BQ25792: Noise issue in PFM mode

Hi Roy,

When PFM is enabled, OutOfAudio bit prevents PFM pulses in the audio range when there is very light load (<<100mA) on SYS/BAT.  If the host disables OOA, there could be audible noise in that range.  I am not aware of switch over on VACx causing a reset but I will go check shortly.  The only recently found conditions to cause audible ringing are:

1. ISYS/ICHG~=200mA-300mA with PFM on and OOA on.  If the host disables OOA or load increases/decreases, this does not occur.  Still investigating root cause.  

2. fsw=1.5Mhz and VBUS~=VBAT just before entering/exiting buck-boost mode.  The converter min controllable pulse width is less than the min duty cycle so the converter starts skipping pulses for a brief period.  The frequency of this pulses can be in the audible range.  Changing to fsw=750kHz prevents this.

Disabling PFM prevents this noise but increase operating Iq which is noticeable at light load.

Regards,

Jeff

Hi Jeff,

If we want to reduce the noise in PFM mode, should be enable the bit or disable the bit? (0 or 1?)

Please help check it. Thank you.

Jeff F said:

I am not aware of switch over on VACx causing a reset but I will go check shortly. 

Regards,
Roy

Hi Jeff,

I see below content in datasheet. Does it mean that the PFM_FWD_DIS bit may reset when we switch the voltage from low to high?

For example, VAC1 = 12V, VAC2 = 5V. Now we used the VAC2 and then switch to VAC1. It will trigger the POR?

Regards,
Roy

Hi Jeff,

Even from low voltage VAC to high voltage VAC? My customer reported that the PFM_FWD_DIS did not change every time. May you help test more times? like 20times?

Regards,

Roy

Hi Roy, we will get back to you on this next week.

Hi Roy,

Is a battery present when the switchover occurs?  If not, does VBUS drop < VBUS_UVLO during the switchover?  If so, then all the registers return to default values.

Regards,

Jeff

Hi Jeff,

Yes, I suspect that there is VBUS drop when doing switch over. I've let customer monitor the VBUS when doing the VAC switch over. And dump the all reg. I believe all of register back to POR value.

And I would like to check the correct action for VAC switch over. 

For example, if BQ uses VAC2 and VAC1 is also presented. If we would like to switch to use VAC1. Just set 0x13h[7:6] from 01b to 10b. correct?

Regards,

Roy

Hi Jeff,

We found that the VBUS voltage would down to 2.5V when we did the VAC switchover. (From VAC2(5V) to VAC1(12V)) Is it normal? It looks like that it needs around 5ms to do switch over and the Vbus cap can't regulate the VBUS voltage. 

Does it mean that the VBUS cap is insufficient at this moment? Do you have any workaround?

Regards,
Roy

Hi Roy,

VBUS=2.5V is below VBUS_UVLO.  I don't understand what you mean by "Vbus cap can't regulate the VBUS voltage".  There must be a delay on switchover in order to prevent connecting VAC1 source to VAC2 source.

Regards,

Jeff

Hi Jeff,

I mean if customer did the VAC switchover, the phenomenon(VBUS drop) is expected? 

I mean for VBUS drop, should we add more cap on VBUS to regulate the voltage when doing the switchover?

Jeff F said:

I don't understand what you mean by "Vbus cap can't regulate the VBUS voltage"

I do believe this is not a normal phenomenon, otherwise we would recommend customer place enough cap on VBUS to prevent VBUS drop doing switchover.

May you help use EVM board and measure the VBUS node when doing switchover? (5V-VAC2 to 12V-VAC1) We tried to add more cap on VBUS and it looks like that the switchover time wasn't fixed and VBUS would down to VBUS_UVLO and then switch the VAC.

Regards,

Roy

Hi Roy,

VBUS drooping is not unexpected because the converter is still running when EN_ACDRx is set to 0 and before EN_ACDRy is set to 1.  If you don't want VBUS to collapse to UVLO, you can first set EN_HIZ=1 then EN_ACDRx=1 then EN_ACDRy=1 and then EN_HIZ=0.  That would prevent the converter from drawing current from the VBUS capacitors during the switchover.  So, in theory, with a big enough capacitance on VBUS, VBUS voltage would not drop below VBUS_UVLO.

I do not understand what the concern is here.  No other customer has asked about this.  What is the issue with VBUS dropping below VBUS_UVLO?  If a battery with voltage > VBAT_UVLO is present, the registers do not reset.

Regards,

Jeff

Jeff,

Thanks for support. I will let customer follow the normal way to do VAC switchover. In addition, for the issue, the system isn't under POR to change the bit. The customer has incorrect action to set back the PFM bit.

Regards,
Roy