pull forward converter. Leave off any secondaries that you don't want. If you are doing a flyback, all bets are off. The transformer calculations aren't the same.