[Prev][Next][Index][Thread]
Real data from real experiments.
All,
I took the time out to repeat two experiments which few seemed to do but
a lot have commented on. I performed the first experiment presented here
over a year ago and the latter experiment 5 years ago. So I repeated
with fresh data and first rate instrumentation.
Experiment #1 2 Capacitor charging problem.
Instrumentation; Keithley 242 precision laboratory regulated supply.
Keithley 615 precision digital electrometer
(1 teraohm input resistance)
B&K Precision 875A digital LCR meter
Beckman 310 digital voltmeter
All test common data:
Two caps are arranged with the minus terminals strapped permenantly
together with both the minus of the electrometer and the minus of the
supply connected to same.
postive lug of the charged cap was connected to the supply with a
temporary hook connector with series 10K ohm resistor to limit inrush
current. This hook could be removed after charge level was determined.
Also permenantly connected to the positive lug of the charged cap was the
positive lead of the electrometer. This was used to determine charge
voltage of the capacitor prior to transfer of charge to the discharge
capacitor. finally, also connected to the positive lug was a single pole
knife switch with was normally left open during charging. the other end
of this knife went to the uncharged capacitor. The uncharged capacitor ws
temporarily shorted with a heavy lead to assure no residual charge and
then removed just proir to the test.
Test A
Capacitors used:
Specially selected (from 30 identical capacitors) modern Mallory
1800 ufd 500 volt electrolytics. 3" diameter by 9" tall Selected using
LCR meter so the two caps used differed in value by 0.4%. (1814 and
1821ufd) Both capacitors assumed to be relativley lossey and loaded with
internal inductance.
The set up was charged a measured 99.95 volts as read by the digital
electrometer. The Beckman voltmeter was used to cross confirm this. the
hook charging lead was disconnected from the supply. The knife was
closed. A loud explosion occured with bits of molten copper bouncing
around on the bench top. immediately the electrometer was read and the
final voltage was 49.76 volts the beckman was then connected and the
reading verified to within .5%. The electrometer was left connected to
the now joined pair of capacitors and after 20 minutes read 49.58 volts.
A second test was made where the only change was a 20K ohm resistor was
placed in series with the knife to allow for slow charging of the
uncharged capacitor. the knife was opened and the hook reconnected to
the supply. a shunt was used to short out the capacitor to be charged
and was left in place for 2 minutes to assure 0 charge retention.
Again the hook was disconnected from the supply when the charged
capacitor read 99.93 volts. The knife was closed and absolutely no
visible spark or noise was noticed. The electrometer fell over the 12 RC
time constants allowed prior to taking a "finished reading". The volatge
at the end of this period read 49.82 volts. Again confirmed to within
.5% by the Beckman voltmeter. After 100 time periods (50 minutes) the
voltage read 49.12volts.
Test #2
Two superb high energy .03ufd discharge storage capacitors which were
custom wound for me by NWL with a measured internal inductance of 25
nanaohenries were now used. They measured .0316 and .0317ufd
respectively. Same setup as above. first test was a solid wire knife
switch circuit.
The charged capacitor was charged to close to 1000 volts.
The main capacitor read 1087.0 volts prior to supply charging hook
removal and knife closure. A rather smallish but noticable pop and
sparking were noted at closure. The electrometer read 543.2 volts.
The 20K ohm resistor ws plased back in the knife circuit. the same
routine of disharge of the cap to be charged was observed and the main
cap charged to 1086.2 volts and the knife closed with no noise or spark.
After many thousands RC time periods. (2 seconds) the electrometer was
read on the electrometer to be 542.8 volts. The Beckman could not be
used here because of its heavy 10 megohm loading of the circuit. Only
the electrometer could play this game. after approximately. 1 million
time constants (10minutes) the electrometer read 520.4 volts.
Conclusion:
The voltage on any capacitor discharging into an identical capacitor with
divide by 1/2 or equal 50% of its original voltage regardless of massive
blasting of copper to molten beads in the exchange or a kinder, gentler
flow of charge through limiting resistances. 1/2 of the original stored
energy (1/2cv^2) is aboslutely and irrevocably lost in this situation to
a plethora of forms of waste energy given off to the evironment and
circuit components. Charge is conversed as well as energy. The energy
transfer to the other capacitor will forever undergo a 50% loss factor.
Experiment #2
A barium titanate capacitor was removed from its epoxy housing. the
block of titanate was .810" thick and 1.77" in diameter. Two 2" brass
disks were used as plates to contact both sides of the slug. The B&K LCR
meter ws used to measure 4.72nf of capacitance (originally it was a 480nf
capacitor dielectric constant k~1200.) A 4 pound lead weight was
consistently used as a compression force device. Two sheets of 5 mil
thick virgin mylar (k=3.5) 2" in diameter were sandwiched between the two
brass disks and a reading of 1.51nf was recorded.
Next the stack consisted of disk-mylar-titanate-mylar-disk (as in a water
capacitor being shielded from the plates by thin membranes of plastic.)
Now the reading for the total group was .99nf.
Conclusion:
Regardless of the dielectric constant of a superior dielectric, any
insertion of a dielectric of vastly lower value, even if very
thin compared to the superior dielectric, (in this case 81 times
thinner), will drag the final value of the resultant capacitor down to a
level even below that of a capacitor made up of only the weaker
dielectric material! There are no free rides. The game is rigged!
DO THE EXPERIMENT! (Benjamin Franklin)
Richard Hull, TCBOR