MOSFET overheating in audio LED controller circuit

Discussion in 'The Projects Forum' started by R.Gregory, Mar 6, 2017.

  1. R.Gregory

    Thread Starter New Member

    Feb 24, 2017
    2
    0
    Greetings,

    I built a circuit which uses line-in audio to control a white LED strip using a MOSFET. the MOSFET overheats when the LEDs are illuminating the brightest i.e. when the music is the loudest. Is it possible the LEDs are drawing too much power? i took the original design form this post:

    https://www.allaboutcircuits.com/projects/make-an-analog-sound-reactive-led-strip/

    I removed the filter right after the input and also removed the switches that control multiple colors as i'm controlling a single color strip. i also swapped out the 47k resistor for a 100k pot.

    I'm driving this strip:
    118", 9.5 watts per foot, 12vdc
    http://www.ecolocityled.com/product/supreme_white_led_strip/led_strip_white

    with this power supply:
    output: 12v, 100w
    https://www.amazon.com/gp/product/B007K4XZPG/ref=od_aui_detailpages00?ie=UTF8&psc=1


    Any tips are welcome!
    Thanks
    Gregory
     
  2. #12

    Expert

    Nov 30, 2010
    18,076
    9,681
    1) Removing the input filter is wrong. Humans can't see an LED flashing at more than 30 times per second. Allowing the audio signal to blink the LED 10 or 20 thousand times per second accomplishes nothing but making the mosfet spend half its time in transition between on and off, which is when the heat builds up.
    2) That 100 ohm resistor from gate to ground is wrong. The op-amp can output V+ minus 1.5 volts. That puts 3.5 volts on the mosfet gate and that seems OK but it also calls for 3.5V/100 ohms = 34 ma.
    Look at page 5, bottom left graph and see that the opamp is losing about 6 volts while trying to deliver 35 ma and it only has 5 volts to start with! Therefore, your gate drive is being wrecked by the 100 ohm gate resistor. Change it to at least 1000 ohms or 2200 ohms.
     
  3. #12

    Expert

    Nov 30, 2010
    18,076
    9,681
    Holey Smoke! The Orange Team allowed that to be published on AllAboutCircuits?:confused:
    Without checking to see if it would work or why he specified a Military Grade amplifier chip?
    I hope they didn't pay him much.:mad:
     
    Sensacell likes this.
  4. Sensacell

    Moderator

    Jun 19, 2012
    2,305
    736
    Totally cringe-worthy.

    A $10.00 opamp to flash some LED's.

    The schematics look like they were drawn by an amphetamine-fueled squirrel.

    Another brilliant bit of AAC editorial.
     
    Sinus23, dl324, RamaD and 1 other person like this.
  5. RamaD

    Senior Member

    Dec 4, 2009
    315
    46
    Of late, LM124 & LM158 in SOIC and TSSOP packages are available under 1$.
     
    #12 likes this.
  6. #12

    Expert

    Nov 30, 2010
    18,076
    9,681
    58 cents at www.mouser.com
    Full Military temperature spec.
     
  7. R.Gregory

    Thread Starter New Member

    Feb 24, 2017
    2
    0
    Much thanks for the quick thorough response. I've swapped the 100k resistor for a 2.2k and the MOSFET shows little to no signs of heating up. The LEDs are responding just fine.

    I am using a graphic EQ before the circuit to have more control over what frequencies are allowed. Do you think the original filter within in the schematic is still necessary?

    The big question i'm wondering is after the exchange of the resistor is this a worthwhile design or should i start from the drawingboard? Ideally i'd like a system that i can keep in an enclosure safely and use without concern.

    Thanks
    R
     
  8. #12

    Expert

    Nov 30, 2010
    18,076
    9,681
    My EQ only goes down to 31.5 Hz. It sure would look funny with one knob up and nine knobs down!:D
    Besides, it's expensive and still too much bandwidth to use a graphic equalizer for each LED blinker so I wouldn't do that when I can design the EQ out with a resistor and a capacitor.
    There are a lot of things I wouldn't do besides designing it so it won't work, like using a Mil Spec chip, a dual amp when you only need one, a 20 volt mosfet in a 12 volt circuit, or a 100 AMP mosfet to control 100 WATTS...unless its way cheap.

    I did a slightly improved version by eliminating the 5V regulator, changing the input impedance to make the capacitor smaller (cheaper) and lowering the feedback impedance so it doubles as a gate resistor for the mosfet. If I was getting paid, like the guy who did the original drawing, I would change the op-amp to a higher impedance version to further decrease the size of the input capacitor and reconsider throwing away 90% of the input signal, only to put it back in with op-amp gain. color organ.png

    So, yeah, start over. Wrong amplifier chip, wrong mosfet, wrong input impedance, wrong feedback impedance. The only thing he got right was the frequency range and you immediately peed in that.:eek: Bad beginner!:D Worse designer.:( Maybe you can live with the input filter now that I've improved it down to a way smaller capacitor.;)
     
    Last edited: Mar 6, 2017
    Alec_t likes this.
Loading...