Re: 304tl experiments...

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Shad Henderson <shenderson@xxxxxxxxxxxxxxxx>

Hi All,

Before I got pulled in another direction, I was running a grid-driven
VTTC on a very small scale, using a TV sweep tube as the main power
tube.

I had problems with the circuit, but have since realized what the
trouble was, and how to fix it.

In a nutshell, here's what I did, and how.


I used a control transformer (120:480) power supply into a full bridge
rectifier and smoothed it with a 2uF 4kV cap.  That was the plate
supply.

A feedback coil (about 2-3 turns of 20ga wound below the primary)ran
into a comparator, which drove a transistor (inverter) and a pair of
TC4420's (needed an inverted input to one of them).  A 556 provided BPS
(~200-1000pps) and "firing time" per pulse, from 70-300uS.

The TC4420's drove a small half bridge of IRF740's, which in turn drove
a ferrite toroid [MISTAKE HERE!] that in turn drove the grid of the
tube.  I had a 1:3 step up ratio on the grid-drive toroidal ferrite, and
a single secondary on it.  I connected one end of the secondary to a
-50v supply, and the other to the grid.  Careful manipulation of the
input voltage to the half bridge allowed me to swing the grid voltage
very well.  Plenty of drive for grid current too.  Beware you don't cook
the grid, though.

The mistake was using a toroidal ferrite core to drive the grid. The
heavy grid current would walk it into saturation very quickly at high
BPS and long "on" times, leading to the waveform going to crap and the
IRF740's popping.

The "fix" is to use a gapped EE ferrite core in place of the toroidal
core that the grid drive transformer is wound on. That'll prevent
saturation, and keep the IRF740's happy.

It sure beat the heck out of using a piece of sillycon to control the
cathode, because by the time I got all that done, I realized I was using
the tube as a big high voltage resistor in series between the B+ supply
and the mosfet.  Disconnecting the cathode from ground just whizzed it
up to not much below the B+ voltage, necessitating a high voltage
mosfet.  The thought of using a tube as a high powered resistor that
consumed lots of power just struck me as, well, silly.  But that's what
it was in that setup.  I was taking *no* advantage of the tube by
switching it's cathode at Fres, and still beating the mosfet silly by
asking it to interrupt current flow (switching losses...).  Much easier
to switch the tube's grid, and eliminate switching losses on the
high-powered end all together.  After all, a tube doesn't care if you
switch it in the middle of a high current cycle.  It just does it.

Once I get my current project off my plate, I'll be fixing the hybrid
VTTC setup and seeing how much I can abuse the little sweep tube.  The
setup is very close to a traditional VTTC, only a slightly bigger tank
cap (I used ~5nF), less primary turns, and the feedback winding has
*far* less turns on it, and less coupling to the primary.  You only need
a few volts feedback.  Be sure to clamp the feedback coil with zeners,
though.  1W or better.

I wanted to spring this on the community as a surprise, with pictures
and a webpage, but since people are contemplating similar ideas, I
figured I'd speak up to share-and-share-alike.  Maybe save others the
hassle of repeating my mistake.

Thanks!

Shad H.




On Tue, 2007-03-27 at 00:09 -0600, Tesla list wrote:
> Original poster: "Steve Ward" <steve.ward@xxxxxxxxx>
>
> Hi Brett,
>
>
> >  It certainly won't be a
> >matter of just plugging in a standary SCR with a 555
> >oscillator on the gate.  I think one of the 1000v 600a
> >IGBT bricks would be sufficient, probably overkill for
> >that duty.  Plus, the driver circuitry wouldn't be any
> >more sophisticated than the usual staccato.
>
> Such an IGBT is completely wrong for this use.  It only needs to run a
> few amps, but possibly very high voltages.  The proper way of
> controlling this would be control of the grid.  Let me give it some
> more thought later...
>
> Steve
>
>
>
>