(  Just for my own purposes, I'm just clarifying my idea again  )

   (  Note : That by using things like graphene and the recently invented very ultralight gels/foams made of metal( or even modified  IC/semiconductor memory chips ),  in order to use less power to get hydrogen from water,  the efficiency of electrolysis is increased.   
     I recently discovered that people have already used these materials to increase the efficiency of electrolysis,  although I don't know if anyone ever tried  modified  IC/semiconductor memory chips  )

    In my idea,  of a  'purely-hypothetical'  device,  with  'no energy losses in the system( no energy leaving the device )'

       -  First,  use  25-Mj  to create 100-cubic-ml-of-hydrogen, 
       -  Then,  use .001-Mj to spark the full amount of 100-cubic-ml-of-hydrogen all in one go( or can  platinum  spark the hydrogen  ?  )

      My idea is that the 100-cubic-ml-of-hydrogen produced,  could contain 'More'  Mj,   than the 25-Mj used to produce it,    and that when the 100-cubic-ml-of-hydrogen  is ignited in one go,  then more than  25-Mj could be recovered as electricity. 
      Or,   that the  burning cloud of '100-cubic-ml-of-hydrogen' could somehow be used to create more than 100-cubic-ml-of-hydrogen.

      So,  with each new repetition of that cycle,  you are producing more electricity and or hydrogen,  and that in turn allows you to produce more hydrogen and or electricity .

     This could be a device that repeats this process for say  4-cycles( maybe each cycle having it's own container of water and combustion chamber ),  before it returns to the beginning.

    So,  in a real device,  any energy-losses( energy leaving the device ) are offset by this effect of constantly producing more electricity and or hydrogen in each cycle .