MC34063 issue. Switches off every cycle

Discussion in 'The Projects Forum' started by karanbanthia, Apr 11, 2015.

  1. karanbanthia

    Thread Starter New Member

    Apr 8, 2015
    5
    0
    Hi All,
    I am using MC34063 in step down configuration but the converter switches off every cycle. Following are the calculations and specifications:
    Vin - 9 to 12V
    Vout - 5V
    Imax - 0.3A
    Rsc - 0.36 ohms
    L - 47uH 1.6A
    Ct - 270pf
    I have tried reducing the current limit resistor value (Rsc), but it didn't help. The usable current is below the designed rated current. I have attached an oscilloscope image showing the output voltage ripple on channel-1 and switching transistor emitter (pin-2 of MC34063A) voltage on channel-2.
    Note that the transistor switches off completely causing a low frequency ripple in the output voltage which is unnecessary.
    I have also attached the parameter and component calculation results.
    Kindly help.
     
  2. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,804
    1,105
    Welcome to AAC!
    Why wouldn't it? It's a switch-mode converter.
     
  3. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    It is called "limit cycling" and is a characteristic of non-linear systems. The system does not have a stable node and the error cannot be reduced to zero. So it has to go a little above the setpoint and then allow the system to go a little below the setpoint.

    I can't tell what the volts per division are on channel 2 which I believe is measuring the output. What is the magnitude of the ripple you are talking about?
     
  4. karanbanthia

    Thread Starter New Member

    Apr 8, 2015
    5
    0
    Hi Alec,
    Switching OFF while changing the duty cycle is fine. That is what it is designed to do. Bu the converter switches OFF completely after fixed interval of time which it shouldn't do since it has to be switching continuously. Kindly check the oscilloscope screenshot which I have attached. On channel 2 is the emitter pin of MC34063A which shows the switching action. The switching freq. measured is around 83KHz and it stops after fixed time interval and then starts again. This is abnormal, isn't it?

    Regards,
    Karan
     
  5. karanbanthia

    Thread Starter New Member

    Apr 8, 2015
    5
    0
    Channel 2 is measuring the emitter pin of MC34063 which shows the switching action, not the output. As per the switching regulator theory, it should be switching continuously, which it is not.
    The final ripple on 5V is around 70mV since it is ridding on the low frequency ON/OFF cycle
     
  6. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    If you're sure there is something untoward going on the best place to start is with the complete schematic. Next is this circuit on a breadboard or a PCB? If it is on a PCB the a blowup of the artwork would be helpful. I don't think 70 mV P-P ripple is bad. What I'm wondering is if the device is experiencing one or more fault conditions and is trying to recover by shutting itself off periodically. Again the limit cycle behavior of fault and recovery may be in play.
     
  7. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    The device has a limit on the peak inductor current. The peak inductor current can exceed the usable output current by a factor of two or more depending on the component selection. Your inductor looks small at 47 uH. Choosing the wrong inductor can cause these parts to misbehave. Layout and and high current paths not properly laid out with an eye to minimum indcutance can also be a problem. Core saturation is also a possibility. High ESR capacitors are also problematical.

    I guess the alternative to going off for a while, is to let your circuit destroy itself, but you probably wouldn't like that too much.

    Where is your schematic?
    Where is your layout?
     
  8. karanbanthia

    Thread Starter New Member

    Apr 8, 2015
    5
    0
    Hi,
    I disconnected the load from the switching regulator and did the following tests:
    1. Connected a 1K resistor as load
    2. Connected a higher load than actual
    In both these cases, there was no change in the switching output. Thus, it is not concerned with the current. Also, the inductor's saturation current is 1.6A.
    Ripple voltage is the next part I am concerned about. The switching regulator should continuously switch, it should never ever turn off, isn't that the theory??
    I am attaching my schematic (calibration resistors have different values here), layout and the actual PCB image.
     
  9. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    No no no. That is not the theory. The regulator may turn off if it detects a high peak current in the inductor. You need to ask yourself why this is happening. It has nothing to do with load current.
     
  10. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    The topology of your schematic matches the example for a step down converter in the datasheet. This is a good thing. The component values are another thing altogether. You have chosen a smaller value of CT, 270 pf in place of 470 pf. You have chosen a smaller value inductor 47 μH in place of 220 μH. Presumably because your switching frequency is higher. Your resistors R61 and R62 are larger by a factor of 2.5 why is this?

    Finally C31 is smaller at 220 μF than the datasheet example at 470 μF. Do you know what the ESR of that capacitor is.

    I'm not real great at reading layouts maybe someone else can spot something.
     
    Last edited: Apr 11, 2015
  11. karanbanthia

    Thread Starter New Member

    Apr 8, 2015
    5
    0
    There is nothing to draw large enough current to turn off the regulator, not even close.
    Also, I have attached my component value calculations sheet in my first post (attachment name: MC34063A 5V calculations.pdf) All the component values are as per these calculations.
    I have also tried changing the component values apart from those calculated in the following ways:
    1. Increasing Ct to 540pf thus reducing switching frequency.
    2. Increasing inductance for reducing ripple current.
    3. reducing feedback resistor to increase peak inductor current.
    4. Increasing output capacitor to reduce output voltage ripple.
    Nothing helped. The regulator still switches OFF and the phenomenon is identically repeatable after fixed time interval. If it would have been due to irregular peak currents, then it would occur once in a while, not repeatedly.
    I don't know the ESR of cap.
     
  12. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    If you have eliminated all the other possible explanations then you are left with the only remaining explanation which is that it is working properly and you don't understand why.

    The output is 5 Volts with 70 mV P-P ripple.

    Q:What is the output of a PWM that is set to zero or 100%
    A:Constant, that is not switching.

    Where is it written that a PWM signal must be beating up and down all the time?
     
  13. crutschow

    Expert

    Mar 14, 2008
    13,042
    3,243
    A hysteretic (bang-bang) regulator, as this appears to be, may intermittently shut off completely when the output voltage gets too high due to a small load current.
    Have you tried it with a higher load current?
     
  14. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,804
    1,105
    Consider what happens once the output cap has charged to the target 5V. What should the converter do then? It can't keep pumping current into the cap or the voltage would continue rising above target; so it stops until the voltage has dropped to just below 5V, then it starts up again. This is the low-frequency cycling you are seeing and is quite normal.
     
Loading...