greg@utcsri.UUCP (Gregory Smith) (12/17/86)
It seems to me that the broadcast power idea wouldn't have worked. If you created a powerful oscillating magnetic or radio field, over the whole earth, at whatever frequency, eddy currents would be induced in any metal object. Huge amounts of power would be required just to drive all these currents, even if no 'antennas' (to tap the power) were in use. Given the number of large metal objects on the planet (bridges, battleships, skyscraper frames, offshore oil platforms, Cadillacs) it seems that the amount of wasted power would vastly exceed the amount put to useful work. Chopping these objects up using insulation would reduce the losses significantly, but I don't think this is practical and I don't think it would help enough anyway. I'm not up much on the history of this. Was the eddy current phenomemon unknown at the time? -- ---------------------------------------------------------------------- Greg Smith University of Toronto UUCP: ..utzoo!utcsri!greg Have vAX, will hack...
yerazuws@rpics.RPI.EDU (Crah) (12/19/86)
The reason for the spark gap is that it's a device to convert low frequency
AC (like 60 Hz) or DC. Otherwise (as previous postings have indicated) all
you get is a big step-up coil and the neat Tesla-type discharges don't
appear.
First principle you need - How a "spark gap" works:
Current in a spark gap behaves very differently than current in a
resistor. More specifically, it takes about 10,000 volts per centimeter
(I know it varies a lot, flames to /dev/null) to throw a spark between
two rounded objects. Sharp points, edges, and etching/blast zappings can
decrease this, but 10 kV is not a bad estimate.
That 10 kV is to START the spark. Once the spark initiates, the ionized
air in the region quickly decreases resistance (takes nanoseconds) and
the voltage drop across the active arc becomes approximately constant at 25
volts in dry air at sea level, irrespective of the length of the arc.
As long as enough current is available to maintain the arc (and the
electrodes don't melt :-) ) the voltage will not exceed 25 to 30 volts,
no matter how long (or short) the arc is.
So, what the spark gap gives us is a switch that in a few nanoseconds
can go from an insulating to a conductive state. (Turning the switch
off takes significantly longer - because it turns off only when the
air in the arc cools and deionizes. So, heavy arcing actually
INHIBITS the spark-gap switch from turning off and on rapidly.).
So what we want to make the high frequency out of DC or 60Hz is a
circuit that works like this-
1) Capacitor charges up -
2) causing spark gap to fire ("switch on") -
3) which rapidly discharges capacitor -
4) falling voltage causes arc to cool and "switch off" -
5) which brings us back to state (1)
...with the whole loop going as rapidly as possible. Note that the
2->3 step should take on the order of 500 nanoseconds or so.
A circuit to do this looks like:
R
---------------\/\/\/\/-----------------
| | |
60 v -------
Hz gap ^ ------- capacitor
| | |
----------------------------------------
and the switch-on speed is controlled by the capacitance of the
capacitor and the parasitic inductance of the wires between the gap
and the cap. The gap firing rate is controlled by 1/RC (assuming
that it's much faster than the 60 Hz supply, which we can view as
DC because it's "on" for so long compared to the spark-gap part of
the circuit.
Why the Resistor "R"? It's so that when the spark gap DOES fire,
the 60 Hz supply cannot put enough current into the arc to keep the
air ionized. If R wasn't there, you'd only get one firing per
half-phase of the 60 Hz (or 120 firings per second).
We can get cuter and more efficient by replacing R with a coil,
such that the time constant of the coil-gap circuit is much longer
than the time it takes for the spark gap to fire and quench. Then,
voltage builds up, gap fires, large V appears across coil, but
because the coil has a decent inductance, the power supply can't put
enough current into the gap QUICKLY to keep the arc from quenching.
So the arc quenches. But current is now flowing from the power
supply thru the coil, charging the cap. As the cap charges, the
voltage across the coil drops. At some point in the future (hopefully
soon) the current through the coil is again less than what the
spark gap needs to arc continuously. At some still LATER point,
the voltage on the capacitor is enough to fire the gap - which
quickly discharges the capacitor. Cycle repeats.
So, that's how the spark gap converts low-frequency AC or DC into
high FREQUENCY AC.
(note - if the coil doesn't have enough inductance, the arc won't
quite quench- and the repeating cycle stops repeating)
Now we need to convert the high Freq. AC current to high freq AC volts.
That's what the air-core transformer is for. We have this pulsating
magnetic field around the coil we put where the resistor was originally.
We can put a second coil, magnetically coupled with the first coil, and
tap off the high freq. energy directly.
So, that's the secondary coil for your Tesla coil.
Now, note what happens as you draw energy from the secondary coil (the
tower/nail-on-top/whatever: the secondary coil, by coupling energy
out to somewhere, impedes the expansion of the magnetic field in
the primary coil. If we impede that magnetic field, the current in the
primary is also impeded. Which means the whole oscillation slows down.
The frequency of oscillation decreases but the power output stays constant.
What if you couple the secondary to some other non-resistive load (like
a fluorescent tube, or a pinwheel, or the ionosphere? Then the
two oscillators (one being the load, one being the spark-gap/coil/cap)
will exchange energy. However, the spark/coil/cap oscillator can
oscillate over a relatively broad band (as long as the cap recharges
before it's called upon to have enough volts to fire the gap) it really
doesn't matter how often or slowly the gap fires.
So: second important principle: Tesla coils TUNE THEMSELVES TO THE
TARGET'S NATURAL RESONANT FREQUENCY (within reasonable bounds).
Generally, the coil has to have a natural oscillation frequency somewhat
higher than the target's but it works the other way around, too.
This tuning action isn't like "we have this small servomotor diddling a
potentiometer"- it's that if the target wants to oscillate slower, it
will hold the oscillation in the coil/gap/cap circuit back. When it
finally lets the spark gap fire, the kick resulting causes the target
to oscillate with the coil/gap/cap.
Now, a few comments:
I've built tesla coils with "habitrail" tubes as the outer coil
forms and toilet plungers (the giant suction-cup darts) as inner coil
forms.
They work.
They make sparks that you thought only Lucasfilm could do.
They make a lot of ozone (irritant/corrosive to rubber and fabrics)
They make a LOT of radio hash.
If you build one, be careful (Hopefully it's really oscillating up
in the MHz, but I don't know a crafty way to really tell for sure).
Don't run it for long periods, or unattended. And realize you
are smashing radio communications over a broad bandwidth- this is
a moderately antisocial act and can get you in trouble if you act
immaturely.
Voltages all around the coil/gap/cap/power supply are LETHAL and
if they're at low frequency, will really cause your heart to stop
and your brain to cold-boot.
(I think I got that all right. Anyone have any changes/amplifications?)
-Bill Yerazunis
{ code, encode, decode, DES, NSA, CIA, KGB, technology export, Iran,
Nicaraugua, Contra, rebel, terrorist, plutonium, botulism
- I realize it's a futile gesture, but it's a gesture nonetheless.
Besides, the poor G-3 in Langley who has to screen these
messages might want to know how to build a Tesla coil. :-) } milo@ndmath.UUCP (Greg Corson) (12/22/86)
I have seen some tesla coils that use multisegment spark gaps...that is, a series of metal disks in in insulating rack with about a quarter inch of space between each disk...you feed power at the ends and it has to jump several gaps in series. Would this give you a richer RF source? Does anyone have any comments on how to "tune up" a tesla coil for maximum output? I have one about 4-5 feet tall wound on a 4 inch core with several large glass capacitors...it gives me about 6"-8" sparks. Is there anything I can do (calculations...etc) that would improve efficiency? Changing the size of the primary, the spark gap...etc? I have been told that the "Q" of each coil should be as high as possible, how would you calculate it in this situation? Greg Corson seismo!iuvax!kangaro!milo
yerazuws@rpics.RPI.EDU (Crah) (12/22/86)
it does have a couple of other advantages: 1) When a disk pits from spark erosion, it can be rotated without having to remove and re-polish it. When the entire periphery is eroded, then you remove it and put it on a lathe for repolishing. 2) Greater thermal mass - helpful in two ways. First reason is that the spark is quenched by cooling- which happens in two ways. First way is by radiative transfer - the bluish light. Second way is by conduction to the electrodes (remember that plasma has a lot of properties we normally associate with metals - and heat conduction is one of them) So, if we have a 1/4 inch spark gap as a single gap, the longest distance from an ion to a cool electrode is 1/8". If we break the 1/4 inch into four gaps, each 1/16" gap will have the same radiative cooling, but now no plasma ion will be more than 1/32" from a nice cool copper disk. Net result is that the spark quenches faster and more dependably. The second advantage of greater thermal mass is that as the electrodes heat up, they lose some of their quenching ability. Spreading the heat load over a number of electrodes helps this (in the short term, at least) Note to people in the Pittsburgh area: Buhl Planetarium had a wonderful twenty-foot or so Tesla coil in operation last time I was there (about 6 years ago). It apparenty was built as an exhibit by some engineers at Westinghouse, materials paid for by Westinghouse. Maybe some CMUer can post a followup article on the coil (is it still there? Still in operation? What are it's vital statistics?). I saw this coil fired up once - the sparks were at least fifteen feet long. Very impressive. -Bill Yerazunis
keithd@cadovax.UUCP (Keith Doyle) (12/23/86)
In article <162@ndmath.UUCP> milo@ndmath.UUCP (Greg Corson) writes: >Does anyone have any comments on how to "tune up" a tesla coil for maximum >output? I have one about 4-5 feet tall wound on a 4 inch core with several >large glass capacitors...it gives me about 6"-8" sparks. Is there anything >I can do (calculations...etc) that would improve efficiency? Changing the >size of the primary, the spark gap...etc? >Greg Corson I've found that more capacitance helped for me. I came across a military oil-filled HV cap (watch those PCB's) that made a world of difference in efficiency. Though I too, would like to hear any tune-up tips. I've been wondering about putting a rotating disk hooked up to a variable speed motor and inserting into the gap area, but haven't tried it yet to see if it works. I did find that using a fan to blast air over the gap helps though makes it a bit more noisy (and it's already too damn noisy). Keith Doyle # {ucbvax,ihnp4,decvax}!trwrb!cadovax!keithd # cadovax!keithd@ucla-locus.arpa
cgs@umd5 (Chris Sylvain) (12/24/86)
In article <162@ndmath.UUCP> milo@ndmath.UUCP (Greg Corson) writes: > >I have been told that the "Q" of each coil should be as high as possible, how >would you calculate it in this situation? > There's two "flavors" of Q: Loaded and Unloaded. Unloaded Q is determined by the series resistance present in both inductors and capacitors. This series resistance dissipates energy in the circuit, and affects the "sharpness" of the response peak of a resonant LC circuit. "Most diagrams of resonant circuits show only inductance and capacitance; no resistance is indicated. Nevertheless, resistance is always present. At frequencies up to about 30 MHz this resistance is mostly in the wire of the coil. At higher frequencies energy loss in the capacitor also becomes a factor. *This energy loss is equivalent to resistance* [emphasis mine]. When maximum sharpness or selectivity is needed, the objective of design is to reduce the inherent resistance to the lowest possible value. The value of the reactance of either the inductor or capacitor at the resonant frequency of a series-resonant circuit, divided by the series resistance in the circuit, is called the Q (quality factor) of the circuit, or: _Q= X/R_ where: Q= quality factor X= reactance in ohms of either the inductor or capacitor R= series resistance in ohms" (1) Q can be used to determine the voltage across the LC circuit: _V= Q*E_ where: E= the voltage being applied to the circuit. _Loaded_ Q: "However, when the circuit delivers energy to a load (as in the case of the resonant circuits used in transmitters) the energy consumed in the circuit itself is negligible compared with that consumed by the load." "The Q of a parallel resonant circuit loaded by a resistive impedance is: _Q= R/X_ where: R= parallel load resistance in ohms X= reactance in ohms." "The effective Q of a circuit loaded by a parallel resistance becomes higher when the reactances are decreased. A circuit loaded with a relatively low resistance (a few thousand ohms [!!!]) must have low-reactance elements (large capacitance and small inductance) to have a reasonably high Q." (2) (50/75 ohms is relatively low to me !) The reactance in ohms of an inductor is: _2 * PI * f * L_ where: PI= famous constant f= frequency of interest L= value of inductor in Henries I hope this helps... --- (1) & (2) _The ARRL Handbook for the Radio Amateur_ pps. 2-32 & 2-34 reprinted and edited somewhat without permission... -- --==---==---==-- .. The jaws that bite, the claws that catch! .. ARPA: cgs@umd5.UMD.EDU BITNET: cgs%umd5@umd2 UUCP: ..!seismo!umd5.umd.edu!cgs