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.

UCC28740: Supply takes a while to start up

Part Number: UCC28740

Hi TI,

I am designing a custom circuit for the UCC28740 that takes universal AC input and outputs 35.7VDC@0.6A. When testing with a very light load, the supply takes a while to start regulating the output, but eventually starts after ~30 seconds. Attaching a scope probe across the AUX winding seems to help speed up the startup. I have increased VDD to twice the value specified by Webench, but should I increase this even more? What could are some reasons why the supply doesn't start regulating after a while. Is the output capacitance too large?

Thanks for the help,

Bryan Grant

  • Bryan 

    If you have not already, we recommend that you use our excel design calculator http://www.ti.com/lit/zip/sluc487 to verify that the component values used in your design are correct.

    We don't recommend changing any values in your design unless we understand what is going on and expect the performance to improve by changing it.  Before changing any component values we recommend you measure aux winding, VDD, CS and DRV pin.  Please review section 7.3.3 (page 12) and section 7.3.4 (page 12-13). 

    Based on the limited information you have provided, a guess as to what could be happening is that at these light load conditions VDD is discharging below VVDD(off) of 7.75V typical during this light load startup.  You need to confirm this by measuring VDD voltage and see if it jumps between VVDD(on) of 21V typical and VVDD(off).  If this is the case, increasing VDD voltage and capacitance will help keep VDD above VVDD(off).  Another alternative option is to add a pre-load resistor on the output to insure that UCC28740 minimum frequency is higher, allowing it to provide enough bias power regardless of load condition.

    Best Regards,

    Eric

  • Hi Eric,

    Thanks for the response.

    I have been using the Excel tool as well, and have verified the component values used except for the differences below.

    The difference I see are that it recommends an output capacitance of 680uF instead of 39uF, as Webench recommended. I think this is potentially due to minimum voltage ripple or transient response requirements. I left those as the default values in the Excel tool, as they are not critical at the moment. And, I'm actually using a 100uF capacitor.

    Another difference I see is that the primary inductance could be different because I wound the transformer myself. I verified the turn ratios and polarity, but I'm not sure about the inductance. I measured it in-circuit with an LCR meter and got about 35mH instead of 1.27mH, as designed by Webench, but it had already been wound with the other coils and soldered into the circuit. Another thing is that the Excel tool recommends a lower primary to secondary turn ratio than the transformer selected from Webench. A red comment shows up that says it might be more efficient, but may limit operation at lower voltages.

    I also verified that VDD is discharging below the UVLO value, and increased the VDD capacitance, but this didn't solve the issue. I already had a 100K resistor across the output of the circuit, but maybe I should add a higher value?

    Thanks for the help,

    Bryan

  • Bryan

    Your actual inductance of 35mH is 27.5x larger than what the calculations specify.  Using such a larger inductance will cause the switching frequency to be much lower than you expect.  At light load you may have some issues since there is no room in the modulation range for UCC28740 to properly regulate the output.

    For VDD voltage, before increasing VDD capacitance please check to see if it's decreasing since the aux voltage is reducing below UVLO turn off or if it's getting pulled actively low.  If it has a very fast discharge to UVLO turn off, there is a chance that a fault condition is being triggered that is shutting down UCC28740.  Section 7.3.4 Fault Protection (page 12) of the datasheet has details on all the faults that can be triggered which would cause such an event.

    Best Regards,

    Eric

  • Hi Eric,

    I just want to confirm my thinking, but, in the downloaded transformer documentation, it specifies the Inductance factor of the transformer to be 144nH. However, when I look up the datasheet for the recommended transformer core (B65651W0000R087), the inductance factor is 3.6uH. This is a 25x difference, which may account for the value I am measuring.

    https://www.tdk-electronics.tdk.com/inf/80/db/fer/p_18_11.pdf

    Could you confirm that this could be a problem, and, if so, I should find and use an alternative core with an inductance value closer to 144nH?

    Thanks again for the help,

    Bryan

    20191115-Tranformer-Design.pdf

  • Bryan

    The key inductance you should be designing around is primary magnetizing inductance Lpr 1.27mH, not Al.

    Inductance Factor (Al) 144.0nH is a material parameter used in magnetic construction.  If you use a core material that has a different Al, what you should do is change the number of turns to make sure that Lpr is equal to 1.27mH.

    This webpage has more information on Al and transformer construction https://www.mag-inc.com/Design/Design-Guides/Inductor-Design-with-Magnetics-Powder-Cores

    Best Regards,

    Eric

  • Hi Eric,

    Thanks for the webpage. I'll take a look so I have a better understanding of transformer construction.

    My concern was that it looks like the recommended core (B65651W0000R087) has an inductance factor of 3.6uH instead of 144nH, as listed in the transformer parameters. Using the inductance factor equation (L=AL*N^2), this would account for the ~25x larger inductance value I am measuring.

    If I use the recommended core, I would have to lower the number of turns from 94 to ~19 to get close to the Lpr, which may make it difficult to meet the performance requirements for the secondary and auxiliary coils, but I will look into it further.

    Thanks again for the help,

    Bryan

  • Bryan

    The spec list from TDK you shared has an "ungapped" Al of 3600nH.  The text right below that table has a comment "other AL values/air gaps and materials available on request – see Processing remarks on page 7."

    If you would like to use the transformer design that came from the webench transformer design tool, you need to add an air gap to this core to reduce Al to 144.0nH.  A gap is required in a flyback transformer to enable proper operation.  Details about why are in page 2-2, 2-3 (15-16) in section 2 and section 5 of the Magnetic Design Handbook https://www.ti.com/seclit/ml/slup132/slup132.pdf

    Best Regards,

    Eric