TPS2116: Reverse leakage current after some time

Hi ,

What is the exact test case in which you see leakage from vin2 to vin1

Regards 

Kunal Goel

Hi Kunal,

The output of the mux is disconnected (trace is cut). The buck regulator that goes to VIN1 is not powered, his power good output (which is connected to PR1) is verified low and as such VIN2 should provide power to the output which it does, but I also measure some voltage (couple of volts, depends on board) on the VIN1 port which should not be the case. With that I get a significant battery leakage current. On a fresh board, it works just fine as it should, but on these "faulty" ones I measure in the hundreds of uA which is not acceptable. 

On another post about this IC, there is mention to decrease the input impedance, so the input isn't "floating", I tried adding a 1k resistor on the buck output, but I still get some leakage. I'm also worried that the buck might try charging the non-rechargeable lithium battery and cause significant damage.  

Hi Krunoslav,

We have some reverse leakage current as per datasheet. 

Regards

Kunal Goel

I'm well aware that some leakage current is expected but I measure >100uA not 0.1uA. 

Thanks. Please share schematic for review. 

Regards

Kunal Goel

Please see the attached schematic. R80 and Q9 are removed so they 100% don't cause any issues. 

Hi,

Please let me check and get back. 

Regards 

Kunal Goel

Hi Krunoslav,

Were you able to figure out the issue?

I'm working on a low power design and planned on using the same IC in a similar configuration (VIN1 is USB + LDO powered and VIN2 is battery powered), so I'm curious to know if there indeed is an issue with the MUX. 

Also, when you mentioned isolating the circuits, are R80 and Q9 still present, or did you remove them as well? Just wondering if they, and any associated measurement / control circuitry, could be loading the battery in some situations. You mention that you don't have this leakage issue initially, and so perhaps this point is moot, but just putting it out there. 

All the best!

No, the issue is still not resolved. Current plan, if TI doesn't have any ideas, is to add a series diode on VIN1 and increase the buck voltage to compensate. This should prevent leakage from VIN2 to VIN1 should it happen again. I'm not too worried about VIN1 leaking into VIN2 as there isn't really any significant leakage when both ports are powered. This issue only happens, at least in my case, when VIN1 isn't powered. I did try to add a 1k resistor to GND on VIN1 to act as a strong pull-down, but my guess is that the IC's get somehow get damaged internally. 

Yes, R80 and Q9 are removed and the leakage still exists. Even with them fitted they don't cause the leakage, but for the sake of faster troubleshooting I removed them. What you see on the schematic above, is all there is for the battery circuit.