• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 663 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.

Original question:

LM5022: Driving GaN Fet with LM5022

LM5022: Multi phase design for LM5155 controller

Marco Degano
Prodigy 180 points
Part Number: LM5155
Other Parts Discussed in Thread: PMP7282, , LM5022, LM5170-Q1, LM5170

Hello, do you have any suggestions on how to make a multi-phase, interleaved boost controller using this device?

I am looking into a way to share equally the currents between the legs

  • 0
    TI__Mastermind 31225 points

    Hi Marco,

    Please take a look at PMP7282's schematic.

    http://www.ti.com/tool/PMP7282

    You can interleave the LM5155 the clocks through a 555 timer and then generate the 180 degree phase shift using a FET in a common source configuration

    Thanks,

    Richard 

  • 0 in reply to Richard Fu22
    Prodigy 180 points

    Hi Richard,

    yes I understand the concept of clock synchronisation;

    however I am worried about the drift of the DC current in CCM between the different phases: for the same reference and duty cycle, nothing will set the currents to the same DC value among the phases (I am planning on having at least four).

    Did you ever encounter this problem?

    Thanks,
    Marco

  • 0 in reply to Marco Degano
    TI__Guru* 78300 points

    Hi Marco,

    You are correct that the load current sharing among the four phased would not be perfectly balanced.  It is a problem for all peak current mode controlled circuit, not only limited to the LM5155 or LM5022. But such a configuration does support higher power. 

    To make four LM5155 in interleaving operation, you need to generate four interleaved clocks each feed to the LM5155's SYNC pin.  In addition, you also need to disable the internal error amplifier by grounding all IC's FB pins, and tying their COMP together.  Then use an external Op-amp as the error amplifier to drive the common COMP node, such that each LM5155's PWM generator references to the same COMP voltage to achieve power sharing. 

    If your power is high, you can consider the LM5170-Q1 configuring in the boost mode.  The LM5170 can achieve well balanced power sharing because it is based on averaged current mode control scheme. It is also powerful and dual phase may be adequate for your application.  If you are interested, you can order an EVM for evaluation.


    Thanks,

    Youhao Xi, Applications Engineering

  • 0 in reply to Youhao Xi
    Prodigy 180 points

    Hi Youhao,

    Can I ask why are you suggesting to use an external error amplifier?

    What would happen if I just parallel the CS and leave all the internal amplifiers working, feeding the same reference to all of them?

    I understand they might differ slightly, but having all of them feeding to the same pin, the average output would always result in the required output voltage, isn't it?

    Regards,

    Marco

  • 0 in reply to Marco Degano
    TI__Guru* 78300 points

    Hi Marco,

    Thank you for asking.  We usually configure the multi phases in master and slave fashion, so you only need one circuit in control and the others just to follow.  It is a conventional way to parallel more ICs of current mode controller.  The COMP pins of these ICs determines the peak current, and use one IC as the master, and all other in slave by tying their COMP pins to the masters, so each phase will have the same peak current.  

    However, you cannot do this with the LM5155 because its COMP pin is a current source (gm output). The master will fight against these slaves.  So, I recommend to put all LM5155s in slave, and use an external Op-amp to control the COMP pins so each IC will see the same COMP, and get the same inductor peak current.  

    You cannot parallel the CS and leave the internal amplifiers working, to achieve the power sharing.

    Hope this clarifies.


    Thanks,

    Youhao