BQ25703A: About disturbance of charging current

Hi Yusuke,

Please send me the schematic and waveforms through a private message. For the waveforms, it would be helpful to see VBUS, SW1, SW2, and IBAT on the same graph.

Also, if you increase the charge current to 1.5 A, does it still oscillate?

Best regards,

Angelo

Hi Angelo,

Thank you for your support.
I sent the file using a private message.
Could you check the file?
And Could you tell me why the charging current is sine-curved?

Best Regards,
Yusuke

Hi Yusuke,

Thanks for sharing the schematic and waveforms. In addition to the oscillating charge current, it's also unusual that HIDRV2 and LODRV2 are switching. With a 19 V input voltage, the charger should be operating in buck mode with Q3 off and Q4 on.

It's hard to read some of the component values in the schematic due to the low image resolution, but please make the schematic match the EVM as closely as possible, especially for the following points. There are other schematic differences as well, but here are the changes that are most likely to cause stability issues:

• The COMP1 compensation network component values must match the EVM exactly.
• VSYS capacitance (the EVM has 6x10 uF ceramic caps plus 1x47 uF POSCAP)
• Remove the resistor from VSYS to /BATDRV
• Change the SRP/SRN and ACP/ACN component values so that they match the EVM

Please make sure the battery is connected with a short wire so that there isn't too much parasitic inductance. For reference, here is a similar E2E thread where the customer reported a sinusoidal charge current:

https://e2e.ti.com/support/power-management/f/power-management-forum/921199/bq25710evm-017-sinusoidal-batfet-gate-and-battery-current

Also, I suggest testing a BQ25703A EVM and comparing the behavior to your board. This will speed up the debug process and confirm whether the issue is related to your board or your test setup.

Best regards,

Angelo

Hi Angelo,

Thank you for your kind support.
>It's hard to read some of the component values in the schematic due to the low image resolution, but please make the schematic match the >EVM as closely as possible, especially for the following points.
I sent the default material submitted by the customer with a private message.

>Also, if you increase the charge current to 1.5 A, does it still oscillate?
Yes. It is not stable at 1.5A, but it tends to improve when the charging current is increased.

>The COMP1 compensation network component values must match the EVM exactly.
Yes. I recommended the same value as EVM.
Customer evaluates the comp pin constant under the same conditions as EVM.

>VSYS capacitance (the EVM has 6x10 uF ceramic caps plus 1x47 uF POSCAP)
Customer adjusted the output capacitor and the system load, but it didn't improve.

>Remove the resistor from VSYS to /BATDRV
Regarding the above, it is installed with the following E2E support. Is there any direct involvement in this issue?
https://e2e.ti.com/support/power-management/f/power-management-forum/956041/bq25703a-about-the-drive-capacity-of-batdrv-pin/3535624#3535624

>Change the SRP/SRN and ACP/ACN component values so that they match the EVM
I will check this with the customer.

>Please make sure the battery is connected with a short wire so that there isn't too much parasitic inductance. For reference, here is a similar >E2E thread where the customer reported a sinusoidal charge current:
Thank you for your information. I will check this with the customer.

Best Regards,
Yusuke

Hi Yusuke,

Thanks for the clearer image of the schematic and the background regarding the resistor from VSYS to /BATDRV. I don't think this resistor would cause the IBAT oscillation shown in your waveforms, but this resistor is not needed for stable operation either. The BQ25703A EVM does not include this resistor. For now, let's remove this resistor in order to match the EVM schematic more closely.

Can you confirm that the COMP1 pin components are identical to the EVM? Based on your last post, the COMP1 components match. However, according to the schematic, R22 and C33 are inside a red box, which means the component values may be different from the EVM. This is important because COMP1, COMP2, and the VSYS capacitance are critical to the converter's stability.

Also, can you please test a BQ25703A EVM and compare the behavior to your board? This will tell us whether the issue is related to your board or your test setup.

By the way, for now, let's do all the tests without any system load connected. Let's only have the adapter and the battery connected so that we can remove extra variables and better isolate the issue.

Best regards,

Angelo

Hi Angelo,

Thank you for your kind support.
I sent the Additional waveform using a private message.
Schematic is set to a value close to EVM.
Could you check the waveform?
Problem could not be solved.
However, the following two were confirmed.

・Disturbance of charging current changes according to charging battery voltage or battery capacity
・When the cable was shortened, the charging flow waveform became stable when the charging current was small.
    However, when the charging current was increased, it returned to the sine curve waveform.

Could you give me some advice on additional check points and causes to solve this problem?

Best regards,
Yusuke

Hi Yusuke,

Thanks for the detailed waveforms. Can you please send me the latest schematic, which shows all of the component values? I want to double check some of the critical component values to make sure that they match the EVM. Some component values, such as the COMP1 and COMP2 compensation networks, must match the EVM exactly in order to allow stable operation.

Also, since the battery's cable length affects the results, it's clear that the parasitic inductance in the test setup is hurting the stability as well. Please test a BQ25703A EVM with your test setup so we can compare the behavior to your board and identify any other areas for improvement.

Best regards,

Angelo

Hi Angelo,

Thank you for your support.
I sent the latest schematic using a private message.
Could you check the private message?

Best Regards,
Yusuke