Transistor gets shorted in Boost Converter

Discussion in 'Power Electronics' started by Xavier Pacheco Paulino, May 10, 2018.

  1. Xavier Pacheco Paulino

    Thread Starter Active Member

    Oct 21, 2015
    543
    45
    Hello,

    I attach an schematic of a Boost converter circuit that I found on an AC motor driver. The driver takes 110VAC from the grid. So I suppose that the boost converter is used to increase the voltage of the DC Bus to control the 220VAC motor.

    The point is that, when the motor has relatively high load, the transistor M1 (which is an IGBT instead of a MOSFET as shown) gets shorted and gate resistor gets open. Sometimes the rectifier bridge also suffers so I have to replace it. The IGBT bridge and recitifier bridge are built-in, that is, all together in a module. The IGBT bridge where the motor is connected doesn't suffer damage. The gate driver is a TC4427. I have replaced the transistor several times, and it still keeps shorting when high load. What could be the problem? It's supposed to stand high loads, so high loads are not the problem.
     
    Last edited: May 10, 2018
  2. danadak

    Well-Known Member

    Mar 10, 2018
    1,459
    297
  3. Marc Sugrue

    New Member

    Jan 19, 2018
    26
    4

    At high loads the peak current through the inductor and transistor will rise, is it possible your saturating the boost inductor putting the transistor under more stress than it was designed for? If it's saturating you are putting a switch directly across thè power which could blow up the switch and the bridge.
     
  4. ebp

    Well-Known Member

    Feb 8, 2018
    1,622
    551
    DO NOT CONNECT AN OSCILLOSCOPE without a clear understanding of EXACTLY what you are doing!

    There is no point in that circuit that is at AC mains ground potential, so there is no point where the ground connection for the scope can be made without use of an isolation transformer for the circuit under test. Differential probes may be used safely, but few scope owners have differential probes. Mediocre results can be obtained by using two channels of the oscilloscope in differential mode, but common mode rejection, particularly at high frequency, is usually quite poor.

    Let me yell that again:
    DO NOT CONNECT AN OSCILLOSCOPE without a clear understanding of EXACTLY what you are doing!

    There are also capacitors charged to quite high and dangerous voltage.

    The circuit is almost certainly an active power factor correction (also called "active harmonic filtering") boost converter that make the input current instantaneously proportional to the input voltage and on a longer time scale inversely proportional to the input voltage. The switch is subject to high current and high voltage. If the timing is off, the switch essentially short-circuits the AC line through the inductor. Since the circuit did work, my suspicion is improper heatsinking of the IGBT. Another possibility is that the inductor is wound on a powdered iron core which as degraded over time due to excessive temperature (the problem is with the "binder" material in that holds the iron powder particles together). The can lead to substantial decrease in inductance and magnetic saturation of the core, which can be catastrophic to the switch. IGBTs are moderately slow, so even though the circuit probably limits current on a cycle-by-cycle basis, if the IGBT can't be switched off fast enough it can be subjected to damaging overcurrent.
     
  5. Xavier Pacheco Paulino

    Thread Starter Active Member

    Oct 21, 2015
    543
    45
    I have a vague thought. What if I connect the main directly to 220V instead of using 110V? Most VFDs require 220V or higher. I guess boost converter wouldn't be needed, right?
     
  6. ebp

    Well-Known Member

    Feb 8, 2018
    1,622
    551
    The boost converter, again assuming it does power factor correction, will still be of value with higher line voltage.

    Without power factor correction, current from the AC line will only flow when the instantaneous line voltage is higher that the voltage on the filter capacitors. Through most of the AC cycle, no input current flows at all, and when it does you get a very high amplitude "pulse" (typically looks sort of like a half of a sine wave at a much higher frequency than the line frequency). The ratio of RMS current to average current is high. Active power factor correction makes the load presented to the AC line look like it is a resistor, so current flows throughout the cycle and in proportion to the instantaneous voltage. The power factor stage also acts as a reasonably good voltage pre-regulator for the rest of the power circuit, maintaining the voltage on the capacitors typically within a few percent, with very little line-frequency ripple voltage. This can be quite advantageous if the rest of the circuit doesn't have sufficient bandwidth to respond well to the ripple that would normally be present.
     
  7. danadak

    Well-Known Member

    Mar 10, 2018
    1,459
    297
  8. Xavier Pacheco Paulino

    Thread Starter Active Member

    Oct 21, 2015
    543
    45
    The module in the circuit is the SKiiP 13NAB065V1, which has 3-phase bridge rectifier + brake chopper + 3-phase bridge inverter. I think that the chopper task is to control the DC Bus voltage? I just want to make sure if it's possible to feed the circuit with 220V and eliminating the boost converter.
     
  9. ebp

    Well-Known Member

    Feb 8, 2018
    1,622
    551
    OK, now I'm totally confused. The module as a 3-phase bridge rectifier, but the circuit shown previously is single phase?

    Yes, you could feed rectified 220 VAC to the the filters, but the ripple voltage will be excessive, the average DC voltage will be low (less than 300 volts versus probably 380 to 400 volts; possibly too low to allow the inverter to operate if it has a low-voltage detection circuit). The deliverable amount of power will be reduced because of the problem with high RMS current from the AC line that comes with simple rectification and capacitive filtering. The stress on the filter capacitors due to high amplitude low-frequency ripple current will probably be significantly worse than with the boost circuit.
     
  10. Xavier Pacheco Paulino

    Thread Starter Active Member

    Oct 21, 2015
    543
    45
    It has a 3-phase bridge rectifier, but it's only using single phase.
     
  11. Xavier Pacheco Paulino

    Thread Starter Active Member

    Oct 21, 2015
    543
    45
    The thing is that I have two AC drivers with the same problem. The heat sink is okay. So I think it shouldn't be the problem. I do not have the scope instrumentation right now to take those measurements. I wish I could find replacement boards.
     
  12. DickCappels

    Moderator

    Aug 21, 2008
    4,868
    1,472
    Some hings to check for:
    Is D1 turning on too slowly and letting a high voltage spike appear on the drain?
    Are D1, C2, and the source of the MOSFET close together so that voltage from fast rising current can be clamped?
    Is the MOSFET rated for more than the current that it is switching?

    I don't see any feedback -how do you know whether the duty cycle is getting to high?
     
  13. ebp

    Well-Known Member

    Feb 8, 2018
    1,622
    551
    I've been assuming that SW1 is a relay contact, and that the resistor it shunts is there to limit peak current when the AC power is applied.

    Is that the case?

    If for any reason SW1 goes open or has excessive resistance during operation, there will not be an adequate discharge path for the boost inductor. This can lead to excessive voltage at the collector of the IGBT. Many semiconductors will fail "instantly", and typically short-circuit, as a result of over voltage.
     
  14. Bordodynov

    Well-Known Member

    May 20, 2015
    1,816
    547
    IGBT transistors turn off slowly, so there will most likely not be an overvoltage, i.e. the transistor does not break through with high voltage. But IGBT transistors, in comparison with MOS transistors (correctly selected), have a greater voltage drop in the on state, and this leads to a greater power dissipated on the IGBT transistor and maybe it overheats and breaks down!
     
  15. ian field

    AAC Fanatic!

    Oct 27, 2012
    6,172
    1,115
    Unless that mains bridge rectifier has a reservoir electrolytic you forgot to draw - its more likely a PFC front end.

    If the transistor gets "punched through" - suspect inductor core saturation.
     
Loading...