What's killing my LM358

Discussion in 'The Projects Forum' started by Yuright, May 19, 2011.

  1. Yuright

    Thread Starter New Member

    Apr 3, 2011
    14
    0
    I finally got the bugs out of this circuit and installed it in the machine but after ten minutes of perfect operation I lost my speed control and found the LM358 had failed. I don't know if there is some kind of protection I can add or if my design is flawed and the IC won't survive in this application. I will upload the latest schematic showing values and additional components. Visit my last thread to get history on this project. Note: ignore my original problem with inverted output. It turned out I needed the outputs to go high when inputs were equal. The Motorola LM358N did this for me and was stable. My question now is what steps if any I can take to protect the IC. I should also add that I utilized a relay that will swap the polarity of the feedback pot when forward or reverse is selected. This was necessary to maintain convergence with the signal from the command pot. thanks
    http://forum.allaboutcircuits.com/showthread.php?t=53838
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You don't seem to have a reverse-EMF diode across the K1 relay coil. There will be a very high reverse voltage across the coil when the circuit is opened; it could be high enough to jump whatever you're using to open the circuit (like a switch), and even cause the power on your circuit board to momentarily change polarity, which would zap the opamp.

    Also, if D1 and/or D3 failed open or were not present, your external solenoids could cause the same problem.
     
  3. magnet18

    Senior Member

    Dec 22, 2010
    1,232
    124
    To add to what wookie said, the UF4007 would be a good diode to use.
     
  4. Yuright

    Thread Starter New Member

    Apr 3, 2011
    14
    0
    Thanks guys, D1 and D3 are in place but I am not sure they are the proper application or not. I need to learn more about diode nomenclature. As for the K1 relay, I will have to add one. One other thing I was concerned about was what happens to the output from the feedback pot when it's polarity is changed by the relay. I wondered about that but didn't know what I could do to protect the circuit. Right now the wiper lead from the feedback pot goes directly to the input of the op-amp. thanks

    One other thing is the power supply to the board comes from a 24volt alternator driven by the machines engine which may or may not be putting out a clean supply. What could I do to protect the board if this were a problem?
     
    Last edited: May 20, 2011
  5. simo_x

    Member

    Dec 23, 2010
    200
    6
    Are you referring to ripple voltage coming from the power supply?

    Normally, I put a diode in series at the entry to prevent errors in cable connections which protect the circuit itself, and off course, a flyback diode to protect the circuit's devices from the relay coil back efm
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    The diodes you selected, BYW100-200, should be fine, as they are rated for 1.5A continuous and 200v. The fast recovery feature is a bonus. You could actually use a standard 1N400x diode (where "x"=1 to 7) as fast recovery is not necessary for your application.

    The alternator supply is definitely "new information". They can output high-voltage 'spikes' when "load dumps" occur, like turning off high-power lamps or inductive loads such as electromagnets, motors, etc. The alternator has a regulator, but when a heavy load suddenly turns off, the alternator can't respond quickly enough to keep the voltage exactly where it should be, so there is a momentary surge.

    These spikes can go as high as 60v in a typical 12v automotive system, but probably much higher in your 24v system - which would zap the LM358 instantly, as it's rated for 32v as the absolute maximum supply voltage.

    Since the transient "spikes" are usually of very short duration, you might be able to get by with simply adding a resistor between your 24v supply and pin 8 of the LM358, along with adding a larger capacitor in parallel with C1 of say, 470uF. The LM358 doesn't use much current, and the resistor & capacitor will slow that spike down.

    However, if it's a 24v system, the alternator is really putting out around 28v (27.6v would be optimal) when charging the batteries. That's awfully close to the 32v maximum of the LM358.

    In addition to the series resistor and 470uF cap, you'll need a voltage regulator and yet another cap on the regulator output to supply the opamp. Since we're getting rid of the spikes using the resistor and the 470uF cap, it's not likely that an LM317's limitation of 40v from input to output will be exceeded, so you can use that regulator.

    Have a look at the attached schematic and simulation. I suggest that R1 be rated for at least 1W. I've selected R2 & R3 to give a nominal 20v output from the LM317, which should be just fine for your circuit.

    Note that you will only power the LM358 from the regulated output, not the rest of the circuit.
     
  7. Yuright

    Thread Starter New Member

    Apr 3, 2011
    14
    0
    Thanks for the great info.
     
  8. Yuright

    Thread Starter New Member

    Apr 3, 2011
    14
    0
    I am still unsure about the inputs to the op-amp. Don't they need to be filtered as well?
     
  9. simo_x

    Member

    Dec 23, 2010
    200
    6
    What kind of feedback you will connect to these input?
    Eventually post a scheme..
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    As simo_x mentions, we don't know what you are feeding into your opamp inputs.

    The inputs should not go below ground, nor higher than the supply voltage to the opamp.

    You can use resistors between the opamp inputs and your signal inputs (100k should work well), and use Schottky diodes (like 1N5817's) to clamp the signal between the opamp supply voltage and ground. You will need 1 resistor and 2 diodes per input.

    I've taken the above schematic and added one signal input resistor and a pair of clamping diodes, and attached it to this reply - just so you can see what one clamping circuit might look like.
     
    Last edited: May 26, 2011
Loading...