PI controller, boost converter control circuit.

Discussion in 'The Projects Forum' started by AbhimanyuSingh, Jun 9, 2011.

  1. AbhimanyuSingh

    Thread Starter New Member

    May 22, 2011
    24
    0
    I am attaching the image as well as the .asc file of my ltspice circuit. The image file of the circuit could be very confusing. The problem I am facing here is with the PI controller at the output of the converter. The output voltage is supposed to be around 37 volts, which it is and hence the voltage across R3 is around 2 volts, or a little less than that. But the voltage at the non-inverting pin of the U2 remains in microvolt as current flows and voltage drops down across R8. The value of R8, R4 and C3 were calculated based upon the compensation needed theoretically. But the voltage at the inverting pin of U2 is very very low, in microvolts and the output of the U2 is very high around 200KV. and the rest of the circuit behaves insanely. The output of U2 is vco, which is used further in the multiplier, B1. What stupidity have I done here with the PI controller? Or anywhere else.
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You're using ideal opamps with no DC feedback path, which is why you are getting crazy high and low voltages. I shorted across C3 to make the VCO a voltage follower, and U2 promptly began behaving. If you want the voltages to be more reasonable, you might use some model components instead of ideal opamps, or at least give them feedback to keep them within the desired range.

    Your .model statement for the switch is not correct, which is why you are getting the SW not found error.
    Change it to:
    .model SW SW(Ron=1m Roff=1MEG Vt=5V)
    Note that I used SW for the name of the model, and SW is the type of the model. That is not an error, it follows the requirements of the model syntax as shown in the help file.
    Note that 'm' is for 1/1000, and MEG is for millions. SPICE does not differentiate between upper and lower case.

    The set/reset flip flop model you used only outputs 0v or 1v; it is not high enough to turn on the MOSFET. Even if it DID turn on the MOSFET, it would turn on and stay on until the voltage got high enough, and then turn off.

    R8 and R9 are unnecessary, as there is a path to ground for V1 through the bridge rectifier.

    C2 is far too small. 10,000uF would be more appropriate.
    L1 is only 5uH? At what frequency did you plan on operating? If 100kHz like V2 indicates, you might use 50uH.

    The schematic image you posted is not identical to the .asc file you uploaded. They should be one and the same, or confusion can arise. You changed the values for R2 and R3 in the .asc file.
     
  3. AbhimanyuSingh

    Thread Starter New Member

    May 22, 2011
    24
    0
    Thanks Sgt.
    Yeah I am learning slowly and ur suggestions will speed it up I guess.
    I figured out the the problem of switch and corrected it. But I have a few doubts.
    You shorted across C3. So, effectively there is no C3 then?? I need a PI controller to be there? How is that possible?

    Secondly, without the controller voltage at the left of inductor used to be rectified sine wave, but after I put this control loop, its no more rectified sine. Rather its a DC with a ripple of 100 Hz. The same used to happen in the circuit without control loop when I increased the value of C2 to 1000uF, but thats undesirable. I need it to be rectified sine. What to do??
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I shorted across C3 because that turned U2 into a voltage follower (gain=1) instead of open-loop gain. You need to figure out a circuit that doesn't use open-loop gain, or you will continue to get outrageous voltages out of the ideal opamp.

    If you just want a rectified sine, then leave C2 at 1uF.
     
  5. AbhimanyuSingh

    Thread Starter New Member

    May 22, 2011
    24
    0
    How the <SNIP> is a PI controller designed? Should I compare the input error with the reference voltage using an opamp and then implement PI compensator across another opamp with non-inverting input grounded (shown in figure 1). Or can I implement the PI compensator across the first opamp giving the referenece voltage in the non -inverting input of the fisrt opamp. I dont have patience to find this out. And this is irritating me. Someone please help me out.
     
    Last edited by a moderator: Jun 15, 2011
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You need to clean up your language, calm down, find patience somewhere, and ask nicely.

    Nobody will want to help someone who is irritable and nasty.
     
  7. AbhimanyuSingh

    Thread Starter New Member

    May 22, 2011
    24
    0
    i didn know it will hurt people here....
     
Loading...