UCC28951: Transformer heating shoot up gradually

Hello,

Your transformer current does not look typical it also has lots of ringing on it.  Your rheostat is a wire wound inductor.  This inductance may be causing the excessive inductor ringing that you are seeing on the primary.  Try using a resistive load or electronic load to see if you can remove it.

When it comes to transformer heating, there is core loss and copper loss.  If your peak current and RMS current is what you told the magnetic designer it should be the magnetic should not be getting hot.  You also need to make sure it was designed for the ambient temperature where the design is being used.

I would double check the RMS current and the peak current in the transformer to make sure it is what you told the magnetic designer. If it is then then the transformer was designed incorrectly.  If it is not then gave the magnetic design the wrong specifications and need to give the designer the correct specifications.

Regards,

Hello,

I reviewed your 4024.power schematic which is the power stage for the PSFB.  With an output power of 3.3 kW I don't believe 68 uF of output capacitance is enough.  You might want to check the size of the output capacitor bank.  The following link will bring you to an application note (slua560) goes through the step by step design process of a phase shifted full bridge (FSFB).  There is a section on sizing the output capacitor based on loading and hold up time.  You can use this to check and size your output capacitor.

https://www.ti.com/lit/pdf/slua560

I reviewed psfb_updated and it looks like you are tying to use an adaptive delay approach.  This is being done by the resistor dividers off the current sense pin to ADEL and ADELEF.   I would recommend using a fixed delay approach because it easier to setup the timing.  How to use this approach is covered in application note SLUA560.

You had mentioned that the design is stable and the only issue is transformer heating.  When I look at the transformer primary voltage vs load current.  The duty cycle seems to be changing every switching cycle.  This is not normal and the feedback loops should not be correcting every switching cycle. In peak current mode control designs and using an opto isolator it is recommended to cross the voltage and current loops to less then 5 kHz.  You might want to double check and reset your compensation to increase phase margin and gain margin.  Typical loops cross over at 2 to 3 kHz.

To prevent a transformer from saturating you need to maintain volt second balance across it.  Since the input voltage of the transformer shows varying duty cycle every switching cycle, the transformer might be partially saturating leading to heating of the transformer.  I would work on stabilizing the feedback loops to get the duty cycle to be more constant.  This should help reduce partial transformer saturation and heating in the transformer.

Regards,

Hello,

Your compensation scheme looks different than PMP6712.  The PMP6712 seems to be using to PSFB in parallel/interleaved.  Your design is different only using a single phase.  

I reviewed the compensation scheme used in PMP6712 and I don't believe the opto transistor is connected correctly (Opto E).   The opto  transistor should be ground referenced.

The following schematic shows the proper connections of the opto transistor and TL431 feedback for voltage.   Please note the opto transistor is ground referenced.

If you are trying to do constant current constant voltage mode control.  The following link will bring you to a 2 kW battery charger reference design that shows a feedback scheme that will work for this application.

https://www.ti.com/tool/PMP8740#tech-docs

Regards,

Hello,

Your circuit is different than any other voltage control, current loop that I have worked on.  Also, the UCC28951 was designed for secondary side control. So there is no application note that I can send you to compensate both loops.  Other customers have gotten the UCC28951 to work with opto isolator feedback, so you should be able too.  I believe your opto coupler is set up incorrectly.  I marked up the schematic showing how the connections should be setup.  This feedback is similar to one previously referenced.

If you remove diode D3 in your feedback circuitry you will only have the voltage loop feedback.  If we assume the opto has a CTR of 1, you can use the excel to compensate the voltage loop.  So you don't run into issues with the opto pole, which generally occurs at 5 to 10 kHz, I would recommend crossing the voltage loop over at 2 kHz.  I would then double check the feedback loop with a network analyzer.  

The following link will bring you to an application note that discusses how to design a PSFB using the UCC28951.  There is a section on compensating the voltage loop in the application note.  The excel tool uses the same equations that are in the excel design tool.  Please make sure you add the correct amount of slope compensation to avoid sub harmonic oscillations in your design.  Setting up the slope compensation is covered in the application note and there are calculations for doing it in the excel design tool.

https://www.ti.com/lit/pdf/slua560

I would recommend compensating the voltage loop and testing it before trying to compensate the current loop.

I reviewed your feedback for the current loop and it looks similar to the 2 kW design, I refenced previously.  I believe it will work with the corrected opto coupler feedback as well.  However, you will have to debug it and compensate this loop.  When you do cross over the current loop, you might want to cross it over at 1 kHz, so it does not fight the voltage loop.

Regards,

Hello,

Thanks for letting me know.

Regards,

Hello,

Could you send me updated schematic for review?

Regards,