+ Post New Thread
Results 1 to 3 of 3
  1. #1
    Member level 1
    Points: 534, Level: 4

    Join Date
    Feb 2017
    0 / 0

    High current source?

    I'm looking for a conception for for construction current source ,
    - current 0-75A
    - output voltage 0-24V
    - output power 0-1800W
    - programmable current with step 100mA
    - full digital control on output current as switching from C1 to C2 level with programable frequence 0-400Hz
    - full digital control on output current as programable slew rate from C1 to C2
    - full digital control on output polarization

    How would you deal with it?

    My first idea.
    I certainly do not see it as real that I could design and mount the inverter (switching) directly from 230V.

    - stage one - Switching power supply 48V 40A, for first test have 2c PS 24V 35A with serial and parallel joining
    - second stage - Many parallel Synchronous Buck DC-DC inventer, for example 3x PS with LM5154 switching frequency between 150-400kHz with full ceramics capacitor and small SMD Instructor any as HCI2212 with high side current monitor and small MCU for control.
    - third stage full high current MOSFET bridge for On/Off switching and polarization change. MCU for control bridge realize too controll full current source and output current control.

    Is it stupid?
    Is not that complicated?
    Better idea?
    Is there a more suitable IC for driving Bock converter in this design then LM5145 ?


    •   AltAdvertisement


  2. #2
    Advanced Member level 3
    Points: 5,177, Level: 17

    Join Date
    Feb 2014
    244 / 244

    Re: High current source?

    So this is AC/DC 0-24V peak(?) 0-400Hz? I design very similar AC output digitally controlled inverters. How tight are your specs for output quality (amplitude accuracy/harmonic distortion). That dictates a lot about the control scheme and topology you need. For example you plan to have a polarity bridge but that's guaranteed to create distortion. Maybe that's fine maybe that's not.

    In general I'd look for off-the-shelf solutions for generating a intermediate DC bus of 36-48V to feed a full bridge, or generate +36/-36V and roll your own half bridge solution (but look up 'bus pumping').

    Note that when you have multiple fets or bridges in parallel you need to make a choice of how you're paralleling them. Direct paralleling (fets sharing gate drives), multiple phases without individual control, or multiple current controlled stages summed together (the first two have balancing risk, the 3rd is more complex).

    When you reference the MCU do you plan on having it driving the PWM signals to the gate drives or drive a DAC that creates a reference for an analog loop? If you're going to have the MCU driving the PWM you'll want to study the PWM timing resolution that the MCU can put out. There are tons of half bridge solutions but they may or may not be applicable to your AC output application and I'm not clear on how they integrate with your digital control.

    Lots of choices that depend on lots of details....

    •   AltAdvertisement


  3. #3
    Advanced Member level 5
    Points: 38,991, Level: 48

    Join Date
    Mar 2008
    1836 / 1836

    Re: High current source?

    If it's unidirectional then I'd look at a current mode controlled
    DC-DC and how well its ISET pin fidelity works for you. Some
    (like ones I designed, but they're too low-voltage) let you
    wrap the sense resistor through the error amp to linearize
    the transfer function.

    You could use a DAC to drive the control input.

    Whether this would have the agility you want, comes down
    to detailed loop design, fSW and so on. Figuring 24V + margin
    for VIN, however, is going to probably keep you in sub-MHz
    technologies unless you're clever (like, use a faster 5V PWM
    controller, a half bridge driver and GaN FETs - current sense
    would then be the main challenge).

    A PFM (hysteretic) control, high side referred with isolated
    control input might be elegant and simpler. But you then have
    a varying frequency content that's harder to filter out (esp.
    at the low current end where pulses are sparse and frequency,

--[[ ]]--