Mosfets and full switched on

Discussion in 'General Electronics Chat' started by clintonb, Mar 25, 2015.

  1. clintonb

    Thread Starter Active Member

    Mar 6, 2011
    42
    0
    I built a circuit that allows a controller to operate a relay coil because of the current being too much for the controller pin.

    Everything works perfectly and nothing set off any alarm bells but now when researching pwm using the IRFZ44N it is said to get very hot?

    The problem apparently is that 5V gate voltage is not enough to switch the mosfet to fully on. What does "fully on" mean when its working with 5V at the moment.
     
  2. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,769
    969
    "Full On" means its resistance from drain to source is a low as possible.. Source/drain is fully saturated..
    You want a "logic level" mosfet in that 5V is all it takes to get it to be "fully on"..

    Without being "fully on" its resistance can be much higher hence the heat/dissipation you are seeing.
     
  3. Papabravo

    Expert

    Feb 24, 2006
    10,140
    1,790
    Check the datasheet for a parameter Vgs(on). That is the gate-source voltage required for "Full ON". Let us know what you find.
     
  4. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,534
    1,251
    For the type of FET you are using, the resistance from the drain to the source decreases as the voltage from the gate to the source increases, so like an NPN transistor it is an inverting relationship. The gate-to-source voltage Vgs needed for the minimum drain-to-source resistance Rds(on) is on the datasheet. Some FETs are designed so that they reach their minimum Rdson with only 5 volts Vgs, and these are called logic level FETs. Yours is an older part, and the typical Vgs for guaranteed minimum Rdson is 10 V.

    Another possible reason for heating is that the FET is turning on slowly. The gate-to-source capacitance in a power MOSFET is large enough that a typical microcontroller output can't charge it up quickly. The static current is basically 0 mA, but the transient current can be over 1 amp (!), and many amps for larger FETs. If you have a scope, you can see if the uC output slope changes when you connect/disconnect the gate. This probably is not the main contributor to your overheating problem, but it is an aspect of using FETs that should be addressed in good design practice.

    ak
     
    Roderick Young and Papabravo like this.
  5. MrChips

    Moderator

    Oct 2, 2009
    12,440
    3,361
    Should your circuit schematic.

    Make sure you are using N-type MOSFET in common source configuration.
     
  6. clintonb

    Thread Starter Active Member

    Mar 6, 2011
    42
    0
    I've gone through the datasheet again and it appears I over simplified things. I saw the vth min and max and jumped to conclusions and in fact you need 10v to saturate.

    I'm trying to undestand the heat calcs. Would I be correct in saying that effectively the mosfet is exactly like a resistor. If you know the gate voltage and the v(ds) you can lookup a current value which you could easily interpret as an resistance. The current through the circuit will be the input voltage divided by the total resistance that is your load+mosfet resistance. Once you know the current through the mosfet, you can then work out the amount of watts and then lookup the degrees celcius per watt?

    I haven't had an overheating problem yet because the coil rating at 12v is only something like 0.063 amps and the gate voltage I'm using is only a fraction of the saturation voltage.

    I don't see a maximum switching frequency though. Is there a limit to what rate you can switch a mosfet at?
     
  7. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,769
    969
    Dissipation in a mosfet that is "fully on" is I^2 * R where R is the Rdson of the mosfet..

    Theoretical example..
    Rdson = 2 ohms
    Current through mosfet = 5A
    Dissipation is
    (5^2) * 2 = 50 watts

    If junction to ambient thermal resistance is 2 deg C/W
    Then temp rise over ambient is 2 x 50 = 100 deg C
    If ambient is 25 deg C then the junction will be 100+25 = 125 deg C

    Typically there is junction to case or junction to ambient thermal resistances.. One for with and one for without..
    Typically the datasheet will also give some guidelines about dissipation with x amount of PCB copper area,etc...
     
    Papabravo likes this.
  8. crutschow

    Expert

    Mar 14, 2008
    13,006
    3,232
    A MOSFET is intrinsically very fast so the limit is usually how fast you can charge and discharge the large gate capacitance of a power MOSFET. For a desired very fast switching time this can require peak currents of more than an amp, so you would need a high-current high-speed driver for the gate.
    One typical configuration is a complementary emitter-follower with an NPN and PNP connected together at their emitters to drive the gate (NPN collector to V+ and PNP collector to ground).
     
  9. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    Your NFET switching ~6A at 10kHz with whimpy gate drive (on purpose). Green trace is the instantaneous power dissipated in the NFET.

    34.gif

    Dissipation in the NFET attributable to three cases, turning on, turning off, and steady-state on. Slow turn on is due to whimpy gate drive, Gate charge, and Miller capacitance. Slow turn off is due to whimpy gate drive and Gate charge. Dissipation during the ON period is due to only 1V of gate overdrive above the Vth=4V (really needs 10V of gate drive).
     
  10. ian field

    Distinguished Member

    Oct 27, 2012
    4,415
    783
    You can get logic level MOSFETs that avoid this problem.
     
  11. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,534
    2,369
    You need the L in the part number as in IRL520 etc.
    Max.
     
  12. clintonb

    Thread Starter Active Member

    Mar 6, 2011
    42
    0
    I'll be using the logic level mosfet but just for clarity, when you say the gate current when the gate is fully open, is that the maximum current assuming no load.

    If I had the mosfet switching a 12v circuit and there was a 2 ohm resistance on the drain of the N channel mosfet, would that mean that 6 amp would go from drain to source and then the dissipitation would be (6)^2*Rds(on)?
     
  13. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066

    As shown in the simulation I posted, the power dissipation of the non-logic level gate version when driven with 5Vgs was about 2W at 6A.
    That would decrease to 36 * Rds(on) if the published gate drive conditions were met. It would take Vds => 10V to get the published Rds(on).
     
  14. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    Here is another sim showing Rds (Blue trace Ω), Vd (Red trace Volts), and PD (Green trace Watts) as gate voltage Vgs is increased from 4.75V t0 10V for the IRFZ44N. Note that this device was never meant to be used with Vgs=5V.

    188.gif

    The device is just barely turned on with Vgs=5V; the slightest reduction in gate drive would blow-up the FET due to overheating.
     
    Last edited: Mar 26, 2015
  15. ian field

    Distinguished Member

    Oct 27, 2012
    4,415
    783
    You might be getting your terminology mixed up - with DC conditions, the only suggestion of gate current is very tiny for charging and discharging the gate capacitance.

    Gate current is just a very tiny pulse as you switch the MOSFET on and off - it only becomes something you need to consider when you're switching it on and off a few hundred thousand or more times per second.
     
Loading...