Use H Bridge to Switch Transformer Load to Generate 100khz Carrier on Primary Coil?

An H-bridge with a 50% duty cycle creates a balanced square wave into transformer primary from a single supply voltage. If OTOH you ground one end of a transformer primary, you'd need a bipolar square wave to avoid saturating the transformer with DC current, which means you'd need positive and negative supplies. Sometimes folks use a capacitor in series with the primary. Besides possibly eliminating the need for a balanced pair of supplies, this also eliminates the need for an exact 50% square wave to avoid any DC current.

In a half-bridge configuration (right) the transformer winding is returned to the tap on a series pair of capacitors. As the half-bridge operates (with a roughly 50% duty cycle) the tap is driven to half the supply voltage. In many designs these capacitors are the bulk storage caps for their rectified AC power.

If you don't want to use an H-bridge, you'll need two switching transistors (one HI and one LO) anyway, and you may need a bulk cap as well, which means you can get a half-bridge virtually for free. The question then becomes, Why use an H-bridge instead of a half-bridge? I could answer that, but I've run out of time!

Thanks Win,

Your comment is helpful. Regard your last question: Why use H bridge instead of half bridge? Can you share the answer to this?

I can understand the unipolar/bipolar drive thing (dc cap n such) but what I really interested to find out before building the PCB is what difference to the outgoing coupled signal on the secondary side for both methods? How does H bridge drive affect the voltage and current signal on both the twisted pair wires? (does H bridge generate more current on the return/gnd wire?)

Do I need modifiy the noise filter attached to secondary coil (used in conventional non H/half bridge method) which is connected to the twisted pair network if In switch to H/half bridge driver?

Regards/CW

The full H vs half provides twice the drive voltage across the winding.

In full H you have dot to V+ and non-dot to V- then non-dot to V+ and dot to V-.

in half H you only get gnd±V across the winding.

H-bridges are commonly used for higher-power (>250W) or efficient power supplies (with Vdc above 100V or so, where the switch voltage drop can be small compared to Vbulk). As such they're in a region where any power-loss improvement matters, even if a few more parts are required.

For example, an H-bridge has the same current drain from the bulk capacitance each half-cycle, whereas half-bridge has twice the current drain from one cap for a half cycle, then twice the current drain from the other cap. The caps work at a lower voltage, but two are needed, hence we may not get to enjoy a lower esr. At any rate, P=I²R tells us we'll have 2x or higher cap esr power loss. This can be thought of as a consequence of using each cap only half the time.

A similar argument goes for the MOSFET or IGBT switches, which are used at higher currents (yet need the same voltage rating), and for only half the time in a half-bridge. Once again I²R tells us we'll have 2x higher R_DS(on) power loss.

regsoft has answered your other questions. The transformer secondary and output filter is unchanged, in general, as they have no idea whether the input is a half-bridge or full bridge. It's possible that inter-winding capacitance-feedthrough issues favor the full bridge configuration.

Many years ago, I built high power devices using H bridge across either 28 VDC or 400 VDC. The problems of turning off either the high side or the low side before turning on the other were annoying at least. Perhaps with the devices built since my turn in the barrel are faster or more tolerant, but I had to conciously drive one off before the other was turned on. The 400 V bridge ran up to 15 KHz at 5 KW with 12 silicon transistors that were at the time $125 each. Not only was failure spectacular, it was expensive.

Luck whichever way you go.

Myself since you need to get up to 100 KHz, I'd personally use a custom transformer with a center-tapped primary and drive it push pull.

Thanks Win, Bill,

Learned alot from this thread. Thanks for guidance.

Regards

Chee Wai