• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 4 answers
  • Subscribers 64 subscribers
  • Views 138 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

UCC28070-Q1: Adjusting UCC28070 OVP Threshold for Higher No-Load Bus Voltage

Sourov Roy
Prodigy 10 points
Part Number: UCC28070-Q1
Other Parts Discussed in Thread: UCC28070, UCC3817, UCC28180

Tool/software:

Hello everyone,

I'm currently working on a design that requires running the downstream converter of an interleaved PFC (using the UCC28070) at a DC bus voltage of 435V during no-load conditions. However, once the load is activated, the DC bus voltage should regulate at 350V.

I understand that the UCC28070 implements overvoltage protection (OVP) by continuously monitoring VVSENSE, and if it rises above 106% of the set regulation voltage (3.18V), the GDx outputs are immediately disabled to prevent excessive voltage. In my case, this means OVP would kick in at 350V × 1.06 = 371V, preventing operation at 435V during no-load conditions.

My question, Is there a way to override or adjust the internal OVP threshold to allow operation up to 435V without triggering shutdown?

I’m considering external circuits to artificially shift the sensed voltage, but I’d love to hear insights from anyone who has tackled a similar issue! Any suggestions or alternative approaches are welcome.

Thanks in advance!

  • 0
    TI__Mastermind 36860 points

    Hello Sourov, 

    The OVP threshold of the UCC28070 is an internal ratio with respect to the 3V regulation reference at VSENSE and cannot be altered externally. 

    Unfortunately, your idea of manipulating the divider to VSENSE input is the only way to accommodate your different regulation levels. 

    As you are aware, when the load kicks in and regulation is shifted down to 350V, the output voltage Vout must drop faster (due to loading) than the sensed voltage at VSENSE input to avoid triggering the OVP.  So the VSENSE divider must be controlled to decrease at a rate that allows the load to pull Vout down. 

    Your load must be capable of sustaining up to 435V stress for a finite period of time until reaching 350V steady-state. 

    Regards,
    Ulrich

  • 0 in reply to Ulrich Goerke
    Prodigy 10 points

    Hello Ulrich,

    I truly appreciate your detailed explanation and suggestion. That makes sense!

    Could you please elaborate on the following point?

    “As you are aware, when the load kicks in and regulation is shifted down to 350V, the output voltage Vout must drop faster (due to loading) than the sensed voltage at VSENSE input to avoid triggering the OVP. So the VSENSE divider must be controlled to decrease at a rate that allows the load to pull Vout down.”

    I’d like to understand more about how to ensure that the VSENSE divider decreases at an appropriate rate in synchronization with the natural voltage drop caused by the load. Are there any recommended circuit techniques or dynamic adjustment strategies that could help achieve this without causing instability or unwanted transients?

    I have a another question regarding the expected output voltage at no-load conditions when the VSENSE divider is designed for a 350V output voltage regulation. Given that the UCC28070 regulates the output voltage based on the feedback from VSENSE, I would like to understand how high the output voltage might rise in the absence of a load. Would the output naturally increase beyond 350V due to the lack of current draw, and if so, what factors would determine the actual no-load voltage?

    Additionally, would it be expected to reach the OVP threshold (350V × 1.06 = 371V) before the controller shuts down the gate drives?

    Looking forward to your insights!

    Thanks,
    Sourov

  • 0 in reply to Sourov Roy
    TI__Expert 8925 points

    Sourov,

    Can you share why you want to achieve this: 435V during no-load conditions and regulate at 350V with load?

    I don't think manipulating VSENSE is a good idea to achieve what you want. 

    Please share more application background so we can help more.

    Ning

  • 0 in reply to Ning Tan
    Prodigy 10 points

    Dear Ning,

    Thank you for your response. I appreciate your willingness to help.

    The application requires the final output to be a bipolar pulse with a 450V amplitude to activate the load. Once the load is activated, the required amplitude reduces to 360V. The output stage consists of an H-bridge circuit, which operates using the DC voltage provided by the PFC.

    The existing design is a 500W system utilizing a single-phase PFC that includes a separate OVP mechanism. However, for the 1000W version, I opted for an interleaved PFC design using the UCC28070, as it is well-suited for interleaving.

    I understand that manipulating VSENSE may not be the ideal approach, and I’d appreciate any alternative suggestions you might have to achieve the required voltage transition while ensuring stable operation.

    Looking forward to your insights!

    Best regards,
    Sourov

  • 0 in reply to Sourov Roy
    TI__Expert 8925 points

    Sourov,

    How fast do you need the Vout to ramp down from 450V to 360V once the load is activated? and vice versa when the load is removed.

    Why not set the Vout to 360V all the time? I don't understand the benefits of 450V at no load condition. 

    Best,

    Ning

  • 0 in reply to Ning Tan
    Prodigy 10 points

    Hi Ning,

    The ramp-down speed is not a critical requirement; it does not need to happen very quickly.

    The reason for the higher initial voltage is that a higher voltage is required to ignite the plasma, and once the plasma is activated, a lower, regulated voltage is needed to safely maintain it. This ensures stable operation while preventing excessive stress on the system.

    I am also open to exploring other controller chips that feature a separate OVP pin, as this might provide more flexibility in achieving the required voltage transition. Could you suggest?

    Additionally, given that the UCC28070 regulates the output voltage based on the feedback from VSENSE, I would like to understand how high the output voltage might rise in the absence of a load.

    Thanks,
    Sourov

  • 0 in reply to Sourov Roy
    TI__Expert 8925 points

    Sourov,

    I assume your input is AC voltage so you need PFC. If your input is DC voltage, then you don't need a PFC IC. An regular boost controller would work here.

    Now go back to your application, I want to clarify few things.

    1) You said the ramp down speed is not critical, but do you have a number? From 440V to 360V in 10ms? 1s? 10s?

    2) I assume the ignition may require higher power so >1kW for how long? Do you have the power profile? This info is helpful for us to provide better suggestions.

    3) How do you determine the ignition is successful so the Vout shall be lowered? By measuring the current?

    4) I understand you want to use interleaved PFC for 1kW but CCM single phase PFC can support 1kW without an issue. You can just scale the power stage.

    The OVP threshold is 106% of regulated voltage as you know. Having said that, just because OVP is triggered, the IC won't latch on fault. It just simply stops switching. You can see the flow below.

    So, you can simply set the Vout = 450V at beginning when there is no load. Then once the ignition is successful, you manually adjust the Vsense network to target 360V Vout. Yes, the OVP will be triggered and switching will stop momentarily (which is what you want). Without energy transfer, the load drains the Vout down to 360V, then the switching will resume. Since you don't care the ramp down time, you can adjust the Vout capacitor to achieve the timing you need based on the load. One caveat, your input voltage needs to be lower than 360V all the time.

    Best,

    Ning

  • 0 in reply to Ning Tan
    Prodigy 10 points

    Hi Ning,

    Thank you for your clarification and for outlining a potential approach.

    To answer your questions:

    • The ramp-down time needs to be at most 100ms.
    • In the worst-case scenario, the ignition power can reach 1.6 times the regulated power, meaning it could go as high as 1.6kW for 100ms before stabilizing at 1kW.
    • Yes, measuring the current is a viable method to detect ignition—once ignition occurs, there is a clear current draw.
    • I understand that a 1kW design can be achieved using a single-stage CCM PFC, and we will opt for that if the interleaved solution introduces risks. That’s why I want to fully understand this issue before making a final decision.

    The existing 500W solution utilizes a single-stage CCM controller with a separate OVP pin, which is set to 445V, while the regulation VSENSE pin is set to 350V. From the scope capture, at no load before ignition, the voltage reaches 440V, and once ignition happens, it gradually ramps down to 350V within 60ms.

    Given this behavior, I have the following questions:

    1. Why does the no-load voltage rise to 440V even though the VSENSE pin is set to 350V?
    2. In the case of the UCC28070 interleaved controller, would the same behavior occur if it had a separate OVP pin?

    Additionally, you mentioned:

    “So, you can simply set the Vout = 450V at the beginning when there is no load. Then, once the ignition is successful, you manually adjust the VSENSE network to target 360V Vout.”

    What method do you recommend for adjusting the VSENSE network manually?

    Considering these concerns, what controller IC would you recommend for both single-stage and interleaved PFC solutions that could best accommodate this application?

    Looking forward to your insights!

    Best regards,
    Sourov

  • 0 in reply to Sourov Roy
    TI__Expert 8925 points

    Sourov,

    I assume this question "Why does the no-load voltage rise to 440V even though the VSENSE pin is set to 350V?" applies to your existing solution. I don't know which controller you are using on the existing design so I cannot comment.

    For UCC28070, if your Vsense is set for 350V Vout, it should regulate at 350V, not at 445V.

    As far as the manual method to adjust the Vsense point, you can consider the following example: use your current sense to generate a signal to turn off one current path (FET is off), effectively increasing the Vsense voltage to trigger OVP after the ignition. Once Vout drops to 350V, Vsense is back to regulation.

    Best,

    Ning

  • 0 in reply to Ning Tan
    Prodigy 10 points

    Hi Ning,

    Thank you for your response.

    The existing solution uses the L4981 from STMicro, which I believe is similar to UCC3817. My preference is to use UCC3817 for this design due to better reference resources.

    I was wondering—if I follow a similar approach as in the L4981, where the separate OVP pin is set to 445V and the regulation VSENSE pin is set to 350V, would this effectively achieve the desired operation as intended? Or would there be any limitations in this approach that I should be aware of?

    Looking forward to your thoughts.

    Best regards,

    Sourov

  • 0 in reply to Sourov Roy
    TI__Expert 8925 points

    UCC3817 should work the way you want. However, I'd recommend UCC28180 if you want a single phase CCM PFC controller.

    I don't think you need to care much about the OVP point. The key in your application is to dynamically change the Vout via Vsense manipulation.

    Ning