VTTC Power Supply Fundamentals - What am I missing?

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Chris Fanjoy" <zappyman@xxxxxxxxxxx>

  Seems I've run into some tough problems with my VTTC power supply. 
I've been using a large MOT, with a simple voltage doubler like this:
----------||-----------+----+------       B+
                        |     |
             ---->|---+     +--||--+
            |                          |
------------+-------------------------+---   GND
  Don't know how well that will display, but anyway....    So far 
I've obtained about 8-9" sparks from the coil, which is driven by a 
pair of 813 tubes. The sparks are very thick and hot, but it seemed 
to me that they should be longer considering the input power. So I 
connected a high-voltage meter (the type used for measuring CRT 
voltage) to the power supply, and was surprised to see a reading of 
only 2kV, even with no load. So I tried a different circuit:
---+----->|------+-----+-------
   |               |      |
   |               C1   C2
   |               |      |
-------------------+      |
   |               |       |
   |               C3   C4
   |               |       |
   +-----|<------+------+-------
  Now, with no load, I measure a good 4kV. But, when the coil is 
connected, it drops to less than half! What's more, the staccato 
controller is seriously messed up, locking into continuous wave when 
the variac is turned above roughly 20V.
  The capacitors used for filtering are all microwave caps, arranged 
in series/parallel combinations as the situation warrants. Both 
circuits I've mentioned use 8 caps in all.
  It seems I'm missing something here, but what? Why can't I keep 
the supply voltage around 4kV? And why does the stupid staccato 
controller keep acting up? The SC works great with the first circuit. 
I should mention that the SC is mounted in an external control box, 
which is connected via a shielded cable to the main unit.
  Thanks for any help.