> >This is extremely interesting discussion.... but.... Could somebody
> >explain to me the physics of how a parasitic oscillation can affect the
> >filament in a tube enough to break it?
> Electromagnetic force occurs at a right angle to the flow of electric
> current. This is why, when arc welding with high currents, the cables
> tend to twich when the arc is struck.
But the cable MELT long before they are torn apart by the tiny little
twitch produced by many HUNDREDS of amperes.
Anyone who remembers looking at a series filament string in an
All-American-Five radio has probably seen this effect. The weakest
and thinnest filament lights up like a light bulb while the heaviest
filaments slowly reach operating temperature.
Filament failures (inrush or otherwise) occur from heat related
fatigue, not magnetic force. Magnetic force is proportional to the
current, thermal energy is proportional to the current squared.
The thinnest (and thus the weakest) area heats fastest when the
filament is cold, and goes over temperature. Excessive heat is
concentrated in the smallest thinnest (weakest) area for the longest
period of time.
That's why series connecting filaments in a high power PA is just as
dumb as series connecting light bulbs.
> >Is there so much current that the filament wire separates or burns in two
> >pieces?
> The filament does not burn in two during an intermittent vhf parasitic
> oscillation. The filament shatters in 811As and 572Bs. In 3-500Zs the
> filament is pushed sideways at the unsupported center of the filament
> helices. This decreases the clearance between the filament and the grid.
> [9/90 QST magazine, "Parasitics Revisited".]'
Can you explain how? The magnetic force is proportional to the
current. Where do all those hundreds or thousands of amperes come
from when the saturated emission current is a few amperes?
> Today I tested 3, 3-500Zs that were removed from an LK-550 that had
> arc-damaged tune capacitor plates.. The grid/filament breakdown
> potentials were 2400v, 2800v and 5100v. However, a normal 3-500Z has a
> grid/filament breakdown of 8kV to 9kV. How about the vacuum? Two of the
> tubes exhibited under 3uA of anode/grid leakage @8800v. The third tube
> was 5uA. In other words, the vacuums were good.
Many things OTHER than spacing control breakdown voltage. The problem
is voltage gradient, not simply overall spacing. What does an arced
capacitor have to do with a grid? Nearly ANYTHING that causes
excessive grid current also causes excessive tank voltage.
Overdrive, mis-tuning, improper load or a load fault. All of these
things can elevate grid current AND tank voltage.'
Many things could have caused the grid to overheat at the same time
the capacitor arced.
> >If that is the case, where does this current come from?
> The pulse of current comes from the cloud of electrons that surround the
> cathode. In a high Mu triode, oscillation with no load results in high
> grid current.
Really? And how much saturated current is available? Does it seem
logical that the "oscillation" could produce hundreds of amperes of
current for enough time to overcome the mechanical or thermal lag of
the filament?
Factually, it is impossible to obtain ten or 12 amperes of cathode
current if the grid is driven to a few thousand volts positive. Where
does the emission current come from?
> >How much
> >more is this current that what is flowing normally through the tube?
> In 3-500Z amplifiers, it is not uncommon to see 1A grid to gnd. chokes
> open during a push-push vhf parasitic. As I recall, the fusing current
> rating of #27 gauge wire it is at least 15a. That is about 100 times
> normal grid current.
The normal cause of choke failure is a HV anode to grid fault. Either
a "barnacle" or gas in the tube. The HV supply dumps to the grid (the
closest element). No way does the cathode emit 15 amperes. We
couldn't get that current if we tried, it's beyond the saturation
current of the electron emitting surface.
. ......
> That's the difference between the push-pull and the push-push variety.
So if I run my PA push-pull, the tubes will last longer????
That sounds very strange. What is the logic behind that statement?
73, Tom W8JI
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|