VTTC Controller

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "sundog by way of Terry Fritz <twftesla-at-qwest-dot-net>" <sundog-at-timeship-dot-net>

Hi All,

  Grab a cup of joe and a donut, this is a long one.  Please snip un-needed 
portions when replying.

    I love Vacuum Tubes.  They are nearly indestructible in TC duty, 
provided you don't blatantly abuse them.  Therefore, after toasting a fair 
number of FETS in trying to get a SSTC running, I decided to run tubes 
instead.  Here's how I'm going about it.

   Tubes require a high negative voltage on the grid to turn the tubes 
off.  The problem with this, is that tubes are normally "on" devices.  My 
way around that...

   I built a -340v smoothed DC supply, by half-wave rectifying the output 
of a 120/480v 500w utility transformer.  I then put a 4.7kohm resistor on 
the -340v line, and ran the resistor's output to the grid of the tube. (in 
my case, an RS1026, 2kW triode).  I then put a -500v, -7A MOSFET from the 
resistor's output to ground, giving me a simple voltage divider when the 
fet is conducting.

    The FETs are controlled by a modified version of Richie Burnett's SSTC 
controller.  The only differences are that I ran pin 13 to Vref instead of 
ground (for push-pull operation), separated the output pins 9 & 10 (for 
complementary drive signals), and am not using a gate transformer on the 
FETS.  Instead I will be trying a 6A peak FET driver, because the IXYS part 
has a massive miller capacitance.

   Basically, the FET turns on, and the resistive divider shorts to 
ground.  The FET has an on resistance of 1.2ohm, but as it's only moving 
100mA or so, it's *well* below both it's -500v rating, and the 7A cont. 
current limit.  The FET is capable of dissipating 180w, but in all my 
testing, the FET has never gotten above room temperature (unlike the R1 
resistor, which is a 25w power resistor and gets scorching hot!)

   Driving a single tube and FET, with a dummy load in the primary (25 
turns primary, 10 turn work coil with a 60w bulb across it), it performed 
flawlessly.  The tube's switching is crisp and sharp, following *very* 
closely the sawtooth waveform shown on the grid.  -340v on the grid is 
plenty to turn the tube off hard enough to prevent it from conducting any 
significant amount.

   Now, here's where it becomes fun.  There are 3 ways to run this 
circuit.  The first is the easiest, which is single-ended mode running a 
single (or parallel tubes) at Fres. The second is to run 2 or more tubes in 
push-pull, at Fres.  The third is to run the tubes single-ended, with a 
tank capacitor across the primary, and run the tubes at 20khz-60khz, 
depending on tank cap size.

   It works, plain and simple.  No grid leak network, no tank cap (unless 
you want one), and full control over the frequency of the coil when not 
using a tank cap.  Audio modulation, staccato operation (without bothering 
with an SCR in the filament transformer's ground lead), is as easy as 
adding a few 555's.  The total cost is pretty low, with the FETS being the 
most expensive parts at $12 each.  The TL494 controller and driver chips 
are cheap, and depending on wether I can wean the grid's necessary current 
down to scores of mA, a hefty cockcroft-walton multiplier powered from the 
mains and a hefty smoothing/reservoir cap may replace the 120/480 tranny, 
with a higher value for R1.

   Now, the good news.  I've looked at most tubes that are used for 
coiling, and it seems the vast majority of them will operate just fine on 
-340v grid supply.  (833's, TL304's, etc).   The FET should never see much 
harsh kickback with the customary RL choke on the grid (at the tube!), and 
the FET is isolated from the HV except for the capacitance of the 
plate/grid (a few pF).  Protecting the grid leads from the controller from 
RF will be more of a task, and I intend to only use a swamping resistor and 
RG-59, with the braid grounded at one end.

   Tubes handle overvoltages and current spikes much better than FETS, they 
are usually much faster to turn on and off (though we're limited by the 
speed of the P channel fets in the divider, ~100nsec in my case).... The 
sharper the gate drive signal, the sharper the FET switching, and even slow 
tubes (833's) can operate up to 20-30mhz, so they'll have *no* trouble 
following the FET's switching.    A major plus is that, barring a FET 
failure from RF or hookup errors, the tubes are constantly biased at -340v, 
or they're "off".  Makes for a much happier tube and less power drawn from 
the plate supply.

   It's a drop-in replacement for your existing grid-leak network (in which 
case you can probably use a 556 for staccato and Fres, controlling a single 
FET driver and FET), or you can run your tube(s) at Fres.

   Maybe not earth-shattering, but it's certainly interesting to me.  The 
real test of this system will be when I hook it up to a GU10A, 10kW tube 
(James, this is gonna rock!), to see if it will control it.  But, the fact 
that my "test" coil is a 4kW push-pull setup, I can see how this will let 
the solid-state controlled VTTC (SSVTTC?) really give us radical power 
levels with easy control of the tubes.
   And yes, I've tested it.  It works.  Still some bugs to work out, (FET 
driver is the problem at the moment), but driving the FET via my signal 
generator worked perfectly in controlling the tube.  It's just a matter of 
getting good gate signals.  From there, your imagination is the 
limit.  Audio modulation, staccato, etc....

   Comments, flames, etc welcome!
Shad

-----------------------------------------------------------------------
Shad (Sundog)
G-5 #1203
"Ever stop to think, and forget to start again?"
-----------------------------------------------------------------------