• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 63 subscribers
  • Views 323 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TPS54821: Discontinuous mode existence (indicated by TINA-TI)

Oscar Mitchell66
Prodigy 170 points
Part Number: TPS54821
Other Parts Discussed in Thread: TINA-TI,

The part's datasheet says:

7.3.2 Continuous Current Mode Operation (CCM) As a synchronous buck converter, the device normally works in CCM (Continuous Conduction Mode) under all load conditions.

In Webench I can't have Vout=0.9V with Vin=8.4V at 1.25 MHz switching frequency (and a 3A load) because of the minimum on time (per 8.2.2.9.1 Minimum Output Voltage)

I then export the design from Webench to TINA-TI for simulation.

When I operate with Vin=5V, Vout=0.9V and 1.25MHz switching frequency the simulation runs and it appears to be operating in CCM. Great.

If I increase the input voltage to 8.4V (no other changes) then the simulation runs and inspection of the inductor current shows it be operating in DCM. This would be marvelous if it really can do that, if it can't I have to do more work.

I don't think this is due to Over Voltage Protection per '7.4.9 Output Overvoltage Protection (OVP)' because DCM appears to be present in the simulation well before the feedback pin reaches 0.6V

So the question is: Will it behave per the TINA-TI simulation and do both CCM and DCM as required?

  • 0
    TI__Expert 7500 points

    Hi,

    When minimum on time is triggered, IC will turn on high side with minimum on time. Then the energy with minimum on time is larger than the needed. So the Vout will rise and trigger OVP. As datasheet metioned, 

    "If the VSENSE pin voltage is greater than the OVP threshold the high-side MOSFET is turned off preventing current from flowing to the output and minimizing output overshoot. When the VSENSE voltage drops lower than the OVP threshold, the high-side MOSFET is allowed to turn on at the next clock cycle." 

    So i think your "DCM" is due to OVP. But this is not a traditional DCM.

    Thanks,

    Lishuang

  • 0 in reply to Lishuang Zhao
    Prodigy 170 points

    But, in the TINA-TI simulation:

    a) this apparent DCM is exhibited in the simulation well before VSENSE approaches 0.6V

    b) the output voltage ends up at exactly the set voltage

    c) the VSENSE pin ends up at exactly 0.6V

    A lot of parameters appear to be passed to the macro like output voltage and switching frequency from the high level so I have no idea what is going on inside.

    I guess the bottom line is that I will have to limit the input voltage with a LDO since the power source is a battery that can reach well above 7.6V.

    (reluctance is due to there being no good fit for this and a discrete one isn't trivial)

    Not sure this can be marked as 'resolved' because I perceive it is currently more 'no high confidence explanation available so work around it'

  • 0 in reply to Oscar Mitchell66
    TI__Expert 7500 points

    Hi Oscar,

    For your application, can you decrease your frequency from 1.2MHz to 700KHz? If so, it will not trigger minimum on time. It can work normally when input voltage within 8.4V.

    Thanks,

    Lishuang

  • 0 in reply to Lishuang Zhao
    Prodigy 170 points

    No, I am stuck with 1.25 MHz.

    But for this problem the TPS54821 is great.

  • 0 in reply to Oscar Mitchell66
    TI__Expert 7500 points

    Hi Oscar,

    I have some error to simulate it in TINA.

    I tested it on bench. The waveform is as below: it does trigger OVP so high side off for a long time. It is matched with design.

    Thanks,

    Lishuang

  • 0 in reply to Lishuang Zhao
    Prodigy 170 points

    Great, thank you.

    That is very informative.

    It does not look like the simulation.

    Clearly real life is more important.

    I suppose that the macro is not great at reflecting the operation of the device.

    Thank you again.