bq24295: Current limiting with standalone charging

Hey Brian,

Just a couple of thoughts here:

You are correct that ILIM controls the input current limit threshold, not the charging current. And yes the charger does have a default charge current of 1A.

1) I did notice that you left D+ pin and D- pin floating on your schematic. This could yield a random input current limit detection, as dictated by what the charger identifies during Input Source Detection. If you short the pins together, does the issue go away? Note that the charger selects the minimum of either the ILIM pin setting or the IINDPM register setting (set after detection).

2) Even though your charger can provide up to 3A, it is possible that a lot of line impedance can push the voltage as seen by the charger on VBUS, to the VINDPM threshold. The default/standalone value for VINDPM is 4.76V, so it would only take 360mV of drop on the VBUS line to cause this. If the input current is ~360mA, we're talking about 1 Ohm of resistance. Did you measure VIN at your adapter or directly on the VBUS pin of the charger?

3) Line impedance from the battery side can also be a factor. If the line impedance from the battery cell to the BAT pin of the charger is significant enough, you may find that the charger falls into the (Constant Voltage) CV mode of operation earlier than expected. I will point to Figure 16. Battery Charging Profile of the datasheet for this. At very high line resistance, the charger will believe it is at the battery regulation voltage much sooner when it provides the 1A of current in (Constant Current) CC mode.

4) Even though you are using the charger in SA mode, I would recommend using I2C communication in host mode to read back the registers on the charger. This would help us investigate what the root cause is.



Regards,
Joel H

Hi Joel,
Thanks for the quick and helpful response. Following up:

1. I will make sure D+ and D- are not floating in our next design iteration but for the time being I think it is not causing the issue - I was able to short the two pins to each other and the behavior did not change.

2. I checked the voltage at the VBUS pin and it was about 5.1 V so I don't think the VINDPM threshold is in play.

3. There does not appear to be any meaningful voltage drop from the BAT pins to the battery terminal

4. This is a great thought - unfortunately we did not break the I2C pins out but I will see if I can get some wires soldered on to the bit of the pins that are exposed on the PCB.

In the meantime, we noticed that the STAT line is exhibiting some strange-seeming behavior (it is driving an LED using I believe the same circuit as in the data sheet application circuit and eval board circuit ). When connected to the charge source and the battery and seeming to charge (power draw from the charge source is in the 50-450mA range as noted), the STAT line is outputting a square wave with frequency of 20 Hz and amplitude of 2.64 V. The duty cycle is 33% (17 ms at 2.64 V and 34 ms at 0V).

Perhaps the chip is seeing no charge fault conditions before initiating charge, then quickly seeing a fault and disabling charge, then when the fault disappears it is initiating charge again, etc etc?

I am digging through the data sheet again but haven't seen anything that seems to explain this. Any suggestions from your end are welcome. Thanks!

Brian

Thanks for your feedback Brian,

One other thing I noticed in your schematic was the lack of a 10uF capacitor on the BAT pin. This is a critical capacitor to help with charging regulation. Please add this to your design close to the IC. 

As far as your comment about the STAT pin, we typically see this when there is no battery connected. It is indirectly stated in the datasheet, but it is not a quantifiable time. When the battery is not making good connection, the charger very quickly enters a charging cycle and terminates. So the STAT pin flips between LOW and HIGH states defined in Table 6 of the datasheet. Adding that capacitor may also be affecting the typical charge cycle.

One thing I would ask for is a waveform of the BAT, SYS pins and charging current into the battery. if possible. 

Regards,

Joel H

Absolutely Brian,

I am glad you were able to find the root cause.

Let us know if you have any more questions regarding the charger.


Regards,
Joel H

Hi Joel,

I had a quick follow-up question about our use of the bq24295. Does this part reset to all default configuration settings whenever it loses power?

For context, we are interested in using it as a standalone charge controller (since it is cheaper than the bq25606 or other PNs that are designed for standalone use). If we change the charge current from the 1A default to e.g. 2A via I2C from an off-board MCU, will that setting persist even if the battery the bq24295 is connected to dies?

I hope the question makes sense and thanks very much for your help,

Brian