Working with microchip MCP73831T chip to charge a single LiPo cell. We are producing by the 1000 unit and getting significant failure of the charger IC. I suspect we need ESD protection on the 5V for hotswap events, or maybe backfeed somehow from the L1 inductor hurting the IC? Does anyone have any clue what could be wrong with my circuit below?

failing circuit

Not sure what that guy meant by non legible.

Product is failing on customers, after 1-3 months of repeated usage. Maybe 5-10% unit failure rate.

USB plug/unplug to charge an 18650 LiIon cell, 2000mAH, at 500mAH per the charger spec sheet. Symptom would be charger IC no longer lights up status LED and does not charge battery or output voltage. Replace IC and works OK again.

If unable to figure out, maybe just dump microchip brand charger for a comparable texas instruments...

enter image description here

  • 4
    \$\begingroup\$ You have a product you make money on (or used to, or want to), that's failing, but you can't be arsed for the five extra minutes it takes to check if your schematic is actually legible? Very bad start. \$\endgroup\$ – Asmyldof Oct 13 '16 at 22:05
  • 1
    \$\begingroup\$ which hotswap? What is swapping with what? \$\endgroup\$ – Ale..chenski Oct 14 '16 at 4:44
  • 1
    \$\begingroup\$ When do you see the failures? At customers, or from production line? \$\endgroup\$ – Ale..chenski Oct 14 '16 at 4:54
  • 1
    \$\begingroup\$ On the schematic it says "connect to 18650 cell" but you say it charges a Lipo cell. So which is it, 18650 or Lipo? What are the specifications of your battery (capacity, C rating, etc.)? Are you charging at 500mA like the schematic says, or some other current? When the MCP73831T fails, what are the symptoms? Can you show us the PCB layout? \$\endgroup\$ – Bruce Abbott Oct 14 '16 at 5:40
  • \$\begingroup\$ updates above show pcb layout \$\endgroup\$ – Zachary Scott Dickerson Oct 18 '16 at 2:09

For problems like this you need to find the failure mechanism. If you don't switching to another controller may not solve the problem.

Make sure your following all of the specs in the datasheet and you aren't exceeding any specs, especially absolute max and min specs. Get an Oscilloscope on the input and output of the controller and set triggers for high and low voltages. Make sure the controller isn't getting into a thermal condition that would cause it to fail. Then run through what a regular customer would do with the product.

If you think the failure mechanism is ESD, then devise a test procedure to test that scenario. Either find a way to generate ESD, or buy a calibrated gun (you should have tested for ESD anyway when your product went to regulatory). Zap the product at different points and see if it fails and how it fails. If it reliably fails when you zap it, then you can solder on an ESD diode and see if it still fails.

If its not ESD, it could actually be a problem with the IC itself. If you've exhausted all avenues and you don't see any ways you are exceeding the specs, and you've actually measured values like voltage and current over a reasonable length of time and conditions. Then reach out to the manufacturer and say hey, this isn't working.


At 500mA a 2000mAh battery will take over 4 hours to fully charge from flat. You have not used the recommended pcb layout with large copper area and many vias, so the IC is probably spending much of that time in over-temperature mode - resulting in even longer charge time. Hopefully your customers are not too inconvenienced by the long charge time, but the continuously elevated temperature may be affecting reliability.

Symptom would be charger IC no longer lights up status LED and does not charge battery or output voltage

Are these failures first being observed at the moment the device is plugged in, or after it has been charging for a while? If the former then it may be an ESD or voltage surge issue, if the latter then it is probably temperature related. In any case I would consider changing to a switch-mode IC that is more efficient and can charge at a higher rate continuously.

  • \$\begingroup\$ I think switch mode chargers will be too expensive for this product, I'm looking at switching to BQ24040 linear charger, higher voltage input. \$\endgroup\$ – Zachary Scott Dickerson Nov 7 '16 at 20:50
  • \$\begingroup\$ "I think switch mode chargers will be too expensive for this product" ...the kind of thinking that results in poor performance, low reliability and a failed product. I have a box full of lipo chargers that I dare not sell to customers because the (normally reputable) manufacturer decided to use a crappy linear charge IC - that overheats and blows up with regularity. \$\endgroup\$ – Bruce Abbott Nov 7 '16 at 22:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.