TPS54KB20: Booted from flyback converter, but unable to be booted from DC supply unit

Hi Hayashi-san,

Could you please provide a more detailed description of the issue? It would be helpful to capture the measurements on the VIN, VOUT, SW, and EN pins for both cases.

If you have the schematic and BOM available (for example, the DCR of the inductor), please share them. Alternatively, you could use the Excel design calculator to check for any design warnings.

https://dr-download.ti.com/design-tools-simulation/calculation-tool/MD-1ikU0qjlfz/1.0/TPS54Kx2x_Excel_Calculator_Tool_and_Checklist_v1p0p0.xlsm

Thank you,

Tomoya

Hi Ide-san,

Customer are going to measure waveform. I'll get back to you after receiving schematic and BOM and waveforms.

REgards,

Hayashi

Hi Hayashi-san,

Understood. I will wait for your response.

Thank you,
Tomoya

Hi Ide-san,

Sorry for late reply. Customer thought it is not a big problem because it works with ACDC flyback converter. However they recently noticed the device doesn't work correctly when with ACDC. So investigation is re-start.

Please see attached VIN, VOUT, SW, and EN pins waveform:DCDC立上がり波形 (1).pdf

And also there is new bahavior, even they uses same ACDC supply, the device works differently chip-by-chip. DCDC立上がり波形2.pdf

Schematics: uLED-Power3 (1).pdf

Inductor datasheet: inductor_commercial_power_vlb10050_ja (1).pdf

Can you please review it to identify the issue?

If there are anything you need to know, please let me know.

Regards,

Hayashi

Hi Hayashi-san,

Tomoya will respond in few days.

Thanks,

Calan

Hi Hideki-san,

My first suggestion is that the part may be latching off due to UVP during startup. Could you increase the soft start time and see if that resolves the issue?

My second thought is related to VCC. Please check to ensure that VCC remains stable at approximately 3V.

Thank you,
Tomoya

Hi Ide-san,

Customer tried to extend SS time (500us->1ms) and the issue doesn't seem. However, they doubt why the change can resolve the issue. Please explain more about the relashion between UVLO timing and SS time in detail?

Please see attached document to see waveform. VCC is stable at 3V.

DCDC立上がり波形3.pdf

Best regards,

Hayashi

Hi Hideki-san,

Tomoya is out today. He will respond next week.

Thanks,

Calan

Hi Hideki-san,

The UV protection should activate as soon as the SS pin reaches the VSS(DONE) threshold of 1.2V (typ), and there is a 70us (typ) UV filter delay.

The waveform appears to have a slight incline on VIN, as shown below. If VIN is close to VOUT, the output voltage will go into dropout due to reaching the maximum duty cycle and will trigger UVP if VOUT stays below the UV threshold for 70us. This may not happen with a slower soft start time because VIN has risen just enough for VOUT to exceed the UV threshold. 

To confirm this, you can capture the SS pin on the same waveform and check what the VOUT is when the SS pin crosses the 1.2V threshold. UVP threshold is as shown below:

Thank you,
Tomoya

Hi Tomoya-san,

Customer updated as below. Left waveforms shows when the device start-up correctly, right waveforms doesn't. Both are measured with SS=550us (Css=22nF) for verification of the root cause.

DCDC立上がり波形4.pdf

When Vss reaches 1.2 V, VIN is between 4.1 and 4.2 V. VOUT is then around 3.7 V.

In terms of calculation, it can be considered that the switching is almost ton(max).

On the other hand, considering UVP, UVP=3.7x0.8=2.96V, which means that protection is applied when the voltage falls below about 3V, but the customer thinks that the acquired VOUT is at least 3V or more and that protection is not applied.

Therefore, customer expect that there is another reason for the protection to be applied, what do you think?

Best regards,

Hayashi