UCC28740: Protection function may stop oscillation at startup

Hello,

The UCC28740 uses primary fault sensing that can be faulsely triggered by aux widing or current sense noise.  This is a common issue with primary fault sensing.

The following link will bring you to an application note that described primary fault sensing and the issues with it.  It discusses these issues and how to resolve them with RC snubbing.  The information in this application note should help you resolve the issues that you are seeing in your design.

https://www.ti.com/lit/pdf/sluaac5

Thank you for interest in Texas Instruments (TI) products.  If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Regards,

Thank you for your response.

I have reviewed paragraph 10 of the document you presented.

Is there any possible solution other than adjusting the CS capacitor value?

Also, is there any possible explanation for this phenomenon, which occurs more frequently at lower ambient temperatures?

I still can't imagine that the overcurrent protection is triggered by noise other than the timing of switching, paragraph 10 of SLUAAC5.

It is unlikely that oscillation would stop at such timing. Also, the startup environment is not bad (no circuits other than the DC-DC converter are running).

The explanation of 10.1.5 CS on the datasheet makes more sense than this explanation.

'A small filter capacitor may be placed on CS to GND, with short traces, to filter any ringing that may be present
at light load conditions when driving MOSFETs with large gate capacitance.'

This is for the following reasons

・The load is very light for about 2 seconds after startup.

・This phenomenon occurs only immediately after startup and is not reproduced from a few seconds after startup.

Q.  Is it possible that overcurrent protection is activated by the ringing described in Data Sheet 10.1.5?

Hello,

It is more likely the aux winding is tripping OVP that is causing your issue.

The CS noise issue would genreally be caused by EMI.  However, if that is the case the best option is to add filtering capacitance to the CS pin.

To figure out why the design is shutting down you need to evaluated VDD, VAUX, CS and Vout with an oscillascope.  The last three switching wavefroms when Vout drops out of regulation.  This will help you determine what the fault is and how to resolve it.  Appliaction note will give you instructions on how to resolve the issue once you have determined which issue it is.

https://www.ti.com/lit/pdf/sluaac5

Regards,

Thank you for your reply.

We actually observed each waveform when oscillation stopped (some kind of protective function operation).

Just before oscillation stops, if there is no CS capacitor, oscillation is unstable.

Looking at the waveform, it certainly doesn't look like it detects OCPs.

However, I don't even think OVP works.

Also, when VS is observed with a probe, it is certain that oscillation stops do not occur.

Which protective function will be activated to stop oscillation?

See pdf for more details.

TI_E2E_Q5_ENG_250110.pdf

Hello,

When trouble shooting I always recommend VDD, VAUX, CS and Vout with an oscillascope.

If you are probing the VS pin directly you are adding 10 pF of capacitance to the VS pin.  If there is ringing there this would help filter it out.

Noise at the VS pin can trigger OVP and if the VS pin is ringing around ground due to noise it will cause the device to shut down as well.  These issues are describe in application note sluaac5 as aux winding noise ringing below ground and triggering OVP.  This can be removed with RC snubbing as described in sluaac5.

Your report shows aux winding ringing below ground.  This is most likely what part of your issue is.  You will need dampen this noise to remove the issue.

I read your report and it seems that there is nois on the CS pin that is causing issues as well.  In the report filtering was added to remove this disturbance.  It is posssible to have both CS noise and aux winding noise to make your design unstable.

Regards,

Thank you for your reply.

Is it correct to say that there is a possibility of false positives for UVLO?（Section 8 of SLUAAC5  (Fig.8-1 'Vaux Ringing Below Ground Can trrip UVLO'))

If so, is it correct to say that the measures in section 9of SLUAAC5 are effective?

The problem now would be to prevent ringing when the FET is on. Is it effective to put a CR snubber in parallel with the output diode (CR16) ?