Amps
[Top] [All Lists]

[AMPS] HV glitch resistor/fuse?

To: <amps@contesting.com>
Subject: [AMPS] HV glitch resistor/fuse?
From: jtml@lanl.gov (John Lyles)
Date: Tue, 9 Dec 1997 21:19:49 -0600
Hey,

When you fuse a HVDC line, remember that when the fuse blows during a
fault, that DC doesn't like to be interrupted. What I mean is that you can
sustain a nice long hot arc. I remember an experiment in college (Virginia
Tech probably had one of the last power engineering labs before all the
kids went digital) in which we had these motor generator sets, and load
banks. We'd rev up the motor, either in series or shunt field config, and
start adding load (resistors to the generator output). The switches were
merely open knife switches. Electrical engineering students had to be smart
and keep a hand in their pocket, and flip these little blades open or
closed by the knob. The floors were made of wood blocks, elevated, to
prevent killing someone. We called it the Frankenstein Lab. It separated
the 'men' from the 'boys' in engineering (sorry if there are females
reading this, it's only a figure of speech!) school.

OK so now you have the setup. I had my experiement really loaded up, and i
believe that the motor was series wired, such that the additional load
caused the field current to increase, which increased the speed, which
increased the current, you get the picture... This thing was screaming, and
the professor came rushing over to tell me to shut it off. I started to
lift the DC load with a knife switch. I don't remember the voltage, but it
may have been a hundred or so. The thing just sizzled and kept going, with
a nice 1 to 2 inch arc continued in space across from the contact to blade.
YIKES! Luckily someone pulled the main breaker, one of those spring loaded
arc-chute Westinghouse things that really looked like it was from Dr.
Frankenstein's lair.

-----------------

I learned that large DC currents are hard to interrupt. Back to the present:
We have this 120 KV fuse here, that opens in case a bank of 7 klystrons
faults. It separates the main power rectifier from the capacitor bank,
which is something like 80-100 uF at 90 KV. The fuse is literally 4 feet
long, and made by Maxwell in San Diego. It is like a big pipe, full of sand
and somekind of element. When it blows, the inside burns up the entire
length, and the sand sort of melts around it.

In another system, we have these 30 KV capacitors in parallel. We learned
the hard way in 1992, when one shorted and all the rest of them (dozens)
dumped their energy into the short. They had 5 Ohm 100 Watt wirewounds in
series with each cap. Every resistor exploded, not to mention having the
fire department, and a big mess of melted steel and oil. Since then, we
have designed a new type of HV fuse, that is spring loaded. It is merely a
#22 wire, suspended in a big 'grasshopper' spring, that keeps a piece of
this wire about 6 inches long, under tension. It the capacitor were to
rupture and short, the wire vaporizes, the spring rips it open, and breaks
the arc. It is helped by a carborundum (now Cesiwid) resistor of a few ohms
in series. We have tested it, and it saves the resistors, and isolates the
faulty capacitor. The grasshopper spring swings over and hits a ground bus,
another safe feature.

------------------

Now, think about using a 3AG or one of those little Bussman fuses. They are
rated for 250 VAC, not much more. People have used them for HV power
supplies for years, on the line side. When you use them on DC, and you dump
all that stored energy in an arc, that fuse will pop, but where does the
current go? it continues until the voltage is low enough to extinguish the
plasma. In this way, a plain fuse is not very good for fast breaking the
circuit during a big fault. Maybe a resistor (long, as someone mentioned)
is better, or a springy thingy, or a thin wire that goes poof. Has anyone
else figured out what the optimal fuse for a 3 to 5000 volt supply would
be?

One more thing, during a tube arc, CPI/Eimac and Burle tubes typically
suggested that the voltage in the arc is about 50 volts. Meaning that the
rest of your power supply voltage during the first instant of the arc is
across the series limiting resistor. You can use this to roughly calculate
the value of R, to keep the current below some specified limit. In the big
tuibes (yes, with handles), these methods are insufficient, and an active
crowbar, using an Ignitron or triggered spark gap, are manditory. Or one of
those fancy low-stored energy power supplies that we were chatting about a
few weeks ago here. We test the crowbar by dropping a #30 wire across the
output (with a vacuum HV switch under remote control). If the wire
survives, then the crowbar is adequate to protect the few mil diameter grid
wires in the power tubes. This is the recommended test by the tube
manufacturers. The TV transmitters have a built in crowbar and some have a
test wire as well. The crowbar is triggered by high current in the power
supply, usually from a shunt and a pulse transformer to step up the voltage
to fire the crowbar.

73
John
K5PRO



--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm

<Prev in Thread] Current Thread [Next in Thread>