Hi,
We made a design to charge a wearable based on the LTC4125 (TX side) and LTC4124 (RX side), and we're having issues making it work properly.
RX side LTC4124 is configured for 25 mA, with 760308101216 coil (7.2 uH) and 91 nF capacitor for 196 kHz resonant frequency. TX side LTC4125 uses 760308101208 coil (13 uH) with 47 nF + 3.9 nH for achieving 196 kHz resonant frequency.
Basically what happens is that, when the RX device coil is placed right on top of the TX one (touching each other), charging works but both coils and LTC4124 heat up significantly. Input current was measured on TX and it's between 0.8 and 1.2A, which sounds too high for the RX current demand.
The moment you leave even an 1mm gap between the coils, power delivery stops. In our application, there is a thin plastic housing between devices so it's mandatory that power delivery works with a small dielectric gap.
The behavior we see when no receiver present is that every 5 seconds (configured time by CTD), a short search ramp starts increasing tank voltage until FB voltage reaches VIN (5.25V), while an square waveform of 196 kHz with varying duty cycle can be observed at the coil terminals during the ramp. If the receiver is present, VFB stops at around 4V and power delivery starts, seeing a fixed square waveform of 196 kHz at the coil terminals. I understand that this is the intended behavior, but:
- We can't get it working without the coils touching each other,
- Input current is too high, causing significant temperature rises.
Here is our schematic. I used the one on the LTC4124 datasheet as reference.
Another strange thing is that, according to some other posts around here, having the voltage reference on IMON should fix the duty cycle, but what I observe is that the search ramps still happen. Maybe the fact that we have that and VPTHM tied to GND is causing issues?
Things we've tried:
- Changing feedback resistor Rfb2 by higher and lower value. Same behavior.
- Using an LTC6990 based TX with the same coil, working at 196 kHz. Works a bit better than the LTC4125 design but mostly the same.
- With the LTC6990, changed resonant frequency on both TX and RX to 270 kHz by replacing tank caps, closer to the RX coil's Q peak. -> Works much better, charges properly through the plastic housing. This makes me think that we need to increase resonant frequency on the LTC4125 to 250 kHz, the max that the IC allows.
Any ideas?
Best regards,
Enrique
