This thread has been locked.

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.

UCC28710 capacity on VS - voltage sense (pin 2)

Other Parts Discussed in Thread: UCC28710, UCC28910FBEVM-526, UCC28711EVM-160

Hello everyone,

currently we are facing a problem with the output voltage of the UCC28710. When our device is influenced by indirect ESD or a burst pulse on the supply cable, the output voltage falls shortly below the reset voltage threshold.

We can not change the layout, but we must use the current design as it is. So we can change components, or add a limited number of components which are already in the layout, but not populated.

Applying larger capacitors (primary and secondary) was not helpful. But when adding a capacitor of 4p7 on the Voltage-Sense pin (2) of the UCC28710 the secondary voltage is stable enough to prevent the problems. While capacitors are generally not recommended on the voltage-sense pin, this is a very small value which does not seem to have a negative effect on the switching behavior.

What are the critical parameters that  we should watch out for in our measurements to make sure that we are not deteriorate the performance?

(We are using a voltage divider of 82k / 20k at the auxilary winding).

Thank you for the advice,

Horst

  • Hello Horst,
    It sounds like you have an understanding of the fault mechanism. As you see, any unexpected disturbance on the VS pin can cause either an overvoltage shutdown or an input under-voltage shutdown. You can filter out this transient with a small capacitor on the VS pin. In your case, adding 4p7 will only delay the VS signal by less than 400nS. This should have no effect in the output voltage set value, since the sampler in the UCC28710 is only active 600nS after the FET turns off. Therefore any slight signal distortion is not seen by the controller. I have seen designs with a 76k8/38k3 and a 10pF capacitor.
    Also, please review your pcb layout. If you look at the pcb layout for the UCC28910FBEVM-526, you will see that the primary side ground is routed under the transformer, effectively shielding the ESD on the secondary from getting into the signal pins around the IC. And if your design uses a common mode choke, it is important to put spark gaps or resistors across both sides of the windings.
    Hopes this helps !

    John Griffin
  • Thank you for the answer. It is very useful to have learned more about what to look out for.

    We will certainly review the PCB layout and try to improve the layout where possible. Hopefully we then can remove the capacitor again.

    But it is important to us to have a intermediate solution.

     

    Best regards

    Horst

  •  

    Hello,

     

    This is a follow up to my previous post.

     

    Back in December I did do tests with the mentioned 4p7 capacitor and could not see any adverse effects.

    However we now received a batch of boards with this modification. Many boards are fine, but some have a low frequency ripple of about 70mV / 1kHz. This usually appears at low input voltage and low load conditions. The ripple disappears when removing this capacitor, but an additional capacitance as low as 1pF can cause the output voltage ripple.

    The ripple on the output voltage also disappeared when we changed the transformer.

     

    As a very small capacitor can cause this ripple we suspect that this capacitor is not the root cause. Looking at the voltage on the auxiliary winding as well as the voltage on the VS pin (2), it is not clear where the problem comes from. (see attached images).

     

    What should we look out for to remove the ripple on the output voltage?

    How does the chip detect the knee point and when does it measure the voltage?

     

  • Hello Horst,

    Hope you have had success resolving this issue. I am observing the same type of output ripple behavior with my UCC28710 application and was wondering if you could share your outcome?

    I find that the UCC28710 is able to operate in regulation but periodically will allow the output voltage to decay for an extended period without the control loop responding. And, just as you illustrated above, it will sharply increase the switching current and frequency to recover the output level. I have seen this behavior on both my board and the UCC28711EVM-160 evaluation board from TI. The issue seems to be dependent on the operating point of the control law profile in constant voltage mode.