mirror-textfiles.com/bbs/KEELYNET/ENERGY/ujtosc.asc

______________________________________________________________________________
|  File Name      : UJTOSC.ASC       |  Online Date     :  12/12/94          |
|  Contributed by : Darrell Moffitt  |  Dir Category    :  ENERGY            |
|  From           : KeelyNet BBS     |  DataLine        :  (214) 324-3501    |
|           KeelyNet * PO BOX 870716 * Mesquite, Texas * USA * 75187         |
|        A FREE Alternative Sciences BBS sponsored by Vanguard Sciences      |
|----------------------------------------------------------------------------|
The following is based on experiments done by Darrell in his work with high
density charge clusters and how they could be generated.  Darrell is an active
researcher who has freely shared his ideas with KeelyNet.  This is an
excellent file with numerous applications for the experimenter.
                                Thanks Darell!
------------------------------------------------------------------------------
                       UJT Relaxation/Pulse Oscillators
                              by Darrell Moffitt

                 (with little thanks to inaccurate references,
                    and many thanks to those which aren't)
------------------------------------------------------------------------------
These following circuits depict unijunction transistor based oscillators.
They are easily modified to vary pulse width and repetition frequency over a
broad range, and are recommended for applications requiring outputs between
1/8 and 3 watts, at current levels between 5 and 250 milliamps.

For example, an npn - coupled circuit will yield 100 milliamps at 7 - 7.5 VDC
(app. 10 VAC) from a fresh 9 V battery.  Source input may be varied from 3 V
up to approximately 30 V, though greater care is necessary at higher voltages
and frequencies.  This will be explained in greater detail in the section
describing circuit operation.

For those unfamiliar with unijunction transistors (ujt's), a bit of background
is in order. The ujt is among the oldest of active semiconductor components,
dating from the early 1950's. It is also the most simple in construction and
use. Its primary function is switching and timing, with a frequency range from
several minutes to several hundred kilohertz.

Some varieties go up to app. 1 megahertz, notably the 2N4947. The more common
varieties are the 2N2604 and the 2N4892, which sell on average between $.30
and $.50 apiece. Both are rated for up to 30 V source input and 50 milliamps
emitter output.

Ujt's consist of a single bar of n-doped silicon, with a positive region which
acts as a gate and control junction. The positive junction is termed the
emitter, while the ends of the negative bar are referred to simply as base-1
and base-2. The positive junction is an ohmic connection, which indicates that
the positive region is attached, rather than diffused.  It is this property
which gives the ujt its special characteristics.

The technical details of ujt function will be addressed later, but it suffices
to note that ujt's function much like the better known silicon controlled
rectifier, with current controlled by voltage.

                                 ______
                                 | N  |---base-2
                                 |__  |
                   emitter   --- |p_| |         Unijunction transistor
                                 |    |
                                 |_N__|---base-1


The crudest form of ujt oscillator requires just three parts, and will output
close to 50 milliamps, at a voltage controlled by the source and the timing
resistor. The waveform is that of a typical relaxation oscillator, roughly a
bent sawtooth. By adding extra bias resistors across the two bases, positive
or negative pulses may be drawn off as well.

This circuit yields ONLY the relaxation waveform, while the following circuits
yield either weak or amplified pulses. Here now is the simplest:

                                 ---- ----  (+)
                             R1  ()   | b2  |
                          <----  ---- <20> ujt # battery
                  output     C1  []   | b1  |
                          <----  ---- ----  (-)

The timing resistor, R1, MUST be set at a minimum of 2K, though specific apps
may require higher values. This is necessary to maintain stable oscillation.

R1 controls both power and frequency SIMULTANEOUSLY, so adjustments of the
timing capacitor, C1, are necessary to achieve a desired frequency and power
output. Oscillation rate is approximately .8/R1*C1.

In these circuits, base-1 is connected to the negative ground line and C1,
while base-2 is connected to the positive rail and R1. The ujt emitter
connects to both C1 and R1 at the junction, and C1 takes any value between
several hundred microfarads and about 500 picofarads.

C1 may be either polar or non-polar - note the polarity! The negative pole of
C1 should always be attached along the negative ground line (base-1) if polar
capacitors are used.

Some circuits require sensitive setting of R1. For most applications, it is
not recommended that one go above 10K, as this reduces power output to
unacceptable levels.

You can usually identify base-1 and base-2 by given specs in metal-can TO-18
models, while the TO-92 plastic case (2N4892, 2N2604) models place base-1 to
the left side of the flat face, and base-2 to the right, as the ujt is seen
head-on.

A pulse circuit merely adds two resistors, typically 47 ohms on base-1, and
100 to 180 ohms on base-2. 100 ohms on each base is typical. A weak negative
pulse is taken from base-2, while a high (positive) amplitude pulse is taken
from base-1. Pulse forms are rounded or "spiked" unipolar, and output is
generally less than that of relaxation circuits.

You will no doubt wish to try low value potentiometers to optimize
performance. Adding a diode between the ujt emitter and R1/C1 junction will
increase the standoff voltage of the device and yield roughly square pulses.
(The anode of the diode connects to the emitter.)

You may also couple R2 through a small coil to C1 on the negative ground rail,
thereby achieving more complex waveforms, or, construct a small auto-
transformer connected such that base-1 connects to the tap, while the two ends
are connected from the end of R2 to the negative pole of C1. This will boost
the output voltage slightly and, by raising the voltage to C1, decreases the
time of the C1 discharge cycle.

For higher performance square pulses, the circuit below adds an npn transistor
inverter, which takes the output of base-1 (+), feeds it to the base of the
npn (direct coupled) and outputs a negative pulse of higher amplitude. In this
circuit, it is MOST important to optimize the base-1 resistor (R2) at a value
between 100 and 1000 ohms, depending upon your choice of npn transistor.

This is due to the tendency of direct coupled circuits to overheat, AND the
necessity to protect the base-emitter junction of the npn from overvoltage.
This will be explained in more detail shortly. Here now is the square pulse
oscillator:
                          (+)
                     |-----------|
                     |           --------- .--->  (+) output
                     |     R1  ()    |     .--->  (-) output
            battery  #          ---- <20> ujt |         < Collector
                     |     C1  []   () R2 --<2D> npn    < Base
                     |           --------- |         < Emitter
                     |-----------|               (npn orientation)
                          (-)

Note again that R2 is connected between the base and emitter of the npn.  That
is, R2 is connected in such a way that one end is attached to base-1 of the
ujt, and the other end is attached to the negative ground line AND the lead
from the emitter of the npn. Another lead goes from base-1 of the ujt to the
base of the npn.

The optimal value of R2 depends upon the setting of R1. In general, R1 is
inversely proportional to R2. Also, high frequency operation (above 10 KHz)
appears to require higher R2 values due to power characteristics which require
more stabilization. It is recommended that you use a 5K-10K pot for R1, and a
1K pot for R2. This allows you to "dial up" desired outputs with different
varieties of npn. When R1 is at a minimum, 220 ohms or higher is suggested for
R2.

This circuit has been tested with several kinds of npn. All appear to work,
but R2 will vary widely. 330-470 ohms is a good general range. When tested for
extended runs (2 hrs.) with a 9-18 V input, and R2 set at 470 ohms, good power
output was achieved with no shutdown.

The npn types tested included common (TO-92 case) 2N3904's and 2N2222's, but
better results were achieved with metal-can 2N2222A's and .8 A, 1.8 W B5J 7539
medium-power transistors (available in Radio Shack "grab bag" packages). SK-
3010 transistors (used in TV's) and similar were successful, but their results
were weak in comparison to those designed for low power supplies.

Again, one should experiment with potentiometers for given applications.  The
output, taken from the positive rail and collector of the npn, may be filtered
through an RC differentiator and diode shunt to isolate the negative pulse.

You may observe while testing that this circuit generates substantial radio
frequency output. A common receiver will detect the output at its RC frequency
- that is, one may place the receiver close (6" or so) to the circuit, and
detect a tone at the RC frequency.

When R1 is set around 3,000 ohm, and C1 is set at 1 nF (giving an RC around
200 kHz), an audio tone may still be detected with the receiver. This is
likely due to  stray harmonics, and may be used or shielded as desired.
(Detection is through the AM band from 550 kHz to 1600 kHz...).

This circuit may be modified to control pulse width by replacing R2 with an
inductor, typically of 100 microhenries or more. The LC constant then
determines pulse width, while the RC constant controls repetition rate.  In
this case, a resistor of 330-470 ohms is connected between the positive rail
and base-2 of the ujt.

Another resistor of 220-330 ohms is connected from the positive rail of the
ujt to the positive rail of the npn, while a resistor of 47-220 ohm is
connected between base-1 of the ujt and the base of the npn (coupling
resistor). By setting C1 low (1 nF) and the inductor at 100-500 microhenries,
you may obtain pulse widths below one microsecond. One circuit of this sort,
with 27 VDC input, output large sparks when the leads were shorted.
------------------------------------------------------------------------------

                                  Ujt Theory

The ujt is a voltage controlled switch, similar to the silicon controlled
rectifier or thyristor. When voltage is applied to the ujt, current flows from
base-1 to base-2 with a resistance of 7-10K. The positive emitter is
controlled by voltage increase across C1, so that when C1 reaches a sufficient
value, the emitter conducts and discharges C1.

The most important values in ujt's are the intrinsic stand-off voltage, <20>, and
the interbase resistance, R-Bb. R-Bb represents the sum of Rb1 and Rb2, the
base-1 and base-2 resistance:

                               R-Bb = Rb1 + Rb2.

With the emitter open, Rb1 is larger than Rb2, so the total voltage across Rb1
and Rb2, V-Bb, is divided proportionally.

The fraction of V-Bb appearing at the emitter junction (between Rb1 and Rb2)
is the stand-off voltage, <20>:

                              <20> = Rb1/R-Bb

                        which may also be written:

                             Rb1 = <20> * R-Bb.

Thus,

            <20> = Rb1/(Rb1 + Rb2), and Vb1 (voltage across base-1) =
                                 <20> * V-Bb.

<EFBFBD>, V-Bb, Rb1 and Rb2 are typically given by manufacturers for specific models
of ujt.

The voltage at which the emitter conducts, Vp, or peak voltage, is
proportional to the product of <20>, and V-Bb plus the diode potential, which
is about .7 V for silicon. Thus,

                          Vp = <20> * (V-Bb + .7 V).

For V-Bb = 9 V, <20> = .78, the emitter conducts at .78 * (9 V + .7 V), or 7.57V.

More extensive theory is available in texts, but a word of caution is in
order. While some texts indicate that the timing resistor of the circuit, R1,
may be set as low as a few hundred ohms, based upon detailed calculations,
this is rarely the case. More to the point, it is often FALSE. For reasons not
immediately apparent, setting R1 below 2k almost inevitably causes unstable
operation. Many thanks to that rare reference which states this explicitly!

The exception to this rule may be those oscillators which employ capacitors
larger than 10 microfarads. In this case, the longer time constants, (and
stabilizing characteristics of large capacitors), may allow one to go down to
700 ohms or so. You will probably find that quick potentiometer testing is
more reliable than theory, in this case. This is especially true when one is
working with "off the shelf" parts of dubious origin.

The most ubiquitous ujt is known to this experimenter simply as the "mu 10",
so called because it is marked with the Motorola logo, resembling the Greek
letter "<22>", and an "I O" along the base. It is probably the 2N2604, and sells
for $.25-.30.
------------------------------------------------------------------------------

                                 Applications

For those desiring a cheap, powerful oscillator to step up voltage with
transformers, this is one of the cheapest and most versatile of all circuits.

With a 9 V battery source, this circuit easily lights NE-2 bulbs from a 12.6-
120 VAC @ 300mA transformer (Radio Shack 273-1385B) and other low frequency
tranformers. It continues to do so when C1 is varied from 47 microfarads down
to 1 nanofarad, with R1 set between 2-5K, and R2 set at 470 ohms.

Furthermore, it continues to light the NE-2 when a second stage, 5-230 VAC @
.4 A transformer is added, giving several kilovolts AC at a minimum 1/25 W.
(CAUTION!)

Another application is in ultrasonics. One may purchase small transducers for
$2.00 or so, and power them from one or more of the circuits described.
Suggested experiments include sonoluminescence and levitation. Both require
the use of opposed transducers.

For sonoluminescence, one merely directs opposed transducers through a fluid
medium, such as water, and observes the light discharge which originates from
cavitation.

Levitation may be observed by opposing tranducers in a small cylindrical
chamber filled with small particles. When vapor or smoke is introduced to the
chamber, one may observe the active interference patterns created by sound.

You should note in both sonoluminescent and levitation experiments that the
effects are due to the properties of longitudinal wave action.

In ultrasonics experiments, current is typically the more important quantity.
For a given power output, one way wish to use lower voltage, high current
sources, such as D cell batteries.
-----------------------------------------------------------------------------
This set of circuits compares well with more sophisticated oscillator types,
such as astable (dual npn/pnp form) multivibrators or Miller, Colpitts and
similar LC oscillators. In particular, it is cheaper and more versatile than
most op-amp or general transistor oscillators.

With an average cost around $1.00, it serves many needs for simple, powerful
oscillators. The ujt, though old and simplistic, will likely continue to serve
in some form for years to come.

                             Darrell Moffitt, 1993
------------------------------------------------------------------------------