From: dhky@shaw.ca   
      
   On 13/01/2014 5:15 AM, Salmon Egg wrote:   
   > In article , Don Kelly    
   > wrote   
   >   
   >    
   >   
   >> On 11/01/2014 6:48 PM, Salmon Egg wrote:   
   >>> In article ,   
   >>>    news13  wrote:   
   >>>   
   >>>   
   >>> There is a lot of knowledge about transformers that is not found in   
   >>> electrical machinery books. There is much about insulation, magnetics,   
   >>> hysteresis, etc. What you will find in such books is ways of testing   
   >>> transformers without having to run them at full power.   
   >>>   
   >> These tests can also give a good handle on an equivalent circuit.   
   >> Sure there are assumptions made so that results aren't exact but within   
   >> the limitations of the measuring apparatus, what is exact?   
   >   
   > The point I was trying to make, maybe not all that well, is that   
   > transformer efficiency is not easy to calculate. With efficiency for   
   > good (large) in greater than 95%, it is not all that easy to compute   
   > losses from first principles. In particular, iron losses are nonlinear.   
   > Skin and proximity effects are not that easy to calculate. If you are   
   > trying to push efficiency toward 100%, it is that kind of loss that has   
   > to be taken into account.   
   >   
   > I am not a transformer expert. but expect that manufacturers do not   
   > carry out such calculations based on first principles. They interpolate   
   > from transformers they have already built. If they need to increase   
   > efficiency, they look for loss sources in what they have built to see   
   > where and how they can be reduced. Good record keeping becomes more   
   > important than fancy calculation.   
   >   
   It depends: the following comments are from a low frequency near   
   constant voltage application  as in a power system - where most   
   transformers are involved.   
      
   While record keeping is important, the particular application of the   
   transformer and  the first principles are important. Manufacturers are   
   well aware of this.   
   The application determines the desired load regime and voltage   
   regulation. This by itself implies optimization of the balance between   
   copper and core losses(the core losses are essentially the non-linear   
   iron losses but as the applied voltage is relatively constant, iron loss   
   variation is small-( linearization is quite valid over the range   
   involved )  as well as considerations of insulation (which can be a   
   major factor in HV and EHV applications in that it affects the coil   
   design to get a reasonable distribution of turn to turn voltages).   
   In addition there is a practical (not technical)balance between capital   
   and operating costs.   
      
   I have tested 3KVA transformers with a peak efficiency near 98%   
   (measuring both core and copper losses)but have knowledge of some of the   
     same rating that have a lower peak efficiency that is lower-say 90-95%   
   but are cheap so that, if they fail, it is cheaper to replace than   
   repair.(these were in the days of 2-bit/gallon  gasoline).  On the other   
   hand a >500MVA transformer with an HV winding at 500+Kv will  be a 'one   
   off" design where peak efficiency in excess of 99% and good voltage   
   regulation is desired and implies a big capital cost.   
   Experimental transformers with "glassmetal cores" and superconducting   
   windings have been built -25 years ago-but cost $$$ and at present are   
   not economically viable.   
   Pulse and audio transformers are entirely different animals but again   
   the balance between first principles and economics is involved and in   
   the sizes involved, first principles reign.   
      
   --   
   Don Kelly   
   remove the cross to reply   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)