Hello All,
I've used the TPS65185 in a design solution to drive an E-Ink display. I've attached the schematic I've created. Note that the inductors used are what is recommended in the TPS65185 datasheet.
A first attempt development board worked fine, it drove the E-Ink display with no issues.
A second form factor board for the project was made, and under no load (no E-Ink screen attached), the PMIC draws very little current, as expected. All rails regulate and the circuit seems to be operating successfully. When loaded (display connected), the TPS65185 will draw over an amp, and then drops to ~500 mA. This is in a controlled case, with just the TPS chip and its required circuitry populated on the board.
I noticed putting my hand or my head close to the board seems to reduce the current draw, down to 100 to 200 mA. This seems very high as well, but leads me to believe there is some issue with the PCB layout. I've also attached labeled screenshots of the Top and Power layers of my PCB I've made. Note that the stackup is 4 layers: Top, Ground, Power, Bottom. The ground layer is a completely solid ground plane underneath the top layer (minus the via holes not connecting to ground).
I've probed the switching nodes and compared them to what is expected from the datasheet, and the seem to look ok. I also don't believe it is the E-Ink display being faulty, as the same display can be used with the previous prototype layout and works fine.
Other observations: I can hear the inductors whining, and the original failure mode was a 3.3V LDO dropping out of regulation because the TPS chip was drawing too much current. I noticed as I put my ear up to my board, the frequency at which the 3.3V rail would oscillate (causing the TPS chip to repeatedly turn on and off) would increase and become stable (drawing less current from the 3.3V LDO, and having it not drop out). This is how I noticed having my body very near to the circuit improves the performance somewhat. I've tried the wet finger trick on some specific parts of the circuit, but that doesn't seem to effect the TPS performance compared to a large body (my whole hand, my face).
I also noticed having a ground lead soldered from an output ground to another ground elsewhere on the PCB effects performance. Putting my hand near this ground loop gives a similar effect to putting my hand/face near the TPS chip.
Has anyone had this issue before? Any tips or ideas on how I could remedy this issue without spinning a new PCB? I can post screenshots of traces if anyone feels that would be helpful for diagnosis, or any other information that would be helpful.
Thanks so much for the help!
Cody
