glitch in inverter output

Discussion in 'General Electronics Chat' started by rocon123, Feb 15, 2016.

  1. rocon123

    Thread Starter New Member

    Jun 26, 2015
    27
    0
    I want to generate AC from DC(an inverter). I am giving Bipolar SPWM(sinusoidal PWM) to H-Bridge. At output I am getting AC signal but with noise. I have attached required screen shots. Hardware.JPG
    schematic of H-bridge.
    IMG_20160205_203719.jpg
    Output between Q2 and Q4 emmiter terminal. I2.jpg
    Output between CN1 and CN2. I4.jpg
    Zoomin output.
    Please help in resolving this issue.
     
  2. tranzz4md

    Member

    Apr 10, 2015
    139
    28
    Doesn't look that bad from what I can see. Your test load matters of course, and the intended load more so. Are you at all familiar with older "solid state drives" or the inverters and VFDs of 10-20 years back? There are plenty of papers out there regarding line and motor heating and losses from noise and waveform distortion. You'll never eliminate either as long as you're dealing with high speed solid state switching.
     
  3. rocon123

    Thread Starter New Member

    Jun 26, 2015
    27
    0
    There must be some way to overcome this.
     
  4. crutschow

    Expert

    Mar 14, 2008
    13,003
    3,232
    What problem does this noise cause?
     
  5. rocon123

    Thread Starter New Member

    Jun 26, 2015
    27
    0
    that noise is not acceptable. It should pure alternating signal. Is this due to ringing effect of IGBT?
     
  6. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,791
    1,103
    Could be ringing due to your scope probe inductance?
     
  7. rocon123

    Thread Starter New Member

    Jun 26, 2015
    27
    0
    No..it is not probe because same frequency pulses at IGBT inputs do not have that effect.
     
  8. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,532
    1,249
    No, it shouldn't.

    This is a switching power supply. Even with a high-Q resonant trap on the output there still will be some components of the switching frequencies. Note in the second scope shot that the noise increases at the sinewave peaks. One possible explanation for this is that the high frequency components of the very high and very low duty cycles are more than the simple output filter can handle. The output filter has a corner frequency of 3.4 kHz, a couple of octaves below the PWM carrier but *way* above the desired output frequency of 50 Hz. Also, the filter's high frequency performance is dependent on the Q of the sole inductor and the impedance of the load on the output. Why is there no inductor between Q2/Q3 and C8?

    Increase the output filter components such that the corner frequency is closer to the output frequency. The closer it is, the cleaner the output. BUT, the closer it is, the more some of the output energy is dissipated in the filter components. I'd say to start with 300 Hz and see if that is better. Also, use a combination of high frequency capacitors, not a single fat electrolytic.

    But before you jump into that, know that your feedback is taken after the filter, so group delay and other phase shifts through the filter will affect the feedback loop, requiring changes to the loop compensation. One solution for this is to leave the circuit as is, and add a second output filter after the voltage feedback pickoff point, between RV1 and F1.

    ak
     
  9. rocon123

    Thread Starter New Member

    Jun 26, 2015
    27
    0
    HI,AK

    Presence of switching frequency component in output of inverter was expected. So decide to modify output filter capacitor according to that noise which is coming.
    But one observation we made is, when ever we vary the load conditions, or by varying output voltage (in open loop) we expected some proportional variation in ringing as we thought it is because of switching of IGBT. But it is not the case. Amplitude and frequency of oscillation remained constant throughout. We tried with different DCO's, different probes, oscillation is constant. Then we monitored voltage across Collector-Emitter of IGBT. To our surprise there are no switching transients.
    So above analysis and observation eliminated predictions of oscillation due to switching or faulty DSC and probe.
    By careful selection of capacitor on output side we were able to filter the output. But our point of interest is the source of ringing and reduce that there it self.
    Because connecting capacitor on output side will not ensure proper operation at all the times. Which makes ineffective.
     
  10. crutschow

    Expert

    Mar 14, 2008
    13,003
    3,232
    That doesn't answer my question.
    What is the application that needs a "pure alternating signal"?
    Why does it need to be pure and how pure should it be?
    Most AC powered devices are not that fussy about the purity of the sinewave.
     
  11. rocon123

    Thread Starter New Member

    Jun 26, 2015
    27
    0
    Calibration devices
     
Loading...