My 100uH inductor is quite good, its a toroid, with what looks like a bifilar winding (it's partially potted). Its leakage inductance is only 0.6uH.
My loads are symmetrical, but I still worried about the possibility of poor tracking (need at least 5%).
Phil Hobbs wrote:
I had designed a PCB meant for the 260kHz LM2671 Simple Switcher, with an LPT-4545 100uH inductor / transformer. It's similar to my sketch, except the IC's ground pin is on the Vneg output rail. Only the negative output is controlled.
Finishing a set of measurements, I'm pleasantly surprised by how well it works. With no load there was some excess voltage on the pos output. From 5mA to over 300mA (my test range), the neg output stays at the setpoint, as expected. And the positive output starts better than 0.1 % at 5mA, and degrades by only 0.4 % at 300mA. This is damn good, far better than my requirement.
In addition to my low leakage inductance, perhaps my high 100uH magnetizing inductance helped, by insuring CCM and reducing the current ripple.