Electronic – Undervoltage protection strategy for Li-ion batteries

batteriesbattery-chargingbmslithium ionsafety

We are using a Li-ion battery with built-in safety PCM (Enix MGL9070), a gas gauge to monitor the battery voltage (STC3115A from ST) and a battery charger (bq24133 from TI).

To protect the battery, the current strategy is:

  • when the gas gauge detects that the battery voltage is below 3.2 V, a warning is shown to the user (low battery) when the board is powered up. The system will go to sleep and the power consumption will then be quite low (100 to 500 uA).

  • the board then relies on the battery's built-in safety PCM to cut-off the power if the voltage goes below 2.75 V.

We initially had also a voltage supervisor at 3.2 V which could pull down the enable pin of the main regulator, so the board power consumption could drop down to almost nothing. This voltage supervisor gave us troubles (bad filtering, high voltage drop when powering other peripherals which triggers it, etc.)

The problem that we would like to solve is this: A user empties the battery down to the low battery warning and then leaves the system unattended. A year or two later, the user tries to charge the system again to use it.

Questions:

  • is it safe to rely on the built-in safety PCM of a Li-ion battery?
  • does it damage the battery to actually reach the built-in safety PCM and can it be then safely charged again?
  • would it be good design to do everything possible (using a voltage supervisor, for example) not to reach the built-in safety PCM of the battery?
  • does a better strategy exist?

EDIT 1:
It is a general consumer product.

Best Answer

The only proper answer to this is really how much confidence do you need to have in your solution. If this is going to be a consumer product like you say, you will probably want to go after UL listing. I tend to always over-spec li-ion circuits, since failure with them is such a big failure.

For UL testing, the product will need to pass "single-fault" testing. This basically means that if any one component/solution fails, the battery shall not present a hazard. So, they would bypass the onboard battery PCM, and then overload/over-discharge/etc the system, and if you have no other protection, the battery would fail.

The "over-kill" approach that I have used is to have the onboard PCM, a PTC fuse, a battery manager (AP9101), and sampling the voltage of the battery with the MCU/gas-gauge like you suggest.

Note: the AP9101 is just an example, currently using it in a few designs

  • The PCM is treated as the last-ditch safety, and the goal is to never need it.
  • The AP9101 is spec'd to activate before the PCM on under/overvoltage. Potentially also on current.
  • Since the AP9101 typically will have a wide range of current when it will activate, I rely on that only for really fast events, and near dead-shorts. For slight over-current events that are out of spec, the PTC fuse is relied on.
  • The MCU sampling the voltage range of the battery allows you to stop charging at a lower voltage than the AP9101 is spec'd at, allowing for longer life of the battery. (This really isn't always needed, but it's a free addition)

Using these three solutions allow for any one to be bypassed or fail, and will still protect the battery. Of course, this adds some extra cost which may not be acceptable, but if properly implemented, it will pass UL.

...and can it be then safely charged again?

Depends on the charger IC. Some won't allow a deeply discharged battery to be charged again, others have a slow ramp up that allows it.

Related Topic