BQ21061: Battery detection mode

Hi Francesco,

This behavior does not represent a battery detection procedure. To better understand this behavior, can you please share the following information?

- Is there other circuitry attached to the BAT pin aside from the protection IC and battery?

- Once the protection IC has closed, what voltage is the battery output?

- What voltage does channel 2 represent?

- What does this look like when just non-open battery is connected and also what does your BAT pin look like when it is just left floating?

When VIN is applied and BAT is floating the charger will attempt to charge the Battery capacitor up to regulation voltage and terminate charging. Soon after the battery capacitor will the discharge below the recharge threshold and charge back up. While the bat pin is floating, VIN is present, and charge is enabled this cycle will repeat over and over again.

This is the behavior expected when the BAT pin is connected to an open battery protection IC. However, most IC's will close after the voltage at their input rises above a certain threshold. So this battery capacitor charging behavior typically allows the protection IC to reconnect to the battery and allow charging. I believe this might be one possible reason for the behavior you are seeing, though usually the protection IC would connect the first time that the voltage crosses the necessary threshold.

For better analysis, it might be helpful to look at your schematic to confirm external components are as needed.

Best Regards,

Juan Ospina

Dear Juan,

thanks for your answer. In the following the answers to your questions:

- Yes, attached to the net connected to BAT pin of the BQ21061 there is a Fuel Gauge

- After IC battery protection has been closed the battery voltage has a value that depend on the residual capacity of the battery. Our firmware inside our MCU opens the battery protection when battery voltage goes below 3.2V (battery almost depleted), in this case, when we power BQ21061 from IN pin, after the successive pulses, the voltage at BAT is around 3.2V. Instead, our MCU can also opens the battery protection on request coming from user and therefore, if battery protection, for example, disconnects the battery when the battery voltage was 3.6V, when we reconnect power at IN pin, after successive pulses, the BAT voltage will be around 3.6V.

- Channel 2 represents the voltage with we power the MCU and other components coming from the LDO output of the BQ21061. The LDO starts with 1.8V as default value on its output but our MCU changes this voltage to 3V because we need higher voltage respect to 1.8V to power a vibration motor with nominal 3V (in a  new release of our firmware we modified the timing in which our MCU changes the LDO output from 1.8V to 3V, in this new release, we will wait about 2sec. after that MCU communicates with BQ21061 on I2C to change LDO output, in this manner, probably when the MCU starts I2C communication, the pulses on the BAT net have been finished and battery was connected)

- about the question regarding this point we didn't perform this type of tests

Our problem is not related to the BQ21061 behaviour since the battery charger performs as we expected, no I2C problem issue, no problem during battery charging, etc, BQ21061 works very well in every conditions but we have some problems with I2C communication with a Fuel Gauge that sometimes does not respond on I2C and this happen sometimes at the power-up after the battery protection has been open the circuit since battery voltage goes below 3.2V. 

We need to understand better the process that BQ21061 performs at power-up when attemps to recharge a protected (open) battery in such a manner that we can start all MCU activities (communication on I2C) after the battery has beSchematic_BQ21061.pdfen connected.

Hi Francesco,

I understand. As mentioned before, with a floating bat pin (or opened protection circuit) the charger will charge the bat capacitor to VREG and cycle the charge/discharge pattern on the VBAT capacitor. I would recommend determining the requirements for your protection IC to close to reallow charging.

Your schematic surrounding the BQ21061 looks good so nothing immediately jumps out as a possible cause for the delay in the protection IC. Based on your protection FET datasheet, it should reconnect the battery to the charger once the voltage on VBAT reaches above 2.8V, which it does on the first pulse. The expected behavior is that it should remain connected at this point.

Can you provide a waveform of the control signal that you are using to disconnect the battery via the MCU next to a waveform of the oscillation you are seeing?

Best Regards,

Juan Ospina

Hi Juan,

our circuit is very miniaturized and therefore it is not so simple to acquire signal coming out from the MCU that drives the protection to open it, however we send on this pin a 5msec. pulse at high logic level (3V). Please not that a high logic level outputs from microcontroller at this pin should immediately open the protection and consequently turn-off all circuits included the MCU (we verified that this happens!). I performed some other tests and if we do not drive the protection but we leave that protection itself disconnet battery with its own threshold (2.8) at the successive power-up no pulses appears on the BAT voltage. Please can you explain me why you asked me to check the control signal that we use to disconnect the battery togheter the pulses at power-up on the BAT net?

From your consideration I seem to understand that pulses on BAT pin at power-up it is not due to the BQ21061 that try to charge battery capacity (you wrote that BQ21061 try to charge capacitor at battery regulation voltage that is 4.2V but the high level of pulses (yellow trace on oscilloscope image already sent) on BAT net at power-up is not 4.2V but much lower, approximately 3.6V) but you guess that these pulses have to do with the protection control signal, is it true?

Thanks

Francesco

Hi Francesco,

Francesco Mariottini said:

Please can you explain me why you asked me to check the control signal that we use to disconnect the battery togheter the pulses at power-up on the BAT net?

I wanted to confirm that the oscillations were not potentially due to a disconnect signal being sent to the protection IC that may be causing it to close repeatedly.

Francesco Mariottini said:

you guess that these pulses have to do with the protection control signal, is it true?

That is my first guess. As you indicated before, a voltage above 2.8V should close the protection circuit. Since the yellow indicates voltage at the BAT output we would expect a constant voltage at 3.2V (BAT cell voltage) and the battery charging. However the pulses that jump up to 3.6V indicate to me that the voltage at this pin is being disconnected from the BAT cell since battery voltage does not change that easily. This leads me to look for possible causes that may cause the protection IC to close and open repeatedly.

Below is a typical waveform of VBAT when VIN is present, charging is enabled, and battery is disconnected:

It doesn't seem to match the pulses that you are seeing, and the high voltage doesn't seem to be equal to VREG. This is why I was curious about other possible circuitry that might be attached or pulling current from the BAT node.

Best Regards,

Juan Ospina

Hi Juan,

the only circuits that are connected to the BAT node are the Fuel Gauge (that is powered from the BAT node) and the protection with a circuit to disconnect it. Our firmware, at power up, does not drive the pin that is connected to the protection to open the discharge mosfet. The MCU drive the signal to open the protection in discharge mode only in two cases: 1. with double click on pushbutton 2. when, in discharge mode, the battery voltage measured from a Fuel Gauge is below a 3.2V (but at power-up the I2C communication with a Fuel Gauge is not yet active, we initialized the communication with Fuel Gauge when the LDO voltage will be 3V).

Anyway the initial pulses on BAT voltage seems to do not influence the behaviour of the BQ21061, also the interpretation of this pulses must be related only to a protection that attemps to connect a discharge mosfet (charge mosfet should be closed and charging should start immediately through this mosfet and the by-pass diode of the discharge mosfet). We performed a lot of test and BQ21061 always shows behaviour as expected. As mentioned in my previous message our problem is Fuel Gauge that sometimes at power up does not respond on I2C bus. This happen after that protection has been open (by the MCU or by itself) and some time has been elapsed (typically we leave the circuit with battery protection open for all the night and the successive morning we try to power the circuit and Fuel Gauge does not respond).

Hi Juan,

no thanks, about BQ21061 no other informations are needed.

Best Regards

Francesco