Hello,
I am currently evaluating the bq24075 and plan to integrate it into a product soon. I am prototyping with a Seeeduino XIAO. I was mislead to believe this device had some sort of advanced power routing features onboard, as it has a separate VIN pad on the bottom of the board which is marketed to be "perfect for powering the board with a LiPo battery." I assumed this implied that it was somehow isolated from the 5V pad. The 5V pad is either input or output. When powered from the USB port, the 5V pad is just a pass-through. Alternatively, you can supply the 5V pad from an external supply, which won't back-feed through the USB port thanks to the inline diode. What I didn't know is that the VIN pad on the bottom of the board is ALSO tied directly to the 5V pad. So essentially there's nothing special about this board whatsoever in terms of power management. It's no different from any other Arduino-inspired board with a 3.3V regulator. Schematic is here. files.seeedstudio.com/.../Seeeduino-XIAO-v1.0-SCH-191112.pdf
Sorry for the backstory. Onto the bq24075. Because of my incorrect assumption, I've essentially been prototyping for three weeks with the IN and OUT pins SHORTED TOGETHER. I was feeding the bq24075's IN from the 5V pad on my XIAO, so I can charge my LiPo when a USB cable is connected. Then I had the OUT of the bq24075 connected to the VIN pad on my XIAO. It seemed like a solid plan and it's been working perfectly for weeks now. But today I finally discovered on accident that the 5V and VIN pads are actually directly connected. I've been studying the bq24075's datasheet to try and figure out why I haven't seen any white smoke. Shorting the IN and OUT together seems incredibly unorthodox after all. However, after looking over the datasheet for a while I've concluded that I think it's actually perfectly okay to do this thanks to the bq24075's Power Path Management. But I really wanted confirmation from TI before I integrate it into my product in this configuration. Is there an adverse effect I'm unaware of - a reason I shouldn't be using the bq24075 in this way?
From what I can tell, Q1 will turn on and effectively short VIN to OUT when VIN is driven by 5V anyway since it would be greater than VIN(DT) + VBAT. But WITHOUT an external supply connected, shorting the IN and OUT pins together externally as I currently have it configured, VIN would be equal to VBAT (since Q2 will be on, to power the circuit from the LiPo), and therefore NOT greater than VIN + VIN(DT), which means that Q1 will remain off anyway. So, even though the bq24075 is back-feeding itself from OUT to IN when running on the LiPo, the voltage will remain too low to activate Q1.
This makes sense to me, but it's so unorthodox that I just wanted to confirm with TI that this is an acceptable circuit with no adverse effects. I sincerely hope to hear from a TI engineer about this soon. I apologize for the lengthy post for what is probably a trivial, elementary question. I'm a self-taught engineer after all. I could avoid all of this simply by powering my circuit with an external 3.3v regulator, driven by the bq24075's OUT, rather than the XIAO's onboard regulator which is tied to the USB 5V rail creating this weird loop. But if I can operate it this way I would much prefer to do so to keep my BOM and PCB footprint smaller.
Thank you very much,
Don Holden