BQ24770: BQ24770 switching step down efficiency

Hi Pawel,

I tested this on an EVM with your operating conditions (CELL pulled high, VBUS = 24 V, ChargeCurrent() = 3 A, and ChargeVoltage() = 15 V). However, I'm unable to duplicate your issue. HIDRV and LODRV are both switching. The waveform below shows the PHASE node and LODRV.

You're testing this on a BQ24770 EVM, correct? Can you share your register settings and waveforms? Also, can you try other ChargeCurrent() settings to see if LODRV is always low?

Lastly, I would like to point out that a 24 V input is outside of the EVM's recommended operating conditions, as shown in the table below. It is not recommended to exceed 23 V.

Best regards,

Angelo

Angelo,

First of all, thank you for quick response. To be honest, we are not using EVM. We have our own design, which is actually copy/paste of reference one. With small difference. We are not monitoring input current, so we made direct connect instead of shunt resistor at input (in BQ24770 pdf, figure 21 it is "Rac"). So both ACN and ACP pins are tied together and connected to "AC" line, which is +24V in our case. Can it be reason of such behavior?

The circuit is functionally OK - output current and maximum voltage is regulated as set up via SMBUS. The problem is related only to low side MOSFET heating.

Dump of option registers is below:

ChargeOption0: 0x654E
ChargeOption1: 0x0A10
ChargeOption2: 0x00C0

Tried all combinations of 12V, 15V and 18V voltages, with 500mA, 1A, 2A and 3A currents. Effect occurs in all of them.

Pawel

Hi Pawel,

Thanks for the information. Your register settings look fine. Since you're not using the EVM, can you share your schematic so I can verify your circuit?

Please also capture a waveform showing PHASE, HIDRV, LODRV, and the inductor current on the same scope shot.

Pawel Cern1 said:

Tried all combinations of 12V, 15V and 18V voltages, with 500mA, 1A, 2A and 3A currents. Effect occurs in all of them.

Just to make sure I understand, LODRV is always low for these test conditions. However, HIDRV and LODRV both pulse occasionally under no load conditions, correct?

Best regards,

Angelo

Our management team is not willing to share schematics.

However, we managed to arrange TI EVM. It was working correctly until we shorted input shunt resistor, marked as "R1" on EVM schematic (to be honest, actually C14 capacitor, which is in parallel). After that, we have exactly the same behavior as in our own project - low  side MOSFET is not driven and warms up significantly when significant load is attached at output. As soon as short circuit is removed, EVM goes back to correct operation.

Basically now we have solution to our problem - we may add input shunt resistor in our project and connect ACP/ACN to its terminals.

Now the question is if we can solve the problem without input current monitoring, i. e. without adding equivalent of R1 in our design?