+ Post New Thread
Results 1 to 12 of 12
  1. #1
    Banned
    Points: 8,202, Level: 21

    Join Date
    Jun 2008
    Posts
    1,180
    Helped
    35 / 35
    Points
    8,202
    Level
    21

    flux walking

    hello,

    how do i stop "flux walking" in an offline push pull converter?

    it can't be done with current sense as this would mean the load current being stopped ?

    •   Alt27th November 2008, 20:53

      advertising

        
       

  2. #2
    Advanced Member level 4
    Points: 21,205, Level: 35
    E-design's Avatar
    Join Date
    Jun 2002
    Posts
    1,275
    Helped
    236 / 236
    Points
    21,205
    Level
    35

    push pull smps

    This was a serious problem with using BJT's but not with MOSFETs



  3. #3
    Banned
    Points: 8,202, Level: 21

    Join Date
    Jun 2008
    Posts
    1,180
    Helped
    35 / 35
    Points
    8,202
    Level
    21

    flux walking full bridge

    Thankyou E-design,

    are mosfets ok because they "break-over" happily when the saturation occurs ?



    •   Alt28th November 2008, 20:06

      advertising

        
       

  4. #4
    Advanced Member level 4
    Points: 21,205, Level: 35
    E-design's Avatar
    Join Date
    Jun 2002
    Posts
    1,275
    Helped
    236 / 236
    Points
    21,205
    Level
    35

    flux walking core saturation

    MOSFETs with their Ron subject to a positive tempco will tend to auto balance any flux imbalance. The other fix would be to use current-mode control.



  5. #5
    Banned
    Points: 8,202, Level: 21

    Join Date
    Jun 2008
    Posts
    1,180
    Helped
    35 / 35
    Points
    8,202
    Level
    21

    push pull converter flux walking

    Thanks E-design,

    ref current mode control, i woul dhave thought it would juts end up cutting off the load current ?

    In the primary, the load current will be quite a bit more than the magnetising current, so how can i just cut off when magnetising current is too high.

    when saturation occurs and current runs away i'll be able to cut-off then but by then current will already have risen badly high ?



    •   Alt29th November 2008, 12:14

      advertising

        
       

  6. #6
    Advanced Member level 4
    Points: 21,205, Level: 35
    E-design's Avatar
    Join Date
    Jun 2002
    Posts
    1,275
    Helped
    236 / 236
    Points
    21,205
    Level
    35

    push-pull magnetizing current transformer

    It is done with "pulse by pulse" current limiting. Most modern chips like the SG3525 and after have this feature.

    Why are you so concerned about this problem? If your transformer is properly designed and constructed and driven with MOSFETs this will be a non-issue. The unequal volt/second problem comes from BJTs with their storage time problems. This is not the case with MOSFETs and the Ron with heating will tend to correct the peak current in each leg of the PP transf.



  7. #7
    FvM
    FvM is offline
    Super Moderator
    Points: 168,043, Level: 97
    Awards:
    Helpful gold

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    26,616
    Helped
    8252 / 8252
    Points
    168,043
    Level
    97

    push pull smps transformer calculator

    I see this options:

    - In a voltage controlled push-pull, assume the design is sufficient symmetrical not to need specific measures to handle flux walking. That's, what most designers do, I think.

    - Simulate or hand-calculate the effect of flux walking with your present design parameters rather than making general considerations. Determine if the possible current inbalance is acceptable. (Although the Elementary Electronic Questions forum is a place for general considerations...)

    - Use current mode control

    - Use voltage mode control with an additional current balancing circuit



  8. #8
    Banned
    Points: 8,202, Level: 21

    Join Date
    Jun 2008
    Posts
    1,180
    Helped
    35 / 35
    Points
    8,202
    Level
    21

    push pull switch mode power supply

    hi,

    i have just built and tested a push pull , using current mode with MC34025 pwm controller.

    It was only a 10 W offline converter but the mosfets immediately got really hot.

    I have changed to do a one transistor forward converter now and it is fine.

    i would have preffered push pull as it's less ripple current in the secondary, but those FETs were frying!

    I am convinced that even with current mode, the push pull is no use, as it will not be able to distinguish magnetising current from load current ?



  9. #9
    Advanced Member level 4
    Points: 21,205, Level: 35
    E-design's Avatar
    Join Date
    Jun 2002
    Posts
    1,275
    Helped
    236 / 236
    Points
    21,205
    Level
    35

    what is flux walking?

    You must be doing something very wrong.

    I have designed many P-P converters for industrial and even military applications using MOSFETs up to several 100's Watts and never had a core stepping problem.

    Maybe the problem is with your peak flux calculation for the P-P transformer or the winding and construction used.



  10. #10
    FvM
    FvM is offline
    Super Moderator
    Points: 168,043, Level: 97
    Awards:
    Helpful gold

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    26,616
    Helped
    8252 / 8252
    Points
    168,043
    Level
    97

    resonance converter smps push pull

    I am convinced that even with current mode, the push pull is no use, as it will not be able to distinguish magnetising current from load current ?
    No need to distinguish it. A current mode controller adjusts the duty cycle to achieve equal peak current for both half cycles which should be sufficient to prevent from saturation.



  11. #11
    Banned
    Points: 8,202, Level: 21

    Join Date
    Jun 2008
    Posts
    1,180
    Helped
    35 / 35
    Points
    8,202
    Level
    21

    push-pull smps

    thankyou FvM, i think i see but.....

    The load current referred to the primary will like as not be of significantly larger magnitude than the magnetising current in a push pull.

    -the current sense resistor must not cut off this referred load current, so it will have to be big enough so that it does not cut off this current...this means it will lack the necessary sensitivity to accurately sense the magnetising current in a timely fashion....

    the magnetising current will thus build up to unpleasant levels before it is curtailed ?

    -do you agree with this ?

    I believe half-bridge and full bridge are also using bipolar flux driven transformers......and no engineer would do them without a bridge cap since they fear flux walking..........so i think also the flux walk should be feared in the push pull.........the problem is, there's nowhere in a push pull where a cap can be placed to stop flux walking ?



  12. #12
    FvM
    FvM is offline
    Super Moderator
    Points: 168,043, Level: 97
    Awards:
    Helpful gold

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    26,616
    Helped
    8252 / 8252
    Points
    168,043
    Level
    97

    flux walking in transformer

    -do you agree with this ?
    No.
    no engineer would do them without a bridge cap since they fear flux walking
    Usually, I'll have a DC current sense to cancel unysmmetries. In an unipolar push-pull circuit, as discussed, achieving either equal mean or peak current in both outputs by the controller has the same effect.

    A cap may be used in half bridge or resonant-converter circuits, where it's present anyway.



+ Post New Thread
Please login