High Current Limiter w/ Opamp, Flip Flop & MOSFET

Discussion in 'The Projects Forum' started by joeegar, Nov 18, 2009.

  1. joeegar

    Thread Starter New Member

    Nov 18, 2009
    I'm currently working on a high current (Max 10A) limiter. I've taken a look around to see what other people are doing and figured the best approach would be to use a MOSFET as a switch and an Opamp/resistor for the feedback. The output from the Opamp (or comparator in this function) goes to the Data input of a 74175 Flip Flop. This FF is clocked at 100kHz and reset at 10kHz. The idea being that if the current goes above the set value the OpAmp goes high (or high enough) and the FF clears the output upon a rising edge of the clock. The inverted output is tied to the MOSFET...

    See image of this circuit here: http://www.solveering.com/freestuff/files/currentLim.jpg

    The issue I have is that the MOSFET (a IRFZ44) gets way too hot...

    What am I missing? I would like to keep the part count to a minimum and am prepared to sacrifice accuracy of the limit but I would like to keep the maximum heat generated by the FET below about 1-2W. Having the 74HCT175 gives a fairly steep rising edge (~700ns) and using a higher current driver (TC1427) gets this down to about 250ns but the FET still gets really hot.

    Any comments? Am I missing something?
  2. Audioguru


    Dec 20, 2007
    The IRFZ44 Mosfet is spec'd with 10V from gate to source. Yours has less than half that so it does not completely turn on and therefore gets very hot.

    Use a logic-level Mosfet like an IRF3711Z.
  3. SgtWookie


    Jul 17, 2007
    What is your value for L?

    Another logic-level MOSFET to consider is the IRLR7807/IRLU7807.

    1) You are using an abysmally slow LM324 opamp as a comparator, which has a 1.1MHz bandwidth unity gain at room temperature on a good day with a tailwind while going down a very steep hill.

    Except you're using it as a comparator in open loop, which means that it has just a tad bit faster response than my Grandma, who died ten years ago.

    2) You have the sense resistor below the source terminal of an N-channel MOSFET, so the only time it can sense current is when the MOSFET is conducting.

    3) There is no path for L1's current when Q turns off, so Q will be blasted by about a kazillion volts, and be vaporized.
  4. Ron H

    AAC Fanatic!

    Apr 14, 2005
    As drawn, if the current is above the set value, the op amp output goes low. When this gets clocked into the FF, /Q1 will stay high, and the current will continue to be too high.
  5. SgtWookie


    Jul 17, 2007
    Here is a modification of your circuit; see the attached.

    For simplicity's sake, the N-channel MOSFET has been swapped out for a P-channel of similar type, and the whole inductor/Rload/Rsense arrangement flipped around. This was necessary because the replacement used for the opamp can sense down to ground, but not up to V+.

    200uH was selected for the value of L1. It is a trade-off between ripple voltage, start-up time and control loop time.

    D2, a high-current Schottky diode has been added across L1/Rload/Rsense to provide a current path when the MOSFET turns off (D1 was in the schematic briefly, but has been removed).

    C1 suppresses transients across D2. In a real circuit, this may or may not be necessary.

    R4/R5 sets the current level; 10A is maximum.

    The slow-to-respond opamp running open loop has been replaced by an LM339 comparator, which is designed for this type of thing.

    Since the LM339 has an open collector output, it was necessary to add R3, a pull-up resistor to act as a current source.

    R1 limits the base current to Q2 to under 5mA. Q2 acts as a level converter to pull the gate of Q1 to nearly 0v when Q0 goes high.

    R6 provides damping for the gate; without it, the gate may "ring" or oscillate due to the LC time constant of the gate wiring (inductive) and the capacitance of the gate.
    R2 pulls the gate to +V when Q2 is not conducting.

    This is not a practical example of a MOSFET gate driver; it is merely a "quick and dirty" to demonstrate that the circuit can be made to work.

    The current through L1 is displayed in the plot on the bottom in green; it's averaging roughly 10A. The yellow plot shows the gate voltage; when it's 15v the MOSFET is turned off; when it's 0v the MOSFET is on.

    The IRL9Z34 MOSFET is not really adequate for this task, but it was available in the library.