BQ24250: Detecting Charger Connection & Fault Conditions

Hi Errol,

Regarding 1, for detecting whether a power source is attached, I recommend using REG#1[b1-b3]=0010 - input UVLO or 1010-Input Fault and LDO low. 

Regarding 2, yes, you should read those register multiple times, continuously, in order to identify the faults and clear them if applicable. For TS fault specifically, you can use REG#6[b0-b2] to determine the exact TS fault. 

Regards,

Jeff

Hi Errol,

Please see my answer embedded below:

• Is it correct that EITHER fault 0010b (input UVLO) OR 1010 (input fault and LDO low) indicate a transition to battery?
  • For 99% of cases, yes.  It is possible but highly unlikely that the LDO output gets shorted to ground, causing LDO low even with input voltage good.
• I have consistently been observing fault 0011b (sleep) when transitioning to battery. Can/should the sleep fault also be used for detecting the transition to battery?
  • Yes.  Sleep isn't really a fault.  That is what we call the state when the charger is powered only by battery.
• What is the recommended approach to detect a transition to charging?
  • The FAULT bits report Normal and the STAT bits report charge in progress.
• How many times should a fault condition generally be read before giving up? If this depends on the fault type, how many times should I read or how long should I wait for each type?
  • I suggest reading the fault register until you get the same fault (or Normal which isn't a fault) 2 times in a row.  That way you have the history of faults and the final fault or normal status of the device.  
• When a fault is indicated, does that mean necessarily that a device is charging, or on battery? Or does a fault give no information about whether a device is charging?
  • During Input OVP or Input UVLO, which will eventually end in Input Fault and LDO Low, the buck converter stops providing SYS output, the device is not charging and the device is powered by the battery.  With Battery temperature fault, the buck converter is providing SYS output but charging stops.  During Battery OVP, the buck converter stops so that the battery can discharge below OVP through the SYS load.  During thermal shutdown, the buck converter stops providing SYS, and therefore, charge current until the IC die temp drops below the shutdown temp threshold. The timer fault is the safety timer expiration and stops charge but not the buck converter.   The No battery connected fault means the battery detection algorithm is running (square wave on SYS and BAT) but, unless TS function is disabled, the Battery temperature fault will stop charge first and prevent battery detection from operating.  The ISET short fault is so unlikely unless there is a board solder or assembly problem (similar to LDO short) that I doubt you will ever see it.
• You mentioned ‘clearing’ faults if applicable. Can you provide guidance on how to do that (or just point me to where it is documented?
  • I suggest reading the fault register until you get the same fault (or Normal which isn't a fault) 2 times in a row. The register stays on the last FAULT status.  Clearing the faults simply means you read the fault and have the host software or user try to fix the fault.  For example, if you have Input OVP followed by Input Fault, the host software could tell the user to remove the input adapter that is too high.
• Is there example code you can provide (or point me to) for handling the bq24250 fault conditions? If so, that would be a good reference.
  • There is a linux driver at https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/power/supply/bq24257_charger.c?h=v5.9-rc6.  It has basic I2C read and write but does not have examples on how to handle faults.   That logic is an end application policy decision that TI cannot provide for legal reasons.

  Hope this helps!

  BTW, just curious why you chose BQ24250 over a newer device like BQ25611D, for example?

  Regards, Jeff