MOSFET switching problem

Discussion in 'The Projects Forum' started by jobertgalea, Dec 29, 2008.

  1. jobertgalea

    Thread Starter Member

    Nov 27, 2008
    11
    0
    Hi,

    I have a problem with the attached circuit. When simulating this circuit on the Proteus software everything works perfectly well. But when I tested the real circuit, I had problems with the switching part using the MOSFET. When it is on and I toggle the switches linked to the MOSFET, the voltages of the whole circuit vary. Even the ground varies when it is supposed to be 0V, it rises to about 0.2V. I cannot understand what is the problem as the simulation worked perfect. Can anyone help me cause I am desperate for a solution as I have a deadline approaching.

    Thanks
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    How is your circuit constructed? What reference do you use to base your measurements from? What power source do you use? Do you have plus and minus 12 volts, or only +12?

    Is it correct to say that the motor runs in one direction at full speed with the left switch closed, and in the other direction at a controlled speed when the right switch is closed?

    Are the op amps 741's, and if so, can you use something better? you might alos find it helpful to place a small resistance in series with the gate of the IRF530.
     
  3. duffy

    Active Member

    Dec 29, 2008
    44
    0
    Ugh, people still use 741's? You need filter caps in that circuit.

    Relative to what? There's your first clue.
     
  4. jobertgalea

    Thread Starter Member

    Nov 27, 2008
    11
    0
    I use a normal power supply and so having + and - and ground. I only need the mosfet to be switched on or off and there is no speed control. Thus I am supplying the mosfet with a +ve voltage signal higher than the threshold to have it switched on and a -ve supply to have it switched off.

    Than what opamp do you suggest me to use instead the 741s?

    And what are filter caps?
     
  5. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    This may be part of your problem. The conduction threshold voltage and the full conduction voltage are very different animals in FET's. If all you want to do is run the motor, what is the circuit for? You could do the same with one DPDT switch with center off.

    To place a FET into full conduction, the gate voltage must be 10 volts or better above the source terminal voltage. 741 op amps can't quite manage to swing the output to within 2 volts of the rail, so you can't get the FET turned all the way on.

    The 741 was designed over 35 years ago. By modern standards, it has a very limited output swing, poor frequency response, and low input impedance. look at spec's for something more modern, like a TL071, an LT1113, or an OPA134.
     
  6. duffy

    Active Member

    Dec 29, 2008
    44
    0
    An MC33202 is a better opamp, and it will go rail-to-rail.

    Once again, relative to what?

    A filter cap for that would be like 220uF or so to filter the power, and maybe a .1uf to get high frequency noise the big one will miss, they go across the supply rails near the op-amps. Another .1uf filter cap would go across the motor to reduce the brush noise. You are going to get all kinds of spikes off that motor, even if it's the size of your thumb.

    With a bigger motor, and considering those are reversing switches, you could also get considerable I*R drops from the load, and that would explain the ground bounce.

    I'm not wild about sending that gate to the negative rail, or relying on the internal diode to protect Q1 from the inductive spikes, either.

    But start with that .2V ground bounce. Like I said, it's your first clue. Look at how you are measuring it. Use ohm's law to calculate how much current it takes to generate .2V from one grounded part of the circuit to the next.
     
  7. jobertgalea

    Thread Starter Member

    Nov 27, 2008
    11
    0
    With reference to the ground problem, the increase in voltage is measured at the resistors where they are connected to ground with reference to the supply ground. To enhance the information, also when pressing the switches, the voltages at the output of the opamps drop down at about 0.1V also when they are supposed to be 1/2V

    And yes I am supplying the mosfet with a voltage bouncing from about +11V to -11V. It is like a digital signal. And I need this type of circuit as it is a control system and a sensor is used. The difference between the reference signal and the signal from the sensor give the amplification needed

    Thank you for the suggestion of using those type of opamps.

    But can you explain what will the filter caps improve in the circuit please?
     
  8. Alberto

    Active Member

    Nov 7, 2008
    169
    36
    They will remove part of the disturbance introduced by the dc brushes motor when running.

    You should also place a decoupling capacitor of 10 nF on the Vdd pin of every OP in your design (both - & +) to improve noise rejection of your system.

    Alberto
     
  9. jobertgalea

    Thread Starter Member

    Nov 27, 2008
    11
    0
    Thank you very much. Thus to conclude I need capacitors in series with every supply of the opamp and capacitors across the supply rails and across the motor. Confirm?

    And finally, do you think that buffers will help if introduced in the circuit to maintain the voltages? As it seems that the motor is taking all the load from the circuit.

    Thanks
     
  10. Alberto

    Active Member

    Nov 7, 2008
    169
    36
    10 nF capacitors not in series to supply, but from Vdd pin and ground.!
    The 220 uF electrolitic capacitor, suggested in post # 6, will act as a buffer. I will use a larger one min 1000 uF better 2200 uF.

    Alberto
     
  11. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    I think the MOS is destroyed because you haven't use a fast recovery diode in parallel with the motor to absorb the back EMF when the motor is switched off. That is why you don't observe any speed control.
     
Loading...