RadioBanter - View Single Post - QUESTION: Fun with Svetlanas or Staying alive with kV power supplies

July 14th 05, 03:34 AM

wrote:
My knowledge of vacuum tubes and kV power supplies is limited. I have
been reading an article in the ARRL Handbook detailing the construction
of a 1kW HF Linear. I'd like to try my hand at building something like
this. I found the article a little intimidating: Ceramic insulators,
parasitic suppressors, thermal and mechanical engineering etc. Is
there some book that details this type of thing with an explanation of
the whys as well as the whats and hows. My priorities a

1) Safety. I'd like to be alive to make my first 1kW QSO
2) Avoiding equipment destruction, arc overs, black smoke, explosions
etc
3) Safetly troubleshooting this kind of equipment, loading testing
etc.
4) How to deal with tubes: warm up, care, etc...
5) Avoiding TVI (ITV), parasitic oscillations etc.
6) Longevity and Duty Cycle issues etc.
7) Costs and sources.

Hope someone can help.

Thanks,

Tim

I built a homebrew KW about 15 years ago using a pair of 4-400 in
grounded grid. It worked fine on 80-20 but I never did get the tank
Q adjusted well enough for good output above that. Not that it couldn't
have been done with a bit more tinkering.

The power supply put out about 4000 volts no load. I used a time delay
circuit with a relay and a power resistor to limit the in-rush current
while the capacitor filter bank charged. It had a relay whose ac coil
was across the primary of the plate transformer that shorted out a power
resistor in series with the primary. As the caps charged and the input
current fell so did the voltage drop across the power resistor until the
primary voltage rose to the relay's pull in voltage shorting out the
resistor. (Idea from ARRL HB).

A second time delay circuit did the same thing for the filaments of the
4-400's. This relay also locked out a second relay that applied power
to the plate transformer. Result, the plate supply couldn't be turned
on until the 4-400 filaments were warm, that delay was about 1-2 seconds
during which a power resistor in the filament transformer primary
limited the 4-400 filaments to half voltage.

The filter caps were 10 330uf 450v computer grade units in series, each
cap bridged by a 50k 25W power resistor to even the voltage across the
caps and act as a bleeder resistor. A 1ma meter with a suitable
multiplier resistor in series served as a voltmeter across the
capacitors (I forget how many meg ohm it was). The meter case was
thick enough plastic to be insulated enough from the chassis (and
and besides the main plumbing was behind several meg ohms). The meter
read 5000v full scale.

I always assumed there was voltage across the caps when I worked on the
rig. First pull the plug(s). Wait till the voltmeter drops to zero.
THEN put a heavy screwdriver with a well insulated handle from ground
to the HV terminal to be sure! (I used to do that to picture tubes
when working on tv sets). Keep on hand in the pocket if you need to
adjust anything when it's hot. BTW an RF burn can be more deadly than
a DC jolt. (Think microwave oven).

HV isn't the only thing that can get you. High current can give you
a nasty surprise. Like the guy replacing some batteries on a golf cart.
He was using a ratchet wrench to tighten the battery clamps and he
ended on the most positive battery terminal. The handle of the wrench
hit the chassis of the golf cart putting it between 36 volts and ground
of some VERY HEAVY DUTY batteries. Ever see a Sears ratchet wrench
glow WHITE HOT? (and melt?)