TPS563208: Switching frequency vs load current at FCCM device

Hi,

The Fsw will slightly change with Vin, Vout and load as it's DCAP control. pls provide circuit schematic for review.

Hi Andy,

Thank you. Is there other DCDC converter IC from TI, which satisfies below?

1. P2P with TPS563208/1

2. OC limit_min > 2.9A (means 2A ver TPS562208/1 can't be used here)

3. Constant SW frequency in light load. (FPWM, FCCM mode)

Regards,

Hayashi

Hi,
Thank you.
I apologize for the continuous questions. Under the same conditions of Vin/Iout/Vout, how much variation can we expect in the switching frequency due to factors such as individual components or temperature? Currently, we measured around 530 kHz for a specific lot, but if there is a possibility of it dropping below 450 kHz, it could lead to issues in the system, so please let me know.

Regards,

Hi Andy,

Please give your insight for Naohiro Mizumata (customer name) questions?

Regards,

Hayashi

Hi Hayashi and Mizumata,

This is a question related with many factors, let me try to answer it one by one.

1. The relation between output current and frequency drop at light load in FCCM mode

With D-CAP control, the buck converter has the control on the on-time and the switching frequency is determined by the on-time control.

But at light load, as the current flowing through the Power MOS is smaller, the time for charging/discharging the junction capacitor of Power MOS is longer. That means it would take longer time to drive the Power MOS, which causes larger error between the "controlled on-time" and "real on-time".

Specifically, one major effect is that it would take longer time to turn off the high side Power MOS at light load. You could refer to Athos's answer in below thread and there's detailed zoomed in SW waveform of a FCCM operated D-CAP buck at light load. It's another part but work with same principle. You could see that at light load, the high side Power MOS turn off in a slow slew rate. During the high side Power MOS turn off, some power would still be flowed from Vin to Vout, which causes "real on-time" longer than the "controlled on-time".

One method to reduce the effect is to use smaller inductance. As when high side Power MOS turn off, that's when the inductor current reaches the peak in a cycle. With smaller inductance, the peak inductor current would be larger with same load current, which could accelerate the procedure to turn off high side Power MOS and make "real on-time" closer to "controlled on-time". 

https://e2e.ti.com/support/power-management-group/power-management---internal/f/power-management---internal-forum/1130653/tps568215-fsw-tolerance-under-light-load-condition/4196136?tisearch=e2e-sitesearch&keymatch=Athos%2525252520Zhao%2525252520Frequency#4196136

2. Factors such as individual components or temperature under the same conditions of Vin/Iout/Vout

For the variation with temperature, I checked the char data of IC during development and it has small impacts. The median Fsw data with 31 samples tested from -40degC to 125degC (other conditions same) varies in the range between 578kHz 583kHz.

For the passive components, just as the principle that I introduced above, the inductor variation matters at light load. But it doesn't mean 20% inductance variation would cause 20% on Fsw at light load, since inductance only affects the Ton difference between "real on-time" and "controlled on-time". To be honest, I don't have the data of impacts on Fsw with inductance variation, as that's more about application determined by customer's BOM. But I think that could be easily checked on bench using another inductor. For example, if using 1.2uH inductor considering the worst case with +20% variation, which is 1.44uH and could use 1.5uH to check worst case.

3. For Hayashi's questions on some recommendation of other parts. TPS563208 is already a FCCM device, but as I introduced the D-CAP control would have some Fsw variation at light load even with FCCM control. I think what you meant is the true fixed frequency controlled devices with internal clock or other solutions avoiding Fsw dropping below 400kHz, some candidates coming out of my minds include:

A. TPS56339

It is a fixed frequency control part with internal oscillator for clock. But the typical Fsw for oscillator is 500kHz, so the minimum Fsw limitation is just 420kHz. It's below the mentioned 450kHz but higher than 400kHz. Low side OCL min is 2.7A. Same pinout in SOT23-6 package. But Vref is different and its Vref is 0.8V. So need to change FB divider on BOM.

B. TPS563249

It's still a D-CAP3 part with FCCM control. But its typical Fsw is 1.4MHz, which makes it have very large margin above 450kHz. OCL min is 3.1A. Same pinout in SOT23-6 package. Vref is 0.6V different with TPS563208, so the FB divider needs to be changed if using this part P2P replacing TPS563208.

Thanks,

Andrew

Hi Andy,

Thank you for your detailed response.

I have one question regarding the switching frequency. When in mass production, I would like to know the extent of the variation in switching frequency, including manufacturing lot and individual unit variations, if possible. For example, the specification states 580 kHz, but I assume not all units will be exactly 580 kHz. What is the expected range of variation?
**As a supplement, I would like to clarify that my question pertains to the variation of the IC itself, excluding any variations in the surrounding circuitry.

Hi Mizumata,

To be honest, this is not a guaranteed data in our datasheet. But it's true that we got some data in development for you as reference.

In our development, we got the char data of devices through automation test and then did calculation for 6 sigma variation in statistics based on it. For 580kHz Fsw typical, the 6 sigma variation based on automation test data (already including temperature variation from -40degC to 125degC) is 540kHz to 636kHz.

That could be given to you as a reference.

Thanks,

Andrew 

Hi Andy,

thank you very,much!!
Noed,I will consider the information provided as a reference.

Andy-san, Hayashi-san,
Thank you for your cooperation.

Regards,