🚀 go-pugleaf

In article <1990Oct7.214431.25748@athena.mit.edu> chuck@mitlns.mit.edu writes:

   In article <1990Oct6.220317.18337@mthvax.cs.miami.edu>, wb8foz@mthvax.cs.miami.edu (David Lesher) writes...
   > 
   >{plan to rectify ac-->dc with cap. filter}
   > 
   >>  This is a BAD idea. If you size the capacitor large enough to get
   >>rid of the ripple you will be running the blanket at the AC peak
   >>voltage not 110. This will mean it will be running at about twice


>       Not at all. Unless there is some current limiting in the
>   supply, such as winding resitance to limit the current, then
>   you will get 20 or 30 amps at the peak to charge the cap and
 
>      (sqrt2*V0)*(sqrt(2)*I0)= 2*V0*I0 == twice the power.
>
>					   Chuck@mitlns.mit.OBedu

Rectifying line AC and putting a capacitor in parallel will in fact 
increase the average power delivered to a resistive load. For any size
of capacitor and resistive load, it's easy enough to calculate the
power delivered by integrating the (square of the) voltage waveform.
The area above the sine wave is the extra boost. 

          ..          .. 
       .     .*    .     .*
     .        . * *        . *
   .           . .          .   *
  ._____________.____________.

      ... is full wave rectified waveform, without any cap
      *** is with a capacitor (not very large compared to load resistance)

About seven years ago, back in India, I used this principle to make a
power booster for a photographic enlarger lamp. Line regulation was lousy,
and the bulb would frequently be too dim. Rather than spend the bucks for
a real autotransformer based regulator, I rigged up a bridge with 4 * 1N4001's
and a salvaged capacitor (100 u or so, from an old valve radio, probably), 
this weird looking DC waveform went through a series transistor and some 
feedback circuit to provide a reasonably constant brightness. 

(It never worked too well, by the way, I must have omitted something. The 
series transistor kept blowing; and that got expensive :-)

/ Jon Sreekanth