Hello, our company designed a board that hosts a Raspberry CM4 and we implemented a solar charger using BQ24650.
The design itself is nearly identical to the one proposed in the TI datasheet but for some unknown reason the board is not working.
Our engineer made some checks and discovered that the pins SRN, SRP and VFB are all connected to ground causing a shortcircuit when given power.
All the board have the V_BAT shorted to GND until we remove physically the BQ24650 from the board.
In fact we tested it and the SRP track blow up. I hope someone can provide some more insight on this problem.
Our current design for the charger:
PixelAgency
Thank you for contacting DigiKey , I am not an engineer but we do have some that monitor this forum , they may be able to give you some in site as to how to remedy the issue you are having .
Thanks Craig
A typical cause of pins that are normally high-impedance linear inputs becoming low-impedance shorts to ground is violation of abs. max voltage constraints. It wouldn’t be unreasonable to use a presumption of such an event to guide investigation.
Noting the location of the failure and any incident component damage or evidence of stress may offer clues as to the path of the fault current.
It’s unclear to me whether this is meant to indicate a repeated observation across multiple samples, or is a reference to observations of a single board.
In general, when a design that a person expects to function doesn’t, it’s because the physical circuit and/or it’s schematic representation are in disagreement with each other and/or the mental model of either/both in the designer’s mind. Take a careful look at the layout and assembly to look for possible disagreements; reliable failures across multiple boards are often a layout error, whereas unreliable failures across multiple boards can point to an assembly problem.
When that process doesn’t reveal the problem or offer enough clues, it can often be helpful to repair a board, attach as many channels of single-capture 'scope to key nodes in the circuit as one has available, set the trigger and time scales as best one can estimate to be appropriate, and blow it again under instrumentation. Rinse/repeat as needed to develop a picture of the sequence of events associated with failure.
1 Like
Our board in designed to have AC and Solar at the same time, switching from one to another as needed. We tried powering two boards with two different lines V_AC and V_PANEL (from a PSU set at ~11V) and got the same damage on both so we excluded a defective IC.
Removing the IC entierly (on a fresh board) removed the shortcircuit and it’s working correctly so we think it’s a problem in that part of the schematic (which we took from the one in the Texas Instrument datasheet):
If is not a clear problem in our schematic then we’ll need to search more elsewhere. I’ll forward the message to our electrical engineer as he’s studying this issue.
Thank you for now and I’ll post updates when I have some 
We conducted some further checks on the PCB, and can confirm that the footprint we used for the BQ24650 is too small. Even if it should be a standard footprint, the actual footprint differs slightly. As a result, the pad designated for thermal dissipation beneath the chip, once soldered, short-circuits multiple pins to GND.
1 Like
