• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 907 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.

LMR38010: PWM, PFM and Forced PWM modes criteria

Sai Bharadwaj P
Prodigy 10 points
Part Number: LMR38010

Tool/software:

Hello, 

   I am using LMR38010 buck converter for generating 12V from 48v as a part of power supply designs for my board. I want to understand the conditions where the transition to PFM and forced PWM happens from the normal PWM operation. I have seen the datasheet/application note but it is not very clear there. I have few basic questions below : 

1. When will the transition between different modes happen exactly ?
2. If I have to change the transition point as above, say at different values of currents of my choice, what should I change in the buck converter design (inductor, capacitor, switching frequency ,etc) and how ?
3. Sometimes when I decrease the value of inductor then the point at which the DCM or I must say the PFM operation happens occurs at lower value of current compared to a higher value of inductor. How is this possible ? 
4. Please suggest corner cases for the above. I need my converter to function in CCM above 0.15-0.2A load current atleast. 

Please reply soon. 

  • 0
    TI__Genius 9570 points

    Hi,

    PFM operation and Forced PWM (FPWM) operation are two different ways to handle light load operation. This depends on the device variant you get, orderable part numbers with the letter F in them are FPWM variants. The non-FPWM variant transitions from PWM to PFM at light loads.

    1. In the non-FPWM variants this transition from PWM to PFM would happen after going into DCM and once peak current goes below IPEAK-MIN given in electrical characteristics table.

    2. You can change the transition point into DCM by changing inductor value and frequency. Higher inductor and frequency means less peak-to-peak inductor current, hence the transition will take place at a lower load current.

    3. That is the opposite of what is expected. Can you measure the ripple in both cases at the same load current and see if it matches expectation (higher with lower inductor) ?

    4. CCM operation with the non-FPWM variant would be difficult at 0.2A, it would require a large inductor. I would recommend going for the FPWM version of the device, this operates in CCM even at light loads.

    Regards,

    Niranjan

  • 0 in reply to Niranjan Nair
    Prodigy 10 points

    Hi Niranjan, 


    Thanks for the reply. 

    The part I'm using is LMR38010SDDAR. What mode does it support - I guess PFM ?

     Regarding the inductor peak current limit of 0.25A, can you please elaborate on what exactly this is ? How will the IC know when the converter has entered DCM ? Is it by checking whether inductor instantaneous current (which I suppose is created from HS and LS current sense, please correct me if I'm wrong) has gone below a certain value or touched zero ? How does this Ipeak-min relate to this mode shift ? 

    Also, when in PFM, what is the basis for turn on and turn off of the switches ? Output voltage is also taken into account ? 

    I'll do an analysis of different inductor values and get back on this point. However, I'm looking forward your reply on the CCM-DCM/PFM transition and also functioning in PFM. 

    Thanks. 

  • 0 in reply to Sai Bharadwaj P
    TI__Genius 9570 points

    Hi,

    Yes that part is non-FPWM (PFM mode at light loads).

    Yes, the mode shift is by checking current sense. After you have entered DCM if you keep reducing the load, the switching continues at the same frequency (PWM) for while until peak current goes lower than Ipeak-min. Once that happens, device enters PFM mode. 

    In PFM mode, the switch turn on is controlled by the feedback loop based on output voltage while turn off is controlled by how long it takes for inductor current to reach Ipeak-min. Thus the pulse width is decided by the time it takes current to reach Ipeak-min while frequency is modulated to regulate the output voltage.

    Regards,

    Niranjan

  • 0 in reply to Niranjan Nair
    Prodigy 10 points

    Okay. 

    The below is the waveform taken between SW pin to GND (Vds of LS switch). Does this mean DCM ? Can I confirm that once the switch waveorm starts becoming like this then it is DCM ?

  • 0 in reply to Sai Bharadwaj P
    TI__Genius 9570 points

    Yes, that ringing in the SW node at an intermediate level is indicative of DCM operation.

  • 0 in reply to Niranjan Nair
    Prodigy 10 points

    Thanks for that clarification. I see that we get more and more ringing like this, and also clamped to some lower/negative voltage level (like below) as we decrease the load current. Please tell what might be causing that and also what does the IC restrict. This behaviour is observed with different values of inductors and switching frequencies. 

    ...

  • 0 in reply to Sai Bharadwaj P
    TI__Genius 9570 points

    Hi,

    I didn't quite get what you mean by IC restricts. This is the expected behavior (DCM) at low load currents. In this part (non-FPWM variant) diode emulation is turned on for the low side FET, which prevents current flow in the opposite direction and causes the device to go into DCM.

    If you use an FPWM variant of the device, negative current flow through the low side FET is allowed and CCM is maintained even at light loads. There you would not see this ringing.

    Regards,

    Niranjan

  • 0 in reply to Niranjan Nair
    Prodigy 10 points

    Can you please let me know what is the cause of this ringing and the flow of current or voltage appearing in this ringing ? This is for a 48V input 12V buck converter and the output current is 0.5A max, inductor is 10uH, but we might increase it further, to achieve ccm in the operating range. At some conditions, this is looking as if the upper device is switching. Please confirm on this. I guess, we are stuck with this confusion right now. 

  • 0 in reply to Sai Bharadwaj P
    TI__Genius 9570 points

    The parts of the waveform where SW node goes high (to VIN voltage) is when high side switch is ON. The inductor current increases during this time and reaches a peak. Once the high side switch turns OFF and low side switch turns ON, the inductor current starts to reduce. This is when you see the low (only the small voltage drop across low side switch will be present). However as the inductor current reduces further and reaches zero, the low side switch stops conducting (due to diode emulation feature present in the non-FPWM variant) to prevent current flow in the opposite direction.

    Thus the SW node effectively floats during this time, until the high side switch turns ON again. This is when you see the ringing, caused by the inductor ringing with the parasitic capacitance present at the SW node.

  • 0 in reply to Niranjan Nair
    Prodigy 10 points

    Thanks a lot for this confirmation. So I take it that the frequency of this ringing is due to inductor and parasitic Cds of switches and inductor parasitic caps in parallel effectively, and this also has the diode in parallel to clamp at certain voltage. 

    The last two points i want to know are : 

    1. Is there any possibility of HS switch turning on internally itself during its off time as well as before the beginning of next switching cycle ?
    2. Can you please suggest if we need to place a snubber circuit  for supressing this, and which location would be the idea ? We want to do as minimum modifications as possible. We need to get away with as minimum components as possible. 

      We are using inductors either 33uh or 47uH currently and are not facing dcm. However, we would want to have the circuit with a lower value of inductance. 

  • +1 in reply to Sai Bharadwaj P
    TI__Genius 9570 points

    1. No, it is controlled by the set frequency.

    2. You can use an RC snubber at SW node to reduce this ringing, but it cannot be removed entirely.

    To avoid DCM and hence this ringing, either a larger inductor has to be used like you are doing now, or the FPWM version of the device has to be used.