A DESCRIPTION OF THE TUNNEL DIODE

AND ITS APPLICATIONS

John W. Wentworth*
Broadcast and Television Division
Radio Corporation of America
Introduction istic in the forward direction. In a
The tunnel diode is one of the most
portion of the forward-bias region (from
A to B on the diagram), the tunnel diode
significant solid-state electronic behaves like a negative resistance; it
devices to emerge from the research is this fact which accounts for most of
laboratories since the transistor. its useful properties.
It is very simple in construction,
consisting only of a p-n semiconductor-
junction with ohmic contacts to the p I
'and n materials.

Figure 1 -I

It differs from a normal junction

Figure 3
diode, however, in that the semiconductor
material is much more heavily "doped* To explain how a tunnel diode
with impurity atoms, resulting in a very operates, it should be helpful to review
significant change in the diode behavior. briefly some of the highlights of semi-
In a normal diode, current tends to flow conductor physics. The treatment here
rather freely if the diode is biased in is non-mathematical and somewhat super-
the forward direction (that is, with the ficial, but it should be adequate to
p-type material positive), but the cur- give the reader at least a qualitative
rent flow is almost negligible if the understanding of tunnel diode behavior.
p type material is made negative relative
to the n type material (provided the Conduction in Metals
reverse bias voltage is appreciably less
than the so-called 'breakdownn potential). The conduction of electricity in
Thus, the characteristic curve of a any solid material depends upon the
conventional diode has this familiar presence of electrons which are free to
shape. move within the molecular structure of
the material. According to the quantum
theory, the electrons in a given atom
can exist only at discrete energy levels,
and can jump from one energy level to
another only by absorbing or radiating
at least one quantum of energy. (The
energy levels correspond roughly to the
4-REVERSE BIAS "orbits" assumed in the older versions
-V V
FORWARD BIAS - of the atomic theory.) In a metallic
conductor, the available energy levels
are very close to each other, and the
Figure 2
agitation of a solid crystal by thermal
energy at all normal temperatures is
sufficient to excite substantial numbers
The tunnel diode, on the other hand, of electrons to energy levels slightly
conducts very freely in the reverse above the average level of the outermost
direction and has an N-shaped character- tshell'i of electrons. An electron energy
diagram is frequently used to represent
*This paper first appeared in Field these conditions.
Engineers' Electronics Digest (RCA)
Vol. 8, No. 1, pp. 2-11, 1960.

21

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
 AVERAGE ENERGY LEVEL OF
- - --- OUTERMOST ELECTRONS.
THIS RANGE OF PARTIALLY
_
_ _ _ _
++++ ++ + ++ + +
OCCUPIED ENERGY LEVELS
INCREASES SLIGHTLY WITH
+ + + + TEMPERATURE.
ALL AVAILABLE ENERGY STATES
IN THIS RANGE ARE COMPLETELY
OCCUPIEO.
/
Figure 4

For every electron which has been

"excited" by thermal energy so as to
occupy an energy level above the dotted
line in the diagram above (symbolized by
minus signs), there is a corresponding
/
'vacancyC, or unoccupied energy state
somewhere below the dotted line (symbol-
ized by plus signs).
An electron at any of the partially-
occupied energy levels is able to move
rather freely through the crystal struc-
ture provided there are Ovacancies" at Figure 5
the electron's energy level in neighbor-
ing atoms. Normally, the "shuttling"
of electrons from one atom to another "forbidden region" vwithin vwhich no
occurs in randoma fashion with no net electron can exist. The available energy
movement of electric charge in any direc- states above and below this forbidden
tion. Under the influence of an electric region, are usually called the "con-
field, however, there is a net movement duction band" and "valence band," as
in the direction of the applied poten- shown in the diagram below.
tial, constituting an electric current.
Most metals are good conductors because HIGH ENERGY ELECTRONS
their substantial range of partially- / __ I CONDUCTION
__ BAND
occupied energy states yields a good
supply of so-called "free electrons" and l FORBIDDEN
of vacancies into which such electrons ELECTRON IREGION
can move. ENERGY
+ + +
Electrical Properties and VALENCE
Semi-Conductors HOLES BAND

In a highly-purified crystal of Figure 6

semiconductor material, the electrons
in the outermost "shell" of each atom
(that is, the valence electrons) are As a result of thermal agitation of the
utilized to form the bonds between crystal, a few electrons accuire enough
adjacent atoms. In a germanium crystal, energy to hurdle the forbidden region,
for example, the four electrons in the leaving "holes" or vacant energy states
outermost shell are used to form the inter- in the top of the valence band. The
atom bonds symbolized by short lines in exact number of such free electrons per
the 2-dimensional sketch shownm in Fig. 5 unit volume is a function of temperature,
(actually, the atoms form a 3-dimensional increasing at higher temperatures.
crystal lattice with double bonds, since Because there are relatively few "charge
each of the neighboring atoms contribute carriers" (free electrons or vacancies),
one bonding electron to hold the atom a pure semiconductor is a very poor con-
in the center). ductor of electricity. In fact, most
insulating materials have the same type
To break loose from its bonding role of energy diagram with such a wide "for-
and become free to move through the bidden region" that virtually no elec-
crystal structure, an electron must trons are able to absorb enough energy
absorb a very substantial amount of to hurdle the barrier at normal tempera-
energy. The energy level diagram for a tures.
semiconductor thus includes a substantial

22

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
"Doping"' of Semi-Conductors normally associated wit,h the bonding
function are able to absorb enough ener-
For practical semiconductor devices, gy to break loose and become charge
the pure semiconductor material is carriers, leaving behind a corresponding
"doped" by the controlled addition of number of vacancies or "holes" in the
specific impurities to modify the elec- valence band. Most of the electrons in
trical properties. So-called n-type the conduction band, however, are
germanium, for example, is formed by "donated" by the impurity atoms. Both
adding an impurity such as arsenic with the electrons and the "holes" are avail-
five rather than four electrons in the able as charge carriers. Electrons move
outer "shell." In moderate quantities, directly, usually in short hops between
atoms of the impurity material replace neighboring atoms. "holes" only appear
some of the germanium atoms in the crys- to move; when an electron from a given
tal structure, and each "donates" one atom moves over to occupy a vacant
surplus electron not needed for bonding, energy state in an adjacent atom, it
as shown by the sketch below. leaves behind a vacancy that has appar-
ently moved in the opposite direction.
Thus "holes" behave as if they were
positively-charged carriers of electric-
ity, even though it is always electrons
that do the actual moving. Holes are
only "minority carriers" in n-type
germanium because the impurity atoms
which donate most of the free electrons
do not form useful "holes." When an
electron breaks loose from one of the
impurity atoms, the atom becomes ionized
(positively charged), but it remains
locked in place in the crystal structure
EXTRA and does not offer an available energy
ELECTRON state into which electrons may move

/ readily. The material remains electri-

cally neutral when viewed as a whole, but
the charge carriers which are free to
move are predominantly negative, hence
the n-type designation.
P-type germanium is formed by the
addition of small amounts of an impurity
such as indium which has only three
valence electrons in the outer shell of
Figure 7 each atom. When indium atoms are sub-
stituted for germanium atoms at some of
the locations in a crystal lattice, each
The extra electrons are readily excited such atom becomes an "acceptor" for one
to energy levels falling within the con- of the bonding atoms from one of the
duction band of the germanium crystal, neighboring germanium atoms, thus creating
so the energy diagram of an arsenic- a vacancy or "hole" into which another
doped germanium crystal is modified as electron may move. Each indium atom
shown below. becomes negatively charged by the surplus
electron attached to it, but these
CONDUCTION BAND ionized atoms are unable to participate
MAJORITY in current flow because they are locked
4 CARRIERS in position in the crystal structure.
(ELECTRONS) Thus the holes (which behave like positive
FORBIDDEN charges, although they are actually
REGION
ELECTRON
ENERGY electron vacancies) are the majority
+ ++ + MINORITY carriers in p-type material, as shown by
VALENCE *- CARRIERS the energy diagram below.
BAND (1HOLES)
The few electrons in the conduction
band represent those that have acquired
Figure 8 enough energy by thermal excitation to
"escape" from the bonding role.
As in the case of the pure ger-
manium crystal, a few of the electrons

23

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
 tend to pass through the junction to the
CONDUCTION BAND MINORITY less-populated valence band of the
4* CARRIERS
(ELECTRONS)
n7type material (this really means, of
course, that electrons form the valence
FORBI DDEN band of the n-type material move over to
REGION occupy some of the surplus vacancies at
ELECTRON MAJORITY the same energy levels in the p-type
ENERGY +V+N+ +B+A+
++ + + + 4 CARRIERS material). The net flow of electrons
+ ++ + + + (HOLES) from left to right continues only until
VALENCE BAND the right side becomes negatively
charged relative to the left by just
the right degree to form an electric
Figure 9 field across the Junction great enough
to maintain a right-to-left electron flow
exactly equal to the left-to-right flow.
The conductivity of either n-type The difference in charge raises the
or p-type semiconductor material can be absolute energy level of the p-type
controlled by varying the concentration material, and tends both to augment the
of the impurity material. For all p-to-n flow (by permitting the electrons
practical purposes, each impurity atom traveling in this direction to Wfall" to
donates one charge-carrying electron or lower energy states) and to diminish the
hole, so the number of impurity atoms n-to-p flow (by requiring the electrons
per unit volume is the primary factor in traveling in this direction to gain
determining conductivity. Tunnel diodes additional energy).
use semiconductor materials with approxi-
mately 1,000-times more impurity atoms When an external voltage is applied
per unit volume than the materials used, to an n-p junction in the reverse direc-
for conventional diodes and transistors. tion (that is, with the n-type material
To complete this brief explanation of positive), the height of the barrier in
tunnel diode behavior, we should examine the energy diagram is soon increased to
the properties of p-n junctions with the point where virtually no electrons
both low and high concentrations of from the n-type region can acquire
impurity atoms. enough energy to cross the junction.
Thus, there is no flow to cancel the
flow of electrons from the p-side to
Semi-Conductor Junctions the n-side, and a net current flows under
the influence of the applied potential,
When a junction is formed within a as shown at the lef t in Figure 11. The
single crystal between p-type and n-type magnitude of the current flow in the
materials, a permanent electric field is reverse-bias direction does not increase
established within the crystal, producing
the energy level- conditions shown in the
diagram below.

4- ACCELERATED BY ELECTRIC FIELD

THERMAL EXCITATION _ "FALL-DOWN HILL")

ELECTRON CONDUCTION BAND
ENERGY

(MEASURED IN
ELECTRON VOLTS; ( -- | FORBIDDEN ELECTRON FLOWS
NEGATIVE REGION ACROSS THE JUNCTION
VOLTAGE REGION MUST
UPWARD) / ++ + + BALANCE WHEN THERE
/+1 +44+ + ARE NO EXTERNAL
++ + + CONNECTIONS

N-TYPE P-TYPE

REGION
VALENCE BAND

Figure 10

The surplus electrons in the con-

duction band of the n-Ttype material can
move across the junction rather freely to
occupy some of the available energy
states in the more sparsely-populated
conduction band of the p-type material.
At the same time, some of the holes in Figure 11
the valence band of the p-type material

24

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
significantly with increasing voltage The combination of the very thin j unc-
because of the very limited supply of tion and the substantially greater off-
minority carriers (thermally excited set in potential on the two sides of a
electrons and holes). When an external tunnel diode junction makes possible the
voltage applies a forward bias to the flow of electrons through the junction
junction, as shown at the right in the region by the process known as "tunnel-
sketch, the potential barrier is effec- ing." (The potential barrier is suf-
tively reduced so that substantial num- ficiently great that relatively few
bers of the majority carriers can acquire electrons are able to cross the junction
enough energy to pass through the junc- by acquiring enough energy to hurdle the
tion, permitting an appreciable current potential barrier.) The great majority
flow. When the forward bias becomes suf- of the free electrons which impinge
ficient to "wipe outn the potential upon the junction region in the course
barrier, the junction becomes, for all of their random motions are reflected
practical purposes, a linear resistance-- back into the same material from which
that is, the current changes become they came, but a small fraction of them
proportional to the voltage changes. are able to pass through the junction
and occupy some of the vacancies in the
Tunnel Diode Junctions partially-occupied energy levels on the
opposite side of the barrier. The
The effective width of the junction probability that a given electron will
region in a semiconductor diode is strong- pass through the junction is very low.,
ly influenced by the concentration of but the supply of "charge carriers" is
impurity atoms (i.e., the degree of so great that even the small fraction of
"doping"u), since the electric field in electrons which pass through can support
the junction region can be maintained an appreciable current flow. When no
only by the ionized impurity atoms locked external voltage is applied, the elec-
into place in the crystal structure. If trons flow in both directions in suf-
the impurity atoms are only sparsely ficient quantities to balance each other.
distributed through the crystal, the When a reverse bias is applied (n-type
space required to build up a sufficient material made positive), there is a
net difference in charge to balance the stronger net flow of electrons from the
electron flows is quite great, and the p-side to the n-side, corresponding to
junction region is relatively wide. point (A) on the characteristic curve
Conversely, a high concentration of shown in Fig. 13.
impurity atoms, as in a tunnel diode,
makes possible the creation of a very
narrow or abrupt junction. At the same
time, the high impurity concentrations
in the two materials used in a tunnel
diode lead to a considerable increase in ,T.
the magnitude of the difference in
potential required to balance the elec-
tron flow across the junction, as shown
by the energy diagram below.

CONDUCTION BAND MINORITY CARRIERS

N-TYPE P-TYPE BECOME INSIGNIFICANT
- _ _ BECAUSE OF HIGH
IFORBDDEN IMPURITY CONCENTRATION Figure 13
REGION
For slight amounts of forward bias
_-,+++++++
_==_=,,,++++++++ in the region between (B) and (C) on the
curve, the tunneling effect continues
to operate and the junction conducts
rather freely. As the forward bias is
ELECTRONS FLOW further increased in the range between
+ ++ ¶ REGION
K JUNCTION THROUGH JUNCTION
WITHOUT SIGNIFICANT (C) and (D), however, the current flow
(VERY NARROW) CHANGES IN ENERGY begins to decrease because the partially-
occupied energy levels in the n-type and
p-type materials are shifted to bring
them opposite the forbidden regions on
Figure 12 the opposite sides of the junction, thus
removing the conditions which make
possible the tunneling effect. At the

25

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
current minimum, point D, the tunneling properties have been so thoroughly ex-
effect has virtually ceased, but a small plored in conjuncti6n with transistors
current flow is possible because the and conventional diodes. Silicon does
potential barrier has been reduced suf- not appear particularly attractive for
ficiently that some electrons are able tunnel diodes because the effective "mass"
to acquire enough energy by thermal ex- of its charge carriers are somewhat
citation to cross the junction. As the greater than for the other materials,
forward bias is further increased to leading to impaired high frequency per-
point E or beyond, the device behaves formance. One of the most promising
essentially the same as a conventional materials is gallium arsenide, which not
junction diode. As noted earlier, it only offers a relatively wide voltage
is the negative resistance in the region swing in the negative resistance region
from (C) to (D) that accounts for most of (approximately 1 volt), but also operates
the tunnel diode' s useful properties. well at temperatures up to 4000C.
Materials for Tunnel Diodes Typical Applications
Some of the basic materials from Tunnel diodes show considerable
which experimental tunnel diodes have promise for a wide variety of applica-
been fabricated are listed below. tions, including the following:
1. Germanium 1. Oscillators (from low audio
2. Silicon frequencies up to the kilomegacycle
3. Gallium Arsenide region).
4. Indium Arsenside
5. Indium Antimonide 2. Mixers or heterodyne converters.
With any given material, the tunnel 3. Amplifiers.
diode characteristics are controlled
primarily by varying the concentration 4. A combined oscillator, mixer
of impurity atoms or the area of the and amplifier (all in a single diode!5.
junction. The voltages at which the
maxima and minima occur in the negative 5. Bi-stable memory element for
resistance characteristic tend to remain computers.
constant for a given material. For ex-
ample, in all germanium tunnel diodes, 6. High-speed. switching element for
the maxima and minima occur at about 50 logic circuits.
and 350 millivolts, respectively, as
shown in the following sketch. 7. Negative resistance element for
filters.
Ge TUNNEL DIODE
WITH GREATER DOPING 8. Rectifier, voltage regulator, or
OR INCREASED JUNCTION AREA small-signal detector. (Diodes for this
service may be processed to minimize the
first current peak, but will exploit the
I ,1E'GALLIUM ARSENIDE
TUNNEL DIODE high conductivity in the reverse direc-
I _~
y
/ XGe
<8TUNNEL
> I (APPROX.) tion.)
/ ~ DI / 6
Performance Characteristics
200 400 600 600 1000 1. As amplifiers, tunnel diodes
MILLIVOLTS offer low noise and high gain-bandwidth
Figure 14 products. In the microwave region, they
are expected to compete closely with
masers and parametric amplifiers with
If the degree of doping on the junction respect to low noise, and will be much
area is increased, the current values lower in cost.
change, but the curve retains essentially
the same shape, as shown by the dotted 2. As computer elements, tunnel
line. Use of a different material, such diodes offer extremely high switching
as gallium arsenide, may move the voltage speed, in the rmilli-microsecond region.
points, as indicated by the dashed line The high switching speed and the broad-
in the sketch. band amplification capabilities result
both from the extremely narrow junction
Germanium will-undoubtedly be a and from the basic nature of the t"tunnel-
popular material for much of the early ing" process, which is not greatly dif-
work with tunnel diodes, because its ferent from ordinary conduction in metals.

26

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
 3. Tunnel diodes operate over a
wide temperature range, extending from
absolute zero up to possibly 40°C (for
gallium arsenide units). This wide
tolerance results from the fact that only
majority carriers are used in the normal
operating regions.
4. Tunnel diodes are relatively
insensitive to atomic radiation (their
tolerance is about 1,000 times greater
than for most transistors).
5. In contrast to transistors,
tunnel diodes will tolerate momentary
overstress well beyond their normal
operating conditions. In many cases, it
is actually quite difficult to dissipate
enough power in the junction to damage
the device.
6.
experimental results indicate that tunnel
diodes are relatively immune to con-
tamination by moisture and atmospheric
gases, thus eliminating much of the need
for hermetic sealing.
7. Because hermetic sealing is
possibly not required, tunel diodes may
be mounted in extremely small., light,
encapsulated "packages," permitting even
greater space and weight savings than
transistors.
8. Tunnel diodes are inherently
low-impedance devices with inherently
high capacitance because of the "abrupt"
junction. When biased for operation in
the negative-resistance region, the
equivalent circuit for a tunnel diode is
approximately as shown below.
-R
Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
Typical values for the negative
resistance (-R) might range from a frac-
tion of an ohm to about 100 ohms, de-
pending upon the material used, the degree
of doping, and the junction area. The
junction capacitance (c) likewise varies
with the construction of the device, but
is typically of the order of 75-100
micro-micro-farads. The resistor r
represents the dissipative resistance
of the ohmic contacts and the semi-
conductor material itself. r is usually
small compared to -R, and can be neg-
lected for many practical circuit
analyses where first-order approxima-
tions are adequate.
9. The tunnel diode operates at a
low power level, both with respect to
input requirements and with respect to
available output. In particular, the
Both theory and preliminary voltage swing is definitely limited in
applications where operation must be
confined to the negative-resistance
region.
10. The two-terminal nature of the
tunnel diode offers both advantages and
disadvantages. On the plus side, most
tunnel diode circuits are extremely
simple, permitting significant reductions
in size and weight and substantial
improvements in reliability. On the
other hand, the lack of inherent isola-
tion between input and output is often
a disadvantage making it impossible, for
example, to cascade amplifier stages in
the usual fashion. (I.t is expected,
however, that new circuit approaches can
be devised to overcome this handicap.)
11. Because of its inherently
simple construction and relatively non-
critical materials, there is good reason
to believe that tunnel diodes can be
produced at very low cost, once tech-
niques for mass production have been
developed.
12. The reliability of tunnel
diode devices is expected to be sub-
stantially better than transistors be-
cause of the combination of factors
listed above as items three through
seven, plus nine and ten.
Basic Circuit ConceDts
A comprehensive analysis of tunnel
diode circuits is beyond the scope of
this paper, but a few simple examples
can be presented to illustrate the
simplicity of tunnel diode circuits and
to show how it is possible to perform a
Figure 15 variety of functions with a two-terminal,
negative-resistance device.
27
Low-Pass Amplifiers When the tunnel diode is in place,
the equivaleat circuit becomes:
A low pass amplifier may be con-
structed by connecting a tunnel diode,
biased for operation in the negative DIODE CAPACITANCE
resistance region, across the load
resistance (RL) of a signal source which
LS. 0

may be symbolized by a constant current e

generator shunted by the source resist- CD

ance (Rs). The bias resistor (Rb) shown
in the following sketch is normally
quite large compared to the load resistor, Figure 18
and can be ignored for first-order cir-
cuit analysis. Its function, of course,
is to establish the DC operating point Because the diode acts as a negative
for thb tunnel diode in the negative reistance, it effectively becomes a cur-.
resistance region; it forms a voltage rent sources contributing additional
divider with the parallel combination of current to the load and source resistors.
Rs and RL. The output voltage under these conditions
(e2) can be calculated as before:
RT (-RD)
SIGNAL SOURCE
LOAD ,
DC SUPPLY e2 RT - RD
A t
The voltage gain of the amplifier is
I
t
<RS tRL
< +>DIODE
Rb
~~TUNNEL 1S+R e2 -RD
el RT - RD
Figure 16
In theory, the amplifier offers
The voltage gain of this simple infinite gain when the parallel combina-
amplifier may be determined readily by tion of the source and load resistors
comparing the performance of the circuit equals the negative resistance of the
as shown with that of the same circuit diode, and the amplifier inverts the
with the tunnel diode removed. If the polarity of the signal if RT is greater
tunnel diode is not in place, the equiva- than RD. In practice, it is desirable
lent circuit becomes: to keep RT somewhat less in magnitude
than RD to avoid stability problems.
(The simple biasing arrangement selected
for illustration also requires that RT
LS X be less than RD.) When defined as the
ratio of the power output with the diode
t l T~~1 in place to the power output with the
diode removed and all other circuit

'lie'1
constants the same, the power gain is
simply the square of the voltage gain as
eR <RLI
expressed above.
The frequency response of the
amplifier is limited by the diode
capacitance (plus wiring capacitances)
Figure 17 shunted across the effective output
resistance. Although the capacitance
may be of the order of 100 uuf, the
The output voltage under these con- resistance values are sufficiently low
ditions (el ) is equal to the signal that the amplifier may have a gain-band-
current multiplied by the equivalent wridth product of 300 megacycles or better.
resistanice of Rs and RL in parallel. (A voltage gain of 10 over a bandwidth
In mathematical form, of 30 megacycles is easily achievedl)
RS RL
el mi sRs =R'is RT

where RT is the equivalent resistance of

the parallel combination.

28

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
Band-Pass Amplifiers or Oscillators Logic Circuits
The following circuit configuration The very simple circuit shown below
can be used for either a band-pass ampli- is so versatile in performing logic
fier or oscillator, depending upon the functions that it is sometimes referred
choice of circuAit constants. to as a "universaln logic circuit.

INPUT/OUTPUT
TANK BLOCKING CHOKE RESET 0 OUTPUT
Li DC SUPPLY
TAPS ON COIL L2
MAY BE USED TUNNEL
FOR IMPEDANCE DIODE RB +
INPUT(S) R
MATCHING IF +
REOUIRED. 4C 1 Rp

Figure 19

For analytical purposes, this cir- Figure 21

cuit may be simplified as follows:
The bias resistor (RB) is sufficient-
ly large that the load line is almost
REFLECTED FROM SOURCE, PARALLEL EQUIVALENT OF
CHOKE AND BIAS SUPPLY
horizontal, as shovn below.
LOAD OR BOTH
. A I

EFFECTIVE I 1 I
IMPEDANCE 2 2 J
_ 2
OF TANK AT > ~~ .~
FREOUENCY o R S-RD CD .R aL
OF INTEREST ° <1 r 1.

Figure 20

The circuit is equivalent to two

tuned circuits tied together showing a
common capacitor (the diode capacitance).
The input/output tank circuit always Figure 22
appears inductive at the frequency of
interest, because the diode capacitance There are two stable operating
is effectively in series with the tuning points, A and B separated in voltage
capacitor. To avoid oscillation in the sufficiently that they may be used to
DC supply circuit, the value of serve as the two states of a binary
1 1 D memory. If the circuit i s operating
- - must be greater than 2 - at point A, it can be switched to the B
Ru (RD) condition by the application of a posi-
tive voltage pulse of sufficient ampli-
If the circuit is to provide stable tude to push the voltage beyond the
amplification, the value of positive voltage peak; the circuit itself
1 1 will then "flip"t rapidly through the
- -must also be greater than negative resistance region until stabil-
RI RD ity is restored at operating point B.
Likewise a switch from B to A may be
2 -. If --- is made less than initiated by a "reset" circuit which
LI RI RD momentarily shorts the tunnel diode or
applies a negative voltage of sufficient
2 $ circuit will perform as an amplitude to shift the diode voltage
below the current minimum into the
oscillator. negative resistance region. In some
applications, the B to A switching could
be accomplished by a negative pulse
applied through a suitably-connected
standard diode. In general, the circuit
has the properties of a bistable flip-
flop.

29

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.
 If an additional input connection for experimental purposes have been
is provided, the circuit can perform as constructed by soldering the se.aicon-
a gate for either land' or "lor" service. ductor pellet between two flat plates
For service as an t"or" gate, the operat- measuring about l/8 by 3/8 inch.
ing bias at point A should be adjusted Insulating spacers and encapsulants are
so that a positive pulse on either of used to hold the plates apart, forming
the two inputs will cause the circuit a "sandwich" about 10 mils thick. This
to switch to the B condition. For "and" structure not only miniimizes lead
service, the bias should be adjusted to inductances, but also lends itself
lower the position of the load line so readily to circuits constructed with
that it takes the presence of both input miiicrostrip transmission line. The forra
pulses simultaneously to cause the factor for production versions of the
switching action to occur. device has not yet been standardized,
but it is highly probably that some form
of flatplate mounting will be used. In
Physical Construction Techiciues general, circuit construction practices
will be similar to those currently used
Because of the very low impedances for UHF devices. Leads must be kept
associated with tunnel diodes, it is very short, and extensive uses should be
necessary to orinloyD construction tech- made of flat strips, in preference to
niques which minimize both the resist- round wires, to minimize unwanted in-
ance aad inductance of interconnecting ductances. The small dimensions associ-
leads. (Lead inductances can resonate ated with tunnel diodes are highly
with the diode capacitance to cause compatible with the Micro-Mlodule con-
parasi tic oscillations.) Tunnel diodes struction technique,.

30

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 19:60 UTC from IE Xplore. Restricon aply.