TPS53681: PWM specifications

Hi

It is US holiday. please expect a answer next week.

Regards

Yihe

Hi,
We are designing Multiphase DC-DC Converters like TPS53681EVM-002.
Thank you for your explanation.
> So min TOFF is setting dependent, and there is no TOFF_max.
Does it mean min TOFF is programmed using resistors or fixed?
I would like to know the minimum time of TOFF_A and TOFF_B below.


Regards,
Nagai

Hi

Please see the image below. 

TBLANK sets minimum time gap between two consecutive CLK_AON pulses. Note that this is TOFF_A + positive pulse width of CLK_AON. The positive pulse width has not any role in TOFF_B. The rising edge of the clock defines PWM firing instant. As an example. The two rising edges near TOFF_A are related to PWM4 and PWM5. (see image below)

TOFF_B: 

In this specific example, mentioning a value of TOFF_B is difficult. This is because, during transients switching frequency changes. However the minimum value is maximum of following two. 

1. MINOFF --> a register setting

2. TBLANK*number of phases - TON of that phase. 

Regards

Vasav 

Hi
Please let me confirm about TOFF_B.
MINOFF is applied to each APWM, however another condition (TBLANK*number of phases - TON of that phase.) shoud be considerd, rigrht?

And I would like to know startup state too.
I checked VBOOT_SR on manual, but I couldn't understand how TON and TOFF pulsewidth is set.

Regards,
Nagai

Hi Nagai,

Understanding on TOFF_B is correct. 

To reply second part (TON/ TOFF) during startup, I need to check device architecture document. I will get back to you ASAP (targeting Thursday). 

Regards

Vasav 

Hi,

Thank you for your answer.

Please let me ask one more question.

The waveforms of the start-up operation are shown below.

The frequency is much higher than 500kHz immediately after power-on.

Is this correct? 

Regards,

Nagai

It seems that some load is connected. Please confirm.

Please remove the load and take the waveform  again. 

Hi,
I tested it again without load but the waveform is same.
Is there a possibility that it operates as a transient operation during the start-up?

Regards,
Nagai

I will check it in our lab on Tuesday, and update to you afterward. 

By the way, I would expect such problem if Cout is quite high. Is it possible to share value of Cout. 

Also, if possible I would like to ensure in layout that it is forming proper LC filter (not the L-C-L-C)

The test condition is that the board is TPS53681EVM-002, but inductor is replaced to smaller value, such as 50nH.

Thanks for the information. 

I will get back to you on Tuesday after carrying tests in out lab.

Hi Nagai,

Attaching the waveforms I obtained from the original EVM. 

Here the pulses are not as close as we seen in images provided by you. 

Regards

Vasav 

Hi,
Thank you for your testing.
Is it okay to use under high frequency conditons such as the waveforms which I sent?
Due to the high current, the filter may be placed in the LC-L-C arrangement.

Regards,
Nagai

Hi,
I understand.
Thank you for your kind answers!

Regards,
Nagai