I designed a LiFePo4 charger based on LTC4013. The schematic is similar to the one of the demo-board (DC2374A Demo Board), with the required adjustments, and I encountered the following problems and I need some help:
- Unreliable start-up: when connecting an illuminated solar panel, the INTVCC voltage will not rise to the nominal value (5V) and the charge pump is not enabling the input MOSFETS. If the panel voltage rises slowly from 0 to nominal, the component does start, with 5V on the INTVCC pin and the MOSFETS are properly driven. The same happens with a bench power supply. If the voltage is ramped up slowly, the component starts normaly. If connected directly to the nominal input voltage, start-up fails.
- MPPT Function inactive: the component does not regulate the input voltage to the programmed value. It oscillates around the Enable voltage, with audible whine (with variations, depending on the feedback compensation network).
Here are the design parameters:
- Solar panel:
- Open circuit voltage (Voc): 22.2v;
- Maximum power point voltage (Vmp): 18.1V
- Maximum power point current (Imp): 1.12A
- Enable: programmed at ~13.5V, using a 120K and 12K resistor divider.
- Mode setting: Li-Ion, Mode1 High, mode 2 float.
- Timing: no timer capacitor, TMR connected to GND.
- Vin connected to VIN_S.
- MPPT and FBOC: programmed using a 545K, 9.1K and 40.2K, using the design example from the datasheet (page 29), with ~40uA divider current.
- Output: set to ~3.6V, using a 110K and 186.8K divider.
- Compensation loop: calculated as instructed on page 26. Tested with multiple values. No matter the values, the input voltage drops (during operation) to the enable voltage.
- Switching frequency: ~500KHz;
- Low battery setting: ~2V, set by a 82.5K resistor.
- Charging current: 3.3A, determined by the 0.015R resistor.
Here is what I attempted so far:
- various values for the enable divider;
- various values for the compensation network;
- various capacitance at input and output;
- various illumination; tested with solar panel (portable) and lab power supply.
- boards, same behavior.
The schematic and the PCB are attached below. The PCB is made on 2 layers, 35um copper thickness.If necessary I can provide the complete project (in Altium Designer). The board also has other circuits on it, but they were not populated.
Obviously, I am missing something, or there is a mistake in the PCB that I cannot see. Has anyone encountered such problems with this component ?