Switching noise and ringing effect in single-phase inverter output

Discussion in 'The Projects Forum' started by tchandoo.76@gmail.com, Jul 8, 2011.

  1. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    hi

    I am making a single-phase inverter with a dc-dc boost converter for my project and with an aim to make it to 1 kw rating and 230 V, 50Hz rms using dc power supply in the lab with 175V and 10A supply.

    I followed guidelines given in 6N136, IR2110 and AN-978 , DT97-3 datasheets and kept Rgon=15 Ω and Rgoff=8.2Ω using anti-parallel diode across Rgon and made individual driver circuits for all switches.
    MOSPETS used : STW20NK50Z
    Switching frequency: 10KHz
    PWM pattern: Sine

    Initially i made the inverter circuit and tested with low input dc voltage around 12V and measured output of inverter circuit i found still ringing effect and switching noise more than double the input dc voltage. some screen shots of output waveform are given as attachments.

    kindly let me know where i am doing mistakes.

    Thanks in advance.

    Chandra Shekar
     
    Last edited: Jul 8, 2011
  2. wayneh

    Expert

    Sep 9, 2010
    12,126
    3,048
    Anyone that might help you will need to see a schematic of your circuit.
     
  3. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Yes, a schematic would be helpful.

    How did you mount this whole circuit? On a PCB? On a protoboard? You have to pay attention to parasitic circuit capacitances/inductances.
    If your transistors are not getting too hot, you can still try to increase your gate resistors ( will increase switching losses though). Or you can use a snubber, I had some good experiences with voltage clamps (RDC)

    cheers
     
  4. RRITESH KAKKAR

    Senior Member

    Jun 29, 2010
    2,831
    89
    Can you please tell about , the transformer here you used for DC to AC conversion??
     
  5. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    the circuit is shown as attachment.

    I made separate high side drivers for all MOSFETS (4 no.s) of single-phase inverter. Optocoupler and IR2110 driver portion of the circuit is on PCB and rest inverter portion of switches are mounted on separate board, connection between driver output to gate and source terminals of MOSFETS done by using wires of around 10-15 cms in length.

    I am not using any snubbers. Suggestions on snubbers are welcomed, as i have less knowledge type and value of capacitor , varistor selection.
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Post a photo of your circuit.
    I do not see any bypass capacitors on the IR2110 or your optocoupler.
    You're not showing all of the connections to the IR2110.
     
  7. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    Thanks for the response.

    I didn't have a camera, unable to post photo of circuit, and do not know about bypass capacitors,how they used and purpose of bypass capacitors. I kept bootstrap capacitors with 0.47μF and 2.2 μF for high side and low side of IR2110.
    As my power circuit ground is different from driver circuit ground, so i kept 4 no. separate high side drivers to drive each MOSFET and using only high side part of the IR2110 IC. The attached driver circuit shows only the connections that I used.
     
  8. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488

    Good question! From what I understand , he's is having a sine wave modulated PWM at the input. So it seems he wants to have 230V/50Hz AC at the output. I've not yet seen a DC/AC boost circuit using only one transistor... Is this only the the DC/DC part of his circuit?
     
    RRITESH KAKKAR likes this.
  9. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488

    Ok, I should have read this before I posted my previous message. :D
    4 MOSFETs, so it's really a DC/AC step-up converter. But this shouldn't have anything to do with the ringing. With bypass capacitor he probably meant the capacitor on the output driver circuit of the 2110. This one is to assure that you have enough power to charge your gate as fast as possible. You have only the 0.47uF cap directly at the output, it's from there you take the energy to charge the gate (the inductance of the wires that go to your power supply does not permit a fast current change through it).
    So first of all, 10 to 15cm would be quite long from my experience.the distance from driver to Mosfet, the wires have to be as short as possible. I don't know about using only the high side, but I know that the IR2110 high side can latch up if you reverse bias its output (can happen when you turn it off having a too large inductance (long wire) at the output). Then it stays always ON and you can imagine what that means for your circuit.

    So first: is there a way to shorten these wires? This done we can investigate further.
     
  10. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    I reduced wire lengths about 50%. Could anyone identify the mistake in my driver circuit (individual high side driver provided for each MOSFET) because when i am increasing the input dc voltage, the pwm pulses from microcontroller are suddenly stopped automatically after 60-70V DC.
    My humble request is please read my posts before commenting

    many thanks in advance.
     
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Read this thread: Decoupling or Bypass Capacitors, Why?
    Also, use this calculator: Inductance of a Straight Wire
    A 24 mil (0.024") diameter wire that is 5cm long has over 50nH of inductance.
    Your IR2110 needs to be as close as possible to the MOSFET gate it's driving and the resistors and diode in the gate path need to have their leads as short as possible. The optocoupler output needs to be as close as possible to the IR2110. You should have 0.1uF ceramic or poly metal film decoupling caps across any IC supply/ground pins with leads as short as possible, and there should be a larger aluminum electrolytic cap nearby.
    You should have the optocoupler IR emitter current limiting resistor right next to the optocoupler; you don't show it at all.

    As others have said, that's not really enough, and your capacitor leads are probably too long.
     
    tchandoo.76@gmail.com likes this.
  12. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Maybe I'm mistaken, but isn't that a fullbridge? Your picture is showing that the source of your transistor IS connected to your drivers circuit ground. Or do you mean something else with "driver"? If it's a fullbridge, why not use the 2110 the way it was intended to ? Or am I missing something here?:confused:

    Also, first you where talking about the ringing you want to get rid of, the "the pwm pulses from microcontroller are suddenly stopped automatically after 60-70V DC." looks like a different story. ;)
     
  13. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    Thanks for the info.
    I am in the process of reducing wire length and change of caps. Actually my dc supply and H-bridge inverter are not directly connected, in between dc-dc stage also exists, so my low side inverter switches are not directly connected to dc source circuit due to which i need to go for separate drivers for each MOSFET. I think my driver circuit has some mistake and also i need info on snubber circuit and its placement.
    Thanks in advance.
     
  14. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    I need to test my circuit at higher voltage (300V), in the process suddenly microcontroller stopped (happened 3-4 times)
     
  15. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Well, I'd love to see a complete circuit diagram :rolleyes:

    However, if the microcontroller stops giving out PWM pulses it could be because of noise generated by your power switches. I had this a few times. The microcontroller can be reset by noise (both conducted or radiated). Can you check this? Does your MC have bypass caps? Are there open ports (every port should be terminated). Is it near your switching components? Is the reset input protected against noise? How about PCB layout?

    Also if you want to simplify your design you can also use HCPL3180 which is an isolated gate driver, works very well, of course you'll need 3x15V for a fullbridge, 2 for High side, 1 for Low side, but I guess you already have this. This way you could take out the 2110 and 6N136.;)
     
    tchandoo.76@gmail.com likes this.
  16. RRITESH KAKKAR

    Senior Member

    Jun 29, 2010
    2,831
    89
    Almost all of the UPS uses Tapping transformer like (x-0-x V) Push and pull circuit are very common and simple, using (x-0V) transformer is not a easy task to work with it...!!
     
  17. tchandoo.76@gmail.com

    Thread Starter New Member

    Jul 6, 2011
    7
    0
    i am using dsPIC30f6010.It is near to my inverter board. how to check microcontroller has open ports and bypass caps. Also how to protect microcontroller reset against switching noise? Could you explain me in detail.
    I increased my bootstrap caps to 4.7 and 10 microfarads with ceramic caps (104) in parallel
    Thanks in advance
     
  18. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    First off all the MC should NOT be near to your inverter board. As far as I could see from the datasheet, your mc is a smd, so I guess you didn't design your own PCB. Am I right?

    There should be bypass caps near the power supply pins of the MC. In general there also is a ground plane underneath the MC.

    Can you post a diagramm / layout (maybe photo) of your MC circuit?

    It also seems you have two or three problems (ringing, 2110 defective and MC reset) which maybe or not related. Let's solve first the reset, as it could affect the other issues. It could be possible for example that your MC PWM outputs get into an undefined state if your MC is being reset by noise. So a layout and schematic of the MC part would be interesting. And a layout of the whole circuit (photo)
     
  19. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Just thought I'd share these very simple pairs of circuits and the simulations of them. It's a demonstration of ringing due to fast rise/fall times, the inductance of a trace and capacitance of an input that makes up a resonant series LC circuit, and the 2nd one using a resistor to slow the rise/fall times sufficiently to prevent the ringing.

    You have some very long traces on that circuit board, and some very thin ground paths, particularly for the high power side. The long traces not only have a lot of inductance, but will radiate the ringing signal and cause havoc on the low power side.
     
  20. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    You still didn't show the whole circuit.

    As Wookie said, if one builds a power converter, you have to make absolutely sure that you follow the rules of layout design.

    Apart from the long gate wires I didn't see you mention where your intermediate DC-bus caps are located (you said you have double conversion). Is there any way u can post a schematic and layout?

    The following links provides useful information about current paths in driver and power stages of converters and Do's and Don'ts

    http://www.irf.com/technical-info/appnotes/an-978.pdf
    http://www.maxim-ic.com/app-notes/index.mvp/id/735
     
Loading...