You are on page 1of 4

~

,.... We have always been taught that
-. ~ nothing can exceed the speed of
~ light. Evidence exists, however, to
. ~~ the effect that this may not al-
ways be the case. Here's a de-
scrip tion of several ex-
. ... periments that seem to dis-
prove the theory of

~
relati vity, and an ex-
p lanat ion of what
. . . . may-or may not-
. " be taking place.

I' VE RECENT LY BE EN ENGAGED IN EX -
perimentation with transmission lines

\
and, in discu ssing my work with other
scientists, if I casu all y h app en to
mention-which is sort offun to do-that
I'm interested in electrical impu lses that
propagate faster than the speed of light,
I'm met with avariety of reactions.

~,
The . most usual is derision-ranging
from skepticism or incredulity to outright
reject ion . On the other hand , there are a
few people who say that the phenomenon
is all old hat, and well known.
The word "well" might be disputed,

~
By
but it is true that in the first decad e of this
HAROLD W. MILNES , Ph.D
century it was already known that electric
waves do propagate in wires at velocities
in excess of c (the veloc ity of light in free
space, equal to 2. 998 x 1010 ern/sec) .
That fact seems to have been obscured by

~
our acceptance of Einstein ' s theories of
relativity so that \;'ery few peopl e-even Properties of electricity
senior gradu ate electrical engineers-are Maxwe ll's equations seem to describe
aware of it. .We are much more familiar the properties of electricity best. They

,~
with idea that the velocity of light has c as predate the theories of relativity by 25
its upper limit and that the veloc ities of years , and were not the only set of equa-
both matter and energy are similarly li- tions proposed in the late nineteenth cen-
mited ; and that no intelligible information tury to explain the behavior of electricity .
can be propagated faster. But they were the ones supporte d by the
Sin ce there is some di spu te as to influential Cambridge schoo l, which was
whether the speed-limit postul ate of the predominant in science at the time .
relativity theories originated with Poin- Though Maxwe ll's equations are very
care or Einstein , we' ll avoid taking sides good where there is a continuous current-
and simply refer to it as the c-hypothes is. flow, they are known to be subject to unmodified equations, the velocity, v, of
It is reco gnized by relativism (the science certain errors , particularly in descri bing sinusoidal waves in transmission lines is
of rel ati vity) that the so-called wave phenome na involving moving isolated given by the following formula: v = 11
c....
velocity of electricity-i-that is to say, the charges. It is precisely when treat ing the /IT. The values of Land C are not the >
velocity at which the crest of a sinusoidal, behavior of those discrete charges that it total inductance and capacitance of the z
c
continuously emitted , electrica l signa l is best to modify the equations to agree line , but represent its specific inductance >
JJ
moves through a conductor-s-can some- with relativity as we currently understand and capac itance-that is to say . its in- -<
times exceed the velocity of light. That it. heren t inductance and capacitance de-
forms an exception to the c-hypothsis. When derived directly from Maxwe ll's term ined on a per-centimeter basis-in
55

the shorter a line you W use. image (b). A very fine wire is selected. In many cases I' ve observe d the (or falling) edge to the next. small. 400 meter s long . 1-0RIGINAL SaUAREWAVE (a) and its planes were co nnected to one another. It's not hard to find a con. each of area 500 square ern . ated from it as the waveform that exists ing boards.J W measur able results.\I j LC (though it plane were stacked benea th those . or the c-hypothesis would no forma tio n indu ced by line -distortion.. __ structe d and placed beneath the first.. then v would be greater pul ses-particul arly the former-make keep the surface area .. and then run a series of ex. and to measure the time delay-if on the screen . exac tly where waves begi n and end. but you are also warned the first's.1 fJ-s/cm . measure inherent capaci tance. Thus . 3-a . a squarewave qu ite satisfac tory. similar. the oscilloscope ' s maximum swe ep-rate separated by I em of air dielectri c. but - .. that CHANNEL CHANNEL A B means using a wire at least 3 x 10 4 cm long .L and C were both in the micro-m icro falling and rising edges of squarewave that. so that the total expose d surface of In ex per iments relatin g to the veloc ity were I Y3 fJ-S. since the input and output point s can be upper estimate can be made by noting that monit ored separately by transfe rring a the circ umference of No . the starting and end ing point s of a particul ar easily discern able displacement of I mm total inherent capacitance must be very signa l. I use lines (J) o about 400 meters ( 1200 feet) long . . 35 steel sound-recording wire (ava il- co ntradictory . You will be able A second . Then a third and fourth ity of the equation v =-. They can be performed by anyone becau se the wires are noninducti vel y with a little knowled ge of electronics. show n in Fig I-a. which would give a delay (under the c-hypo thesis) of at least one us. Obvio usly. and the boards were separated periment s to see whic h theor y hold s between the images of those points. ~ o : ' indeed. one em apart. prop agate in conductors at ve loc. respectively. Pulses are measured from one rising forming a co ntinuous transmission line trains). For a I-MH z pulse .. plane was co n- to see for yourse lf that electric pulses do . 35 wire is . to position it so that pulse is considered from one successive ance of one -ohm-per-em. it is essenti al to use would imply under the c-hypothsis. hence the in- than c. A d ua l-trace other. An air ga p of one em separated that the results do not establish the valid. It is so that the values of L and C are qu ite farads-well below the picofara d range possibl e . it is necessary to define FIG. which allows for pulses somew hat longer than a DELAY CIRCUIT Z oa: microsecond. 300 I Malmo Dr. and wound. however . but a crude ure. the better . The c-hy pothes is. 6 Ifprecise measurements are to be made o « (beyo nd ju st determining that something a: is takin g place). In most show n in Fig . osci llosco pe is not an abso lute necessity . {3 ' the two planes . and a co ndition would arise where goo d refere nce-poi nts that ca n be easily herent capac itance. despite its high resis t- ads) per ce ntimeter. taken into acco unt) . if it were applied to the FIG. as a line 400 meters long the 400 -meter line is 1000 square em . Arlington Heights.on e. and waves that have been gener. to avoid such difficul. with is more likel y to be true than would be the b the direction of the wires alternating . The total length was sho w that there are a numb er of ways to exac tly 100 meters. strung tightly back and forth. forming a panied in the output waveform by sharp se ries of 100 parallel lines. for the two are mutua lly Th at is illustrated in Fig . and f3' are images of the falling and rising able from Fidelitone. That' s what I' ve been doing ..15 line ca n distort the signal it carries . squarewav e 'pulses or trains of pulses. then you would obtain an ca pac itance of 4. mark ing points are nearly always acco m. able. I am awa re of no practical way to the ca pabilities of my equipment to meas. henri es and farads . spee d of prop agation of electrical wave .. as shown. where a ' No . The value of L was kept quit e small sion. 2-DEVICE USED BY THE AUTHOR to compare a delayed signal with the original. and Two plates .425 x 10-11 farads.025 Requirements probe from one to the other. The experi ment shows one way to accomplish If. and.106 x 10.. so as to (10. much less than that. a transm ission The delay circuit should be co nstructed basi s it is less th an 1. have a Doin g so make s it ea sy to determine the was . eac h one Basis for experiments ove rshoo t spikes immed iately following meter long (when the turns at the end are Th e ex peri me nts described bel ow them .introduce d .. no---. The from one another by 99 ern. IL 60005) which is the range of micro-mi cro fara ds (picofar. sma ll. Since too long a line ca n I- o distort a wave form. O n the face of the osci lloscope. The wire was water.. On a per-centi meter any. the capaci - ties by using a line long enough so that the del a y pre d ic te d by th e c -hy po thesis would exceed the time period of a sing le cycle. 2. The hookup for studyi ng the delay is illustrated in Fig. ductor whose inhere nt capacitance is in edges a and 13 of the original wave. Number 40 either the unm odi fied Maxwell equat ions traced thro ugh progressive stages of de. I-b . get the results needed to reach a co nclu. its inherent inductance is in the range of falling (or rising) poi nt to the nex t similar As shown in Fig . and the direction of current now do not requ ire a large cash outlay-all in anyo ne line the reve rse of that in the yo u need are an oscill osc op e and a two adjace nt lines. of electrica l signals . the trace of the output signal from the line would then be displace d at least one full wave length with respect to the trace of the input signal. with its wires running perpendicular to ities faster than c. Resistor R is used to proxim ity of the wire surfaces to one an- instances the spee ds have been beyond pro vid e a sig na l at B . squarewave generator. a hundred notch- micro-micro henries (picohenries) per poi nt. speed of pr o pagati on of squarewave The capacitive effec t of the planes is trains to be gre ater th an 100c. ju st as long as you can get R . If the delay em . es were cut . co pper wire is the finest genera lly ava il- would fail.12 ) range. 3-c and depends on the hundred times the speed of light. Also. 56 . but it is so fragile that I prefer use longer apply. The procedure used in the first mentioned earlier. in two insulat- ce ntimeter.

. .oo' B ~ IV " il' i '''I ~ il' -.1.- . .. a I I I A ex . 5-a-5-c for three different frequencies.a-4-d for three dif. The line > JJ the precision that the instrument permits....... equiva lent il' capacitance is shown in c. tI'''..8340 MHz (b) .".. I'll describe a seco nd experiment I performed .. -.~300.. ex f\ .. > to the input waveform.. The displacement of the ~5( . the deformation is continuous in not be shielded or termin ated without at the highest sweep-rate of a 15-MH z between ... ex il I ex ~ ..... il \ B \ r i\ \ I' :.. be displaced by at least 13 em with respect frequent ly coiled on them selves to en.. 4.7" /-. ~ :' II\. (-.. . r-. but it is unnecessary to do so./...1. \ . well-known delay lines and their core s are c. :/ \ -. 100.l _ V .Tn that ex periment. b B \ I'\... I i I I BI .000 ohms .. 35 steel wire in a gia nt loop once around the city bloc k where I li ve . difficu lty in obtaining clean signals be- end of the device ....~ ~ IV I ~ il' B v- I il' I il' c a b A A ex I B1:. the output wave form should was 40.. I :--1 v B4 "- I . I'"i\ V ex' I .! ~ f-. the value of R was 3500 ohms.-J-ex' uJ.. 5-SIGNALS... are the same as those ind icated for Fig. but the o ' and W A second experiment hance the delay effec t. I I • \ \ .. ' I . shie lded cab les are hypothesis.e. 200. -fV FIG. the transit time to the 200. 4. In case you think that bro adc ast sig na ls th at even the input are show n in Figs . In that. and 480-meter conti nuous loo p.. 0.. J lil' -" 'f\... V b d c FIG. /' A lil 1\ Br 'I"' / ··. 300.. A r-. 4-c shows images probed at 100. I ran 480 meters (1600 feet) of No . '\.. ". 000 ohms. I/'" H" -1 B3 ... Further more.... . / I I / :. Th e spe cific in- ducta nce of the loop can be co nsidered so small as to be negligible . / . I v "..-. Frequenc ies used 0.} \ . --11... and the inherent capac itance eve n less than that in the first y_--------99 cm------~g expe rime nt. 3-APPARATUS USED TO CONSTRUCT a 400-meter delay line is shown in a and b.... I ' .. Fig. ferent frequencies...... . design of the wire array. /\' V V fl. ~ It r ". and 400 meters from the input ca use of noise interfe rence. and the waveform s are show n in Figs."t.3579545 MHz (e). 4-d the traces are show n at 0. ~ fY ' r . The value of R was set at 5000 altering its esse ntial L and C characteris- o sc i llo scope. depending on the z c point s match the (X and 13 point s to within The apparatus ju st described is the least design of various manufacturers.. ohms and the total resistance of the device tics. I I I \J :\ . tance could be further reduced by separat. AI ex il I .. A 1\ . cumbersome for laboratory use that I have picked up so much random noise from -< Typical input and output wa veform s develop ed so far. j ex f\..... I il I \ { il If 1\ I I \. returned using FIG. 4-0RIGINAL AND IMAGED signals for three different frequencies: 1 MHz (a).... AND IMAGES. In case you feel that signals should have been at leas t 16 em if ing the wires and the planes by more than the precise alignment of the curves is th e waves prop agat e d at ve loc ity c one cm.l a ' il' .. the line cou ld other end is so brief that it is und etectable point . Accord ing to th e c.... somehow related to the length of the line. and th e resist ance of the lin e 48. . a The results were esse ntially the same . .... . There was considerable When a squarewave signal is fed to one in Fig.--. the observed effect is dependent on the squarewave was fuzzily indistinct on the 57 .. and 400-meter points. through the line .../ \ . I il ( . .

eve n if 15 of the units.-. The test freq uenc ies used were: f) = using two TV delay-lines in series. enon being discussed .. If the strip is simply left floatin g. are connected in series. tive delay . When that strip Facts VS . which checks with the velocity of light . 6-a. T = n / f . so f 1/f2 = nl/n2' The numb er of tinctly depende nt on wire size and surface waves. = ineffective . in a 480-meter area. one ern (O. a co ntinuous gra da. and its setting varied from fre quency .Electronics readers . and ground. They depe nd en- TV receivers are prob ably fam iliar to The a' and Wpoints iden tifyin g the orig. They are: "the integer relationship to the others. yet. under the c-hypothes is . so me of the peaks pre. in who le or in part . tube blew . f2. independent . unfounded in fact. the latter is obviously false. 6-COLOR-TV delay line (a) and equivalent fiction. oscillosco pe . the y are almos t e lectro mag netic waves on freq ue ncy) tic rep resentation (Fig. tirely on the effect s of dispersion (the most Radio.. my limited made to pass through that co nductor at a tion of effec ts can be followe d ." corr esponded may be deduc ed by using What is free space? It is a mathem atic al some eleme ntary arithmetic and from the FIG..e. Of some 15 obliterated. velocity co nsiderably grea ter than that of The original signal beco mes more and Th e analyses of both Brillou in and light. when co mb ined with the small in cond uctors are cont radictory. for we can find current texts that lead to only one conclusion : An electrical point " S" (in Fig . If a one. or whether a delay long . " Exp erime nta l evi de nce diti on s of very low L and C values. along with its schema. One is inal wave remain evid ent for a long time d ep end en c y of th e v e l oc i ty of shown in Fig . the fact that the input and output waves b velocity of light in free space. in du e co ur se . a freq uen. 6-b). the foil was left floating and pagated down a line without distortio n. but . about 27 meters (80 feet) defo rma tio n of the old. that is . where f is delay of two fL S . For an out.--- . Inherent capac itance is dis- n2/f2.with velocity of propagation is independent of the correspondin g graphs show. for a coax ial cable would not be a more effec - would have entered the line. and f." 58 . If the first statement of length/c. circuit (b ). co mes a matter of subjec tive opinion Typically. Beneath (/) the windings lies a strip of foil covering ~ about a third of the tube . no two have vestig ial. but to what extent. off the air. ilarly insulated. invalid . th e ab o ve e qua tio n we re tru e .1 = explanation of curves. my graphs illustrate . those of empty and in- the transit time of the line were T = wire . delay is prod uced." and "a pulse can be pro- them prod uced a measurable displace. perime ntal det erm inations of t he er . and wellian and relativistic theories both sup- Delay lines it is hard to decid e where the orig inal port the c-hypothes is. th e n of the line . f2 = 0 . It really be. How is the \' = l/ jLC form ula o suits. se pa rati ng th e co nd uc to rs .the product LC is independent of --. velocity depends on signa l in a co nductor . under suitable con. more def~rmed. At certain times. " ment of the waves. of course . none of the foil grounded to prod uce a delay. To frequency. Now . there would be no point in put wave to appea r witho ut displacement coiling its core to produce a more effec- with respect to the input wave. to lead the duct ors is approximately 3 x 108 output prob e along the loop as had been R meters/second. R-E had precisely the same characteristics . 7-0RIGINAL SIGNAL and images returned The very analysis that persuades us that 1/2f2.: local TV stations and airpor t beacon were electric constant and permeability of the insulating medium. get curve B. To obtain the clea n signals works says: shown. its properties con- frequencie s.4-inch) in diameter. as well. Cur ve A shows the input tive delay line than any other wire sim- equal to the signal-frequency. " Radio is ju st like TV-only the picture C? unconn ected to gro und-no measurab le Johnson . I found that it was necessary to c:-r~ -~ ----~ .-- . See text for \' equals 1I/CL-a result that may be I MHz. can be one megohm to zero . It co uld not occ ur at a different overlay of one substrate of a print ed- frequ ency . terstellar space. Y2. crea ted to suit the results of Max- fact that no significant displacement oc. The assurance that all is well and line would be either one or two. and seem to ignore ex perimental evi- o of the signal passed through the line re.. as signal. In that time n = Tf waves only two delay lines were used . valu e 1//[C . c . which checks with ex- done in the previous experi ment. they are of the devices I' ve stud ied . unless n2 = Tf2 circ uit board upo n another would also be wer e also an integer.357 9545 MH z.8340 MH z.has other conseq uences 0 . as ca n be see n in curve C. it can be seen that the a ' and Wpoints In real life . and the phenom - Th e one-u s delay lines used in co lor. either f) = f2. for air-insulated con- It was imp ossib le. literature o is c o nnec ted to ground. though . Howev. recon ciled with the c-hypothesis? W. 6-b). eve n in curve D . and where the delayed wave is. wave is. or f) = FIG. And. curve 0 the output signal. well' s equatio ns. mislead- u:J inductance of the winding. involving some 405 meters ( 1300 feet) of media: at others . betwe en the foil state " . where length is the fixed length To obtain the results shown in Fig . in matter. in Transmiss ion Lines and Net. agrees. Somm erfeld cla im to ex plain why Max- dominate. none of them bears an that are rather surprising . although they always remain and. curred at sev era l different . wound as a single layer on a form of the input signa l has take n place . is ju st means that f)/f2 would be equal to 1. 2. or that-an assurance . f 1 = 2f 2. It ca n mego hm poten tiomet er is inserted at true . if the velocity of electric- cy would have to be chose n that would al pulses were unaffected by the distance make n a whole number . equipment cannot measure accurately. venientl y simulate those of co nducting Suppose that .. Nor is the former critical one I have so far performed . not o nly Let us assume . for instance .-- !5J1 perform the ex periment between 3 AM the size and separation of the con- and 4 AM on a Sunday morning when the ductors and depends only on the di- ~. but was the case for one of the frequ encies. . the in here nt The statements made in the literature g: cap acitance of the line is increased to the relatin g to the veloc ity of electric signals frl point that. other delay devices that depend on an say f). they co nsist of a co il whether the new wave is merely so me of fine wire. if dispersion is not assumed . as I regar d this experime nt as the most match the a and ~ points. nearly correc t. dence . a one-us de lay ing. The numerical J. that that would the TV delay-device not work . 7..