TPS2121: Slow Switchover Time with Load Connected

Oscar Qu

Part Number: TPS2121

Tool/software:

Hello,

I am experience a slow switchover time (around 25 ms) when using the TPS2121 chip. We are currently using the XCOMP mode and prioritizing IN1 over IN2. During testing, IN1 is 12V and IN2 is 16.8V. According to the calculator, PR1 and CP2 are above 1.06V, meaning fast switchover time should be enabled. When no load is connected, the chip is performing as expected. However, when we are drawing more than 1A, when IN1 and IN2 are both on and IN1 is turned off, Vout experiences a drop to 0V and stays there for around 25 ms. I've tried changing the capacitor connected to SS to 0.1 uF, which helped a little but the switching time is still slower than expected. We've also tried adding capacitance to the output which didn't help. Something interesting that happened was that when we proved the PR1 and CP2 pins on the oscilloscope, the timing changed and wasn't consistent when we kept on testing. In the picture, the blue waveform is IN1, green is IN2, and yellow is VOUT. Would anyone have any ideas about what could be happening? Thank you in advance!

11 months ago

0 Kunal Goel 11 months ago

TI__Guru* 79181 points

Hi Oscar,

Let me check and get back on this.

Regards

Kunal Goel

0 Oscar Qu 11 months ago

Prodigy 10 points

For context, here are some more scope screenshots. The first screenshot shows the baseline when no load is connected. As far as we can tell, it is operating as intended under these conditions.Channel 1 is Vout of the chip, Channel 2 is Vin1 (12V source), Channel 3 is Vin2 (16V source), and Channel 4 is the ST pin.

In the second picture, we have connected a load drawing around 1.4A. After IN1 and IN2 are both on and IN1 is being prioritized, we turn IN1 off. At this point, the chip should seamlessly switch to IN2 (so from 12V to 16V output). However, there is a 25 us delay which can be observed. Here, Channel 1 is Vout, 2 is IN1, 3 is PR1, and 4 is CP2. We then tried adding input capacitance, and output capacitance to no avail. We have also tried desoldering the pullup for ST and using hysteresis of a 1MOhm resistor from ST to PR1, which still didn't fix the issues.