New, corrected, Sweet (and sour) spot tests, etc.

 Malcolm, John C, Bart, Richie, all,
 **** All my recent tests were in error.  There was a short in the
       secondary coil that was affecting the results (I think).  I first
       tried installing a 3" by 10" toroid and running the coil at 280
       watts to see if the spark increased (which would mean that
       the 4" by 13" toroid was too large).  It seems that the spark 
       did increase a little, but I noticed arcing on the secondary.
       I knew the secondary had burned spots, but I didn't think it
       was affecting the results.  But I think it may have.  Today it
       was very humid, and this may have been increasing the spark
       lengths too, I don't know.  Anyhow, I removed the burned turns
       and repaired the secondary and redid some tests (below), and 
       got results more in line with what I originally expected.  I'm 
       sorry about the confusion.******

 New tests:

 In these tests, I ran the TC at various power levels, and with four
 different toroid sizes, to see how far the spark length would vary
 from the square law predicted lengths.  Spark length shortfalls from
 predicted lengths might mean that the coil is becoming inefficient in
 some cases.
 All tests at 120 bps.  42" is a reference max spark length.  (This 
 table can be more easily analyzed if it's printed out on paper, so
 the headings don't scroll out of view as you scroll down.)
                                   pri   predicted
 Watts    cap W  Top"   tap    length       length actual    shortfall%
   280       210     3x10   18       29"               27"                 7
   280       210     4x13   19       29                27                  7
     (note above: the smooth toroid size no longer mattered)
   400       318     4x13    20      35                 35                 0
   530       440     4x13    21      42                 42                 0
   620       498     4x13    21      44.5              45                  0
      (note: now TC efficiency is constant from 35" to 45" spark and
       with same toroid)
   280       210     4x17    20       29                24                 13
   400       318     4x17    20       36                34                  
   530       440     4x17    21       42                40                  5
      (note above how the 4x17" toroid is now inefficient, it may be
       emitting corona from its corregated surface.  The smooth spun
       4x13" toroid seems better despite being smaller.)

  Here's a re-test of the test of Malcolm's suggestion of using a large
  toroid to hold the voltage constant when increasing the power to see
  if the spark gets longer:
  280      210       3x10"    18     29                 27                7
  560      440       6x26"    26     42                 38                10 
  These results show that by doubling the power input, and using
  a rather large toroid to attempt to keep the output voltage constant,
  the sparks still increased by 41%, even more than in the original
  test.  This suggests that streamer current is very important too
  for creating long sparks.  The sparks were very hot and strong 
  using this large toroid, BTW. 

  Then I tried higher power:
  750                  6x26     26      48                  47                
   (here, the TC can be seen to be remaining on the square law curve,
    the 2% shortfall is within measurement error)

  Next, I tried the 4x13 toroid at 750 watts: 
  750                                        48  (spark did not reach 48")    
      (I will do more tests of this) 
  In any case, the results above show that:

 1.  Just ramping up the power (or maybe the break rate too) may
      give false results; the coil has to retuned for each power level
      (nothing new here).
 2.  The toroid size may be less critical than the last tests suggested,
      it's still very important though.  Smooth toroids seem much better,
      in some cases, and the smooth toroid easily produces just one
      streamer when a bump is installed, even when the sparks are
      3.4 times the toroid diameter.
 3.  Sour spots occured now only at very low power levels.  The 
      square law held, from 400 watts to 750 watts.  I am pleased
      with the results of these new tests which fit my original 
      expectations, and added new data.
 4.  Thus, when making comparisons of 120 and 240bps, it may be
      acceptable to let the spark length vary, as long as the TC is
      known to be "on the curve", from tests such as the above.  I
      will re-do the 120 vs. 240 bps comparison tests. 

 John Freau