Hello,

We have assigned this thread to the appropriate engineer who should be able to respond by Thursday.

Regards,

Alex

Perry Wang said:

Is there any other GUIs or drivers that i can try to set the PMIC mode default to FPWM not by TDA4(currently we are limitated by our software) ?

The only way to do this is by changing the NVM. This isn't recommended, as you are essentially creating a custom NVM.

Perry Wang said:

as far as i understand it should keep on PWM mode for heavy load when it set to automactic mode defaul

That is correct, while in AUTO mode the BUCKs should be in PWM mode for heavy loads and PFM mode for light loads. How heavy is the load when this happens? Your scope shot is not a typical PFM signal. My concern is that you're seeing a source of instability.

-Mike

Hello, Michael,

Thanks for your response, 

Yes, The test scope seems not a typical PFM mode waveform, It's more likely the power regulator always switching between from PWM and PFM mode;

The load current transient we measured will jump from 1A (light load) to 5.5A(peak value) around (heavy load),You can refer to the picture i attached,

you concerned about the source of instability is that you suspect the input of the PMIC is not stable?  

Thanks

Hello Perry,

With the holiday weekend coming up. Mike may not be able to respond to this until Tuesday. Apologies for any delays.

Regards,

Alex

Hello, Alex

Thanks, But it is urgent for us to clarify what is the root cause for the BUCK output instablity as it makes some capacitors sream!  And i find it is not easy to program NVM to set the power to FPWM mode, Please increase the prority internal,

Thanks for your graet support!

Hi Perry,

Please be assured we are giving this topic the proper urgency but we will need to wait for next week when Michael is back in office.

Best regards,

Layne J

Can you provide a scope shot like the very first one that includes load current, VCCA, and output voltage as well? Seeing all 4 signals on one plot would be very helpful.

Are all 3 phases using 0.47uH inductors?

How much point of load capacitance do you have have located near the processor?

Hello, Michael

Thanks for your reply,

1, Please check the scope i captured when the AVS module works up, From the scope it is obviously that when the current is low the PMIC set to PFM mode,When the current jumps high(the heavy load) Then the PMIC switch to PWM mode.

2,All the three phases using 0.47uH inductors - Yes!

3,How much point of load capacitance do you have have located near the processor? - aorund 20,Check below the capacitors near processor.

Thanks and waiting for your reply

Hi Perry,

The product expert will get back to you tomorrow, today is Memorial day in US. 

Thanks.

Regards,

Tomi Koskela

Hello, Tomi

Any feedback from TI product expert?

Thanks

Hello Perry,

For the '1212, the threshold for switching between PWM and PFM mode is around 500mA.

With that in mind, your latest scope shot does not look abnormal to me. The middle section is clearly a light load so the PMIC goes to PFM mode. To the right is a steady state PWM situation at about 1A DC load. The left side, the switch node is reacting to the load requirements of processor as expected.

Hello,Michael

Thanks for your reply, So it is indeed the PMIC (TPS6594-Q1) default setting mode is auto mode!

Is that available to set the PMIC to FPWM mode default by our custome SW through I2C configuration during each power up cycle as the NVM is not suggested to be changed?

Thanks!

Thanks,Michael

Just set the mode by the I2C configue right?

Thanks!

Hello, Michael

There are still some problems with the PMIC power mode, As we discussed before,when PMIC in AUTO mode the BUCKs should be in PWM mode for heavy loads and PFM mode for light loads,But when we try to set the current to 3A (even up to 5A) on the output of the BUCKs with the DC Elctrionic Load device, The Buck seems still switch between from PWM and PFM mode, It is very stange and abnormal from my understanding, From my understand when the load current reach to 5A then the PMCI should always working on PWM mode. No swapping power mode on heavy load.

Please advise!

Many Thanks! 

Hello Perry,

Due to the US holiday we will be a bit delayed in our e2e responses this week. We should be able to respond by Thursday. Thank you.

Regards,

Alex

Hello Perry,

  Please check "Buck Phase Adding and Shedding" specs below. Buck3 may be still working in PFM mode when the 3-phase has 4A load. If 3 bucks has unbalance issue with PCB layout, then one buck could be still working in PFM mode even at 5A load. 

Thanks!

Phil

Hello Perry,

  If you think your question has been answered, would you please click "This resolved my issue" button to close this thread?

  Greatly appreciated!

Thanks!

Phil

Hello, Phil

Thanks for your reply and i am sorry to feedback you late, Currently we are still not seting the power mode to FPWM mode successfully through PMIC GUI tool, We are trying to change the PMIC to FPWM mode to check when the AVS runs up(heavy load) then the PMIC always working on the PFWM mode,

Because we have occured some Capacitance scream on the AVS power rail ,3v3 input power rail,Vbat power line, We are suspected it is caused by the PMIC output mode is not on FPWM mode. By the way i have a  question from the datasheet: it is recommended to set the PMIC to FPWM mode, But why from NVM confuguration these BUCKn are always set to AUTO mode?

Thanks!

Hello, Phil

Sorry for late reply, As i describled above, We have scope the BUCK SWs node waveform and confirm the PMIC is not in PFM mode and we believe the capacitance screaming is caused by the PMIC BUCK, i will attached the waveform for your reference:

From your advise "MCU can easily change it to FPWM mode by I2C writing"  but it seems not easy from our side, We have tried to set the FPWM mode when power up the PMIC, But it seems not working, As i understand the PMIC would  upload the NVM value from startup everytime,

Even we try to change the register value successfully after power up the PMIC, It doesn't come into effect on current cycle, As Michael suggested as below:

Please advise if you have any comments to add , Thanks!

Hello Perry,

   I just tried again on 3 BUCKs with the EVM I have; they all have proper response for the mode change by I2C writing in current cycle.

   Even without load at all, I can tell by their output voltages: PWM mode has tighter output than PFM mode as showing by table below from datasheet:

  The waveform you captured looks like caused by the "Buck Phase Adding and Shedding" feature since the load change at that moment. 

   Please also change the bit below from 0 to 1 to forces the BUCK you test to operate always in multi-phase and forced PWM operation mode:

  Please close the thread if you think your question is answered. 

Thanks!

Phil