...
Rev1 Testing
During testing, multiple issues were found with this design. Most notably the P-Ch MOSFET Q1 had the drain and source wired in reverse, and the latch latched high in the beginning. This is an issue as initially when the HV battery is off and power supplied to the board, pre-charge will close the main relay as it is currently configured. To fix this, it is proposed that an AND gate be added to the comparator output and the negative input of the comparator. This way, the output will only be high if the comparator is high AND the battery relay is closed.
By observing the opto-isolator output, it was observed that during the transition from high to low to close the main relay, it bounced between the two states multiple times for more than a second. This was remedied by adding a positive feedback between the comparator output and Pins 5/6 as recommended in Figure 1. As seen in the plot below, bouncing was significantly reduced with 240mV below 12V of positive feedback.
From this, we can also see that after adding hysteresis, switching takes almost double the time. This can be remedied by either unbalancing the resistor divider and moving the switching point earlier, or experimenting with different values of hysteresis for different voltage thresholds. Because the unique way in which it is implemented as recommended by TI, adding trim potentiometers and slightly unbalancing the resistor divider is the easier and more reliable solution.
Changes for Rev2
- Fix diode silkscreen on the bottom of the board
- Remove R2, replace R1 with a smaller value
- Q1 should have 1/2/4/6 and 4 swapped
- Ultra-fit and Dura-Clik connectors instead of Clik-Mates
- Consider adding XOR AND gate between Vcc negative input and Comparator output
- Add passthrough for relay sense line (shorted when relay is closed, open circuit when relay is open)
- Fix logic issue with the latch
- Add inverter or source opto-isolator that does not invert the logic
- Fix voltage drop with MOSFET stages
- Add hysteresis to avoid bouncing