Hi all, working on a half-bridge series resonant converter for wireless power transfer (~104kHz resonant frequency, 12V bus, SG3525 PWM controller + UCC27714 gate driver, SiZ319CDS FETs). Two questions - one about a gate drive issue and one about the coil waveform.
Gate drive issue (CH1/CH2)
The setup is two complementary SiZ319CDS MOSFETs in a half-bridge driving a PCB spiral coil as the resonant inductor. Gate resistors are 5 ohm in series (R33/R34) and 10k pull-downs to COM/GND (R31/R32).
When I scope the low-side FET gate (CH2) the off-state sits at ~2.4V instead of 0V. High-side (CH1) looks clean and switches between -0.8V and 12.8V. Both FETs are identical with symmetric pull-downs so I expected them to behave the same.
My theory is dV/dt induced false turn-on (Miller effect). When the high-side turns on, the switch node slews rapidly upward and couples charge through Q8's Cgd onto the low-side gate. Since the low-side source is fixed to GND, the full gate lift appears as Vgs and pushes right up to the threshold voltage (~1.5-2.5V for this FET). The high-side doesn't show the same issue because its source floats with the switch node, so Vgs stays roughly constant during the transition - effectively self-compensating.
The 10k pull-down isn't helping because the Miller transient is only ~2-5ns wide but the RC time constant is 10k x 500pF = 5us - way too slow to bleed the injected charge in time. Planning to drop R32 from 10k to 3.3k as a first fix.
Coil waveform issue (TP4)
I also scoped the switch node / PCB coil (TP4) and the waveform looks nothing like the sustained sinusoid I was expecting from a resonant converter. Instead I'm seeing short damped ringing bursts (~10-15ns duration) at roughly 104kHz repetition rate, with the signal sitting near 0V for most of the period (~85% of samples are below 0.3V).
The gate drive switching frequency from the SG3525 is currently ~71kHz (pot at max resistance) vs the designed resonant frequency of ~104kHz. My current theory is the converter is operating well below resonance so the tank never builds up sustained oscillation - it just gets a short kick each cycle and damps out before the next one. The burst repetition rate matching the resonant frequency rather than the switching frequency also suggests the tank is ringing at its natural frequency rather than being driven continuously.
Questions:
- Does the Miller diagnosis look correct? Is 3.3k the right pull-down value or would a dedicated Miller clamp be better?
- Does the coil burst waveform match what you'd expect for a resonant converter running well below its resonant frequency?
- Any other reasons a series resonant WPT tank would show this burst behaviour rather than sustained oscillation?
Attaching scope captures of both gate signals (CH1/CH2), the coil switch node (TP4), and the schematic. I'm a newbie electrical engineer, so any input would be very helpful. Thanks!
