UCC28740: UCC28740

Hi Jasvant,

Thank you for the query on UCC28740.

Please share the excel calculator for the design along with the schematic and targetted power levels for the design to help identify the problem.

Regards,

Harish

Hi Harish,

Please find the calculation sheet and schematic with the complete design details including power levels as you requested. Kindly revert at the earliest.

Thanks,

JasvantUCC28740_Schematic.pdfCalculation_sheet_UCC28740.xlsx

Hi Jasvant,

Thank you for forwarding the schematic. From my initial observations and the fact that your max. Fsw is limited to 35khz, the magnetizing inductance seems very high because Fsw is limited to a design constraint of 35khz. You are correct in your reasoning about the feedback. The controller has max. Fsw of 100khz and you might probably need to decrease the magnetising inductance to 800uH to push max. Fsw to 80khz or something as per the attached calculator.Copy of UCC28740_worksheet.xlsx

What is the purpose J2 short connecting it to PGND? I checked your calculations it looks fine. There could could be leakage inductance effects on cross-regulation suggested in this video: https://training.ti.com/multiple-output-flybacks-how-improve-cross-regulation

But as first step, I would recommend reducing the Lm by adding some paper/something at the air gap and running the test.

Please let me know your observations.

Regards,

Harish

HI Harish,

Thanks for the reply.

However, I have some doubts as follows:

1. The maximum frequency was selected following TIDA-010000 reference design where it is chosen to be 32kHz. So it should be okay, Right?
2. The board did turn-on initially for 2 days so, why high Lm did not create any problem that time?

And to answer your questions

• J2 is not short in board, it just provides an option for non-isolation if needed similar to TIDA-010000 schematic. 
• Leakage Inductance is less than 2% of Lm so that should not come in picture. Also, first board should turn-on without restarting then we can focus on cross-regulation improvement.

  Thanks, 

Jasvant Singh

Hi Jasvant,

Thank you for the reply. 

TIDA-010000 I understand operates at a low switching frequency of 32Khz as you mentioned because it is for a motor driver application and they want to minimize the overall switching losses by reducing the turn off switching losses which occur at peak current. So I take back my claim on reducing Lm. I was not informed earlier that the board did start up and run in your earlier testing.

The UCC28740 is trying to start and is unable to complete it before VDD reaches the UVLO turn-off threshold.  Since the turns-ratio of the +24V (reference) winding to the AUX winding is 0.5 and the 24V only reaches about 5V, the AUX winding reflected voltage is unable to sustain the VDD to keep running.   So, something is preventing the output from rising further than 5V, and this needs to be investigated and debugged. There could be excessive load on +18V, or excess load on any of the other outputs. Can you increase the load slowly and let us kow your observations. Please be careful to check for OVP at light load.

With this limited frequency range the device may operate in a restricted range - in the first FM range and some AM range, not all features like the CC mode may be available. You may additionally need to increase the Vdd capacitance or increase the Auxwinding turn ratio to gain higer VDD voltage.

You had informed about not getting an OCP fault. I would suggest double checking this by reducing the Rcs value. If not Vcs might need to detect minimum voltage requires its first pulse can make CS voltage reach 200mV within 12us. You might want to check ths as well.

Lastly please refer section section 2.1, 2.5 which talk about UVLO and startup issue.

https://www.ti.com/lit/an/slua783/slua783.pdf?ts=1665562790845

Let me know your observations on the above points. 

Regards,

Harish