[Prev][Next][Index][Thread]

Re: Light Bulb Experiment (Followup)

I previously said (in part):

> > Using 420 PPS, with each toroid-ground current surge lasting only 1.5
> > uSec, the total time "ON" time per second is about 420 x 1.5 uSec or 630
> > uSec, implying a duty cycle of only about 0.063%. The toroid-ground
> > current peaks which would have to flow to "average" 400 MA can now be
> > estimated: Isurge=(0.400)/(630x1e-6) or about 635 Amps(!). However,
> > since the actual current surges are exponentially decaying during each
> > 1.5 uSec shot, the actual current peaks are probably significantly
> > greater than 1000 Amps(!).
> >
> > This explains why these discharges look so mean and evil... they ARE!
> >
> > As always, flames, brickbats, and snickering are always welcomed.  <:?)
> >
> > Safe (and rubber-booted) coilin' to ya!
> >
> > -- Bert --
> 
> Bert
> This is neat stuff but I think one more measurement needs to be made. I
> have a major problem in believing that the filament will actually take a
> 1000 amp pulse. I think a controlled test dumping a 1000 amp pulse of
> the prescribed 1.5 usec  pulse duration thru the lamp to see if the lamp
> will actually live would help to bolster your conclusions.
> 
> Skip

Skip,

Hmm... I see your point and agree with you. 635 or 1000 amps is a goodly
amount of current, even if only for a very brief amount of time. Are you
thinking that the high amplitude pulses would cause the filament to open
up thermally (overheating or even vaporizing)? Or would the filament be
stretched beyond the breaking point due to magnetic forces? A
combination of both? 

Realistically, I'm not sure I have the appropriate hardware to be able
to generate large, controlled, 1.5 uSec current pulses of sufficient
amplitude to actually run such a test. A charged capacitor, discharged
through the lamp and a spark-gap (or Thyratron) might do it. The "cold"
DC resistance of the 100W quartz-halogen filament is about 12 Ohms. The
measured "warm" value, at 400 MA, is about 80 Ohms.  Because of the
highly non-linear behavior of the filament WRT temperature, a single
isolated current pulse would behave very differently than the 10,000th
one (i.e., after the filament has stabilized to the "average" 400 MA
temperature). 

A cold filament would require >12 KV to reach a 1000 A instantaneous
peak, not including any gap/thyratron losses. When "warm", it would need
>80 KV. Skin effect would make the effective AC resistances and applied voltages for each case significantly higher. The time constant for the exponentially declining discharge would need to be about 500 nSec to match the 1.5 uSec total ring-down time observed.

Anyone have any thoughts on a practical way this could be done?? 

 -- Bert --