Balun design for differential PA output at 2.4GHz?

Need a balun for differential PA to single-ended antenna at 2.4GHz. Marchand vs LC balun - what do you use in production? Insertion loss budget is tight.

For a differential PA to a single-ended 2.4 GHz antenna with a tight IL budget, production choices usually land like this:

- Marchand (coupled-line) on PCB: good bandwidth and repeatability if your stackup is controlled; IL often ~0.5–1+ dB depending on loss tangent and length. Needs real estate and a well-defined ground reference.
- LC lumped balun: smaller, cheaper, narrower band; IL can be competitive if inductors are high-Q, but part tolerance and self-resonance hurt yield at 2.4 GHz.
- Discrete transformer / LTCC balun: common in modules when you want a vendor-specified IL and balance; watch power handling and amplitude/phase imbalance into the PA.

Before picking topology, measure or simulate balance: 0.5 dB / 10° imbalance already hurts even-order and load-pull symmetry. Also confirm the PA wants a real differential load — some "diff PA" eval boards already include the balun, and swapping topology changes the load the transistors see.

Note: this thread is under Test; for layout deep-dives the Components / PA boards may get more eyes.

Open question: in your production builds at 2.4 GHz, what insertion-loss number do you actually budget for the balun + matching, and which topology hit it most reliably?

— rfforum 编辑部 / Editorial Team