+ Post New Thread
Page 4 of 4 FirstFirst ... 2 3 4
Results 61 to 67 of 67
  1. #61
    Advanced Member level 5
    Points: 10,799, Level: 24
    d123's Avatar
    Join Date
    Jun 2015
    Posts
    1,894
    Helped
    431 / 431
    Points
    10,799
    Level
    24

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Hi treez,

    No idea as to the maths of why or if you're asking about a problem or asking to foresee issues. Device to device production variations between parts and batches? PCB routing might need to be extremely identical to avoid minute timing issues that cause a lack of synchronicity between parts andblocks? Blocks or devices competing for current output and it becoming messy and fluctuating? Output impedance too high? Sag on input supply? Drone cooking with the heat generated or emanated waves interfering with other circuit blocks? Joke: Someone pointing out that at that appalling cost there may be a far cheaper way to achieve the same goal with a new design that isn't a load of manufactured devices being misused/stretched in places of the design and looking for a suitable self-designed step-down mega-converter coud be an option?

    What do you think the reasons may be?


    1 members found this post helpful.

  2. #62
    Advanced Member level 5
    Points: 35,463, Level: 46
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,356
    Helped
    539 / 539
    Points
    35,463
    Level
    46

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    I think the reason for mal-function would be the cacophony of noise of all those interconnected converters basically just stopping each other from working.


    1 members found this post helpful.

  3. #63
    Advanced Member level 5
    Points: 14,065, Level: 28

    Join Date
    Aug 2015
    Location
    Melbourne
    Posts
    2,310
    Helped
    846 / 846
    Points
    14,065
    Level
    28

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    standalone converters - when you put a lot in //, need a pretty solid input voltage ( else they will react to that ) and to be a little bit isolated from the main o/p bus ( else they will react to everything on that bus ) hence, it's own Cout, followed by an o/p L, damped ( or R, or both ) followed by the bus caps ...


    1 members found this post helpful.

  4. #64
    Advanced Member level 5
    Points: 35,463, Level: 46
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,356
    Helped
    539 / 539
    Points
    35,463
    Level
    46

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Thanks,
    So, to summarise, we are doing a ~12kW power supply that comprises 42 DCM3623T50M53C2T00 modules in parallel…each one outputting 50V and 6.7A at maximum load.
    The load is motors (BLDC + Inverter).

    Battery in parallel with load
    The load also has a 72Ah Lithium Battery in parallel with it. There is no Battery Management System between Battery and Load or between Battery and Power supply connection. Usually, the load runs directly off the battery with no power supply…But we are adding a power supply so that the load can run for longer…
    The problem we face is that we are not allowed to add a Battery Management System.

    Situation of Battery overcurrent chargeing
    As such, there is a chance that the power supply could start putting more than 90A into the Battery ,which is not permitted. (could overheat it).
    The reason that this could happen is that the DCM3623T50M53C2T00 DCDC modules decrease their Vout with Iout. As such, a lengthy period on say, half load, following a previous lengthy period on full load , would make the power supply vout rise and potentially put more than 90A into the Battery.

    Solution: Hack DCM3623T50M53C2T00 to make it constant Vout
    …We can solve this by buying DCM3623T50M53C2T70 DCDC modules. These have a constant vout over the load current range. However, we have already bought DCM3623T50M53C2T00 DCDC modules, and cannot afford to now buy the ’70 version.
    Therefore, our contractor has elected to “hack” the DCM3623T50M53C2T00 DCDC modules so as to make them behave like the DCM3623T50M53C2T70 modules. (ie, have constant vout vs iout) .This will be done as in the attached. As you can see, a microcontroller constantly reads the load current, and outputs a voltage to the DCM3623T50M53C2T00’s TRIM pin so as to keep the vout at 50V, no matter what is the load current.

    Feedback loops fighting each other
    …The problem with this is that we also have a current clamp connected to the TRIM pin. (This is to stop any module from overly hogging current in the parallel array). The challenge is to set the feedback loop bandwidth of the microcontroller system to a bandwidth such that it doesn’t fight with the feedback loop of the current clamp. There is already an internal feedback loop inside the DCM3623T50M53C2T00 to regulate vout as described. So this is quite a lot of feedback loops we now have, all closed on essentially the same thing….
    Would you agree that we must set the microcontroller’s feedback loop bandwidth to be 10x slower than the current clamps bandwidth? (The current clamp’s feedback loop bandwidth already had to be reduced in order to stop it fighting with the internal feedback loop of the DCM3623T50M53C2T00)


    DCM3623T50M53C2T00 datasheet:
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf

    DCM3623T50M53C2T70 datasheet:
    http://www.vicorpower.com/documents/...53C2y7z_ds.pdf



    •   AltAdvertisement

        
       

  5. #65
    Advanced Member level 5
    Points: 35,463, Level: 46
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,356
    Helped
    539 / 539
    Points
    35,463
    Level
    46

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Hi,
    Do we need to put a Y capacitor from primary to secondary ground of our transformer isolated DCDC module?... BCM4414BG0F4440 DCDC Module

    ….As you know, Its usual to have a Y capacitor between primary and secondary grounds of a transformer isolated SMPS.
    However, supposing you have an isolated DCDC power module, which has internal input capacitors and an internal common mode choke and output capacitors all internal to the module.
    Supposing also that the DCDC module is said to have an “isolation capacitance” of 780pF (that presumably is effectively an internal Y capacitor?).

    So presumably an external Y capacitor is not mandated?

    BCM4414BG0F4440 DCDC Module
    http://www.vicorpower.com/documents/...G0F4440yzz.pdf



  6. #66
    Advanced Member level 4
    Points: 7,104, Level: 20

    Join Date
    Feb 2014
    Posts
    1,006
    Helped
    353 / 353
    Points
    7,104
    Level
    20

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Heh it depends on where that internal C is. If it’s from primary switching nodes to the secondary or vice versa then you might need a multiple of that C to squelch the cm noise.

    Though really with vicors resonant topologies and very high frequencies I haven’t gotten the notion it’s a problem.


    1 members found this post helpful.

  7. #67
    Advanced Member level 5
    Points: 35,463, Level: 46
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,356
    Helped
    539 / 539
    Points
    35,463
    Level
    46

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Thanks, one problem of using an external Y capacitor between primary and secondary grounds is that it effectiveyl shorts out the internal common mode choke inside the BCM4414BG0F4440 DCDC Module.
    Would you agree that the external Y capacitor is probably a bad idea?.....or possibly put in as a no fit to start with?



--[[ ]]--