BEAM HEADS 101 - a beam circuit tutorial

Rev 1.0 11/99 - wilf rigter

WHAT'S A BEAM HEAD

A BEAM HEAD is a mechanical and electronic assembly (Fig 1) that has the=
simple function of detecting and tracking a light source. It does this by =
rotating a set of photo sensors so that both receive an equal amount of lig=
ht. The head is usually a circuit board containing the electronics mounted =
on the body of a small gear motor with the shaft of the motor, as the neck,=
attached to fixed support or mobile robot body. While seeking or tracking =
a light, a head continuously rotates around its neck. Since an umbilical po=
wer cord could lead to strangulation, the power source (battery or solar) i=
s usually carried along with the electronics on the head.

By adding a second motor and a duplicate set of electronics, the head ma=
y also look up or down. A head rotating in a 2D plane is referred to as a 2=
Degrees of Freedom (2DOF) head and a head with the additional up down moti=
on scanning a 3D volume is called a 3DOF head.

ELEMENTARY DESIGN DEAR WATSON

In any design project, we chose and combine elements from various older =
designs blended with new ideas to meet specific requirements. The choice of=
one element may influence the choice of the other elements. For example, a=
large motor may require the addition of a separate motor driver. The Power=
Smart head circuit is a simple solution to meet the design requirements for=
a small efficient 2DOF head controller. It has some unique design ingredie=
nts but it owes its concept and general form to Mark Tilden and others and =
ultimately to nature itself.

 THE BASE

Of all the head elements, the base or support structure is the least sen=
sitive to influencing the other head elements. It can be almost any shape t=
hat does not obscure "vision" and can made of any material sufficiently str=
ong to hold up the head. Needless to say, for humans as well as robots, the=
overall aesthetics of the head are an important part of the design and whi=
le the simple tripod in this example will suffice, don't limit your imagina=
tion in your own design choices. It is also possible to wall or ceiling mou=
nt a head with the appropriate support. The details of a mobile base are be=
yond this introductory tutorial, but I hope to explore the design of a head=
mounted on a robot in a future article.

THE MOTOR

In the PowerSmart head design, an efficient 30 rpm gear motor was chosen=
but unmodified and modified hobby servos have also been used albeit with a=
different electronics design. Your choice of motors will have an influence=
on the design of the electronics and for now we will concentrate on small =
gear motors, typically found in video camera lens assemblies (ie Nihon - BG=
Micro), which can be driven directly from 74HC/AC logic. The motor I used w=
as mounted with the shaft pointed vertically down. A small brass collar wit=
h a setscrew was attached to the end of the shaft and soldered to the coppe=
r wire tripod base.

THE BRAIN

The head has a brain made with various electronic components including r=
esistors, capacitors, diodes, LEDs, solar cells, photo sensors, a voltage t=
rigger and a single IC chip. The brain controls the motion of the head, in =
response to the location of a light spot within the field of vision. There =
are many ways to construct a head circuit: I used a Radio Shack 2"x2" squar=
e perfboard mounted with two stiff 1/4" wires connected to the solder termi=
nals on the top of the motor body. The two photo diodes (PD) or light depen=
dent resistors (LDR) inside small black tubes are mounted on one side of th=
e board facing out, like a pair of eyes, in the same direction. To help you=
understand the basics before tackling what goes on in the brain of a HEAD,=
I have also included a short appendix with a simple description of electro=
nic principles and useful mechanical models of electronic components used i=
n this tutorial.

WHAT IS A POWER SMART HEAD

The POWERSMART (PS) head circuit was designed as an improved version of =
the bicore head. The operation of the PS head circuit will be described in =
more detail as an example of BEAM heads in general. The basic PS head schem=
atic in Fig 2 contains everything except for the SE part. The PS head uses =
a 74HC240 Octal inverting buffer organized as 4 inverters in 2 groups each =
with a separate group output enable. If the second group of inverters is no=
t yet used, they should have pins 1, 2, 4, 6 and 8 temporarily connected to=
0V. Because of its power saving features the PS head is easily solarized b=
ut for some applications, the basic circuit plus batteries is all you need =
and of course there is another set of inverters in case you want to try bui=
lding a 3DOF head. If want to solarize the head, one of the two the solar e=
ngines can be assembled and tested. The PS head is controlled with pin 19 c=
onnected to the SE enable output instead of 0V. The operation of the circui=
ts is described in the next few sections following the order in which the P=
S head can be build and tested.

Fig 2

THE SOLAR ENGINE

The power supply for the SE head can be 3-6V Alkaline batteries but for =
long term autonomous operation power can be generated by a solar cell and s=
tored in a large capacitor or nicad batteries as shown in Fig 3.

Both the bicore heads and PowerSmart heads can be solarized this way. Th=
e solar cell should be chosen to be compatible with the capacitor or rechar=
geable battery voltage rating. With sufficient light the solar cell will sl=
owly charge the nicad battery or capacitor. A Ge diode is recommended place=
d in series with the solar cell to prevent the discharge of stored energy w=
hen the light level fluctuates. When the voltage of the stored charge is su=
fficient to operate the circuit the SE triggers. The output of the SE enabl=
es the head circuit and head will attempt to find and lock on to a bright l=
ight. For the PS design we look at two SE options: one using a MC34164 volt=
age supervisor in a simplified Chloroplast SE circuit and the other is a 13=
81 SE adaptation by Darrell Johnson shown also in Fig 10. Of course only on=
e of the two SE circuits should be used.

1381 SE

The 1381L voltage supervisor in Fig 4 is used to sense if the voltage on=
the power supply capacitor has sufficiently charged up (about 3V) and its =
output goes active high when the 1381 triggers. The 1381 has very small hys=
teresis and will quickly reset so its output is used to SET a latch made wi=
th two inverters of the second group. The inverted output of the latch prov=
ides and an ENABLE signal to hold the 74HC240 tri-state control pin19 activ=
e low. The latch is held SET by 2 diode in a positive feedback loop. When t=
he supply voltage drops below the forward voltage of the 2 feedback diodes,=
the latch resets disabling pin 19. Because the PS head uses very little po=
wer when locked on, the charge on the solar cap can last a long time. When =
the motor runs, the voltage on the storage capacitor drops until it is less=
than the forward voltage of the 2 SE latch diodes at which point the latch=
resets. The voltage on Pin 19 will rise to V+, disabling the PS motion gen=
erator circuit and the power supply capacitor starts to charge up again fro=
m the solar cells.

Fig 4

THE CHLOROPLAST SE

The Chloroplast SE (CSE) design in Fig 5 is much simpler and uses a MC34=
164 voltage trigger instead of a1381. No latch is required for the CSE sinc=
e it cleverly uses a series input resistor to increase hysteresis between t=
he set and reset voltages of the trigger. The 1M resistor is adjusted to se=
t the trigger point and the reset occurs near 5V. Instead of the usual MPS1=
2A darlington motor driver transistor in the original design by Craig Mania=
rd, we use a 2N3904 for logic inversion to drive the 240 tri-state enable p=
in 19 active low. If you intend to build a 3DOF PS head, then both pin 1 an=
d pin 19 would be enabled.

Fig 5

PS CIRCUIT DETAILS

In the next paragraphs, we describe each section of the PS head design a=
s you build it. I recommend that you initially use a 3V to 6V battery or be=
nch power supply to test and debug each step before proceeding to the next =
step. Even if you intend to build the solar version, you should build and d=
ebug the PS head circuit separately. Once the PS head circuit is set up, ad=
d the SE to solarize the circuit.

PS SENSORS

In Fig 6, we see a "photo bridge" with two light dependent resistors (LD=
R) or Photo Diodes (PD) used as the PS head "eyes". Either type of sensor i=
s ok. Even green LEDs can be used instead of photo diodes but are inferior =
in terms of sensitivity and can only be used in bright conditions.

The photo bridge serves the same function as photo diodes used in bicore=
head circuits but unlike the bicore circuit, the photo bridge acts as a vo=
ltage divider with a midpoint "ratio-metric" output voltage which is largel=
y independent of the absolute light level. When the top (turn right) LDR/PD=
is exposed to more light the midpoint voltage rises and when the bottom (t=
urn left) LDR/PD is brighter then the voltage at the midpoint drops. So if =
both are exposed to the same light level the midpoint voltage is about 1/2V=
+. The LDR or PD pair should be carefully matched so that their resistance =
is approximately the same value when each is exposed to the same light leve=
l. Radio Shack currently sells a package of 5 LDRs for a few dollars but yo=
u may need to buy two packs to get good matched pairs. You can use a digita=
l voltmeter connected between the midpoint of the photo-bridge and 0V to me=
asure sensitivity. With both sensors pointing at a diffuse light source, th=
e midpoint should at about 1/2V+. Don't worry if they are not perfectly mat=
ched, since this will only cause a error in the angle of the head when lock=
ed on to a light. One other supplier of LDRs and PDs is Solarbotics who sel=
l premium matched pairs ready to use. Optional 1K resistors can be added to=
each leg of the photo bridge to provide a minimum resistance between +V an=
d 0V to avoid excessive current flow if the LDR resistances decrease to ver=
y low values from exposure to very bright light. The important point to rem=
ember is that this part of the photo bridge produces an output voltage at t=
he midpoint, which is proportional to the difference in light level on each=
sensor. That makes the circuit work over a large range of light levels wit=
h little change in sensitivity. The output voltage is used to control the d=
uty cycle (on/off ratio) of the following oscillator stage in a very intere=
sting and useful way.

PS HEAD HI/LO OSCILLATOR

The brain of the PS head is a classic CMOS astable oscillator with a dif=
ference: it has an additional input resistor, which provides control of the=
symmetry of the oscillator waveform proportional to the input voltage. Thi=
s is a form of pulse width modulation over a range, which includes 0-100% o=
n/off ratios. This is sometimes called a saturating oscillator since the ou=
tput will be continuously high or low when the control voltage is higher or=
lower than the linear control voltage input range. The resulting circuit i=
n Fig 7 is therefore called a High/Low Oscillator (HLO).

First let's examine how the HLO without using the IN signal, behaves lik=
e a classic CMOS astable oscillator. The oscillator consists of 2 inverters=
with the input of the first inverter serving as a summing node for several=
signals. As can be seen, a 5M potentiometer (R1) is connected from the out=
put of inverter A back to it's own input. Because of the signal inversion, =
this is DC negative feedback, which tends to drive the input voltage of inv=
erter A towards its 1/2V+ threshold. The output of inverter A is also direc=
tly connected to the input of inverter B. The output of inverter B is conne=
cted with a capacitor back to the input of inverter A. Because of the even =
(2) number of inverter in that feedback path, this is positive AC feedback,=
which drives the input to 0V or +V. This feedback capacitor (C1) has a sim=
ilar function to the two capacitors in the bicore circuit described later. =
The inverter A initially has a voltage equal to V+ across the feedback resi=
stor which discharges the voltage on the capacitor and the input of inverte=
r A and eventually causes it's own output voltage to change from high or lo=
w towards the center linear region. When the output of inverter A enters th=
e linear region of the inverter B input the signal passes inverted through =
inverter B and its output starts to change. The output signal of inverter B=
, capacitively coupled back to the input of inverter A, will rapidly drive =
the changing input and output voltages to their saturated, supply voltage l=
imited, high or low states. Once stable, these levels are again subject to =
dc negative feedback and the process repeats but in the reverse direction. =

To recap, the voltage at inverter A output connected to the feedback res=
istor is opposite to the voltage on the capacitor at the summing input of i=
nverter A, so the capacitor discharges through the resistor with the input =
voltage of A eventually reaching the threshold. This causes the output of i=
nverters A and B to change and the capacitor sends the change back in phase=
to the summing input of A thereby rapidly accelerating the change. When th=
e outputs have changed state, the whole process repeats. The result is two =
square wave signals at the output pins of the inverters with one signal inv=
erted with respect to the other. The frequency of the squarewave signal is =
inversely proportional to the product of the capacitor and resistor values.=
If you adjust the pot value to a smaller value the frequency increases. Be=
fore the final adjustments leave the pot in the center position

PULSE WIDTH MODULATION

As described in the photo bridge section, that circuit generates a volta=
ge proportional to the difference in the left and right LDR light levels. N=
ow we convert this voltage to a useful waveform to control the motor by con=
necting the photo bridge output to the IN terminal of the HLO as shown in F=
ig 8. While separate fixed resistors can be used, the adjustable potentiome=
ters R1 and R2 are recommended for reduced parts count, to explore various =
head behaviors and to compensate for variations in voltage/logic thresholds=
. The 0.022 cap shown in the schematic can be two 0.01 caps connected in pa=
rallel.

The potentiometer R1 is connected to pins 15 and 17 and the IN terminal =
to the midpoint of the photo-bridge to adjust the sensitivity, oscillator f=
requency and degree of pulse width modulation. Since the adjustment inheren=
tly affects all three functions, it requires tuning to give the right respo=
nse. The output voltage of the photo-bridge is proportional to the differen=
ce of light on the photocells. That voltage is connected via part of R1 to =
the input of inverter where it is "summed" with the positive and negative f=
eedback of the capacitor and the feedback resistor, which is the other part=
of the potentiometer. The photo-bridge difference voltage (Vp) results in =
an offset current which generates equal left and right pulse widths at the =
oscillator outputs when the Vp is balanced at 1/2V+ (or near the switching =
threshold) but changes the left right pulse width if Vp is unbalanced. It i=
s useful at this point to also connect R2, C2, inverters C and D and the mo=
tor as shown in Fig 2 to observe the effects of the photo bridge and adjust=
ment of R1 on the motor. For this part of the circuit testing, potentiomete=
r R2 must be adjusted to zero ohms so that inverters C and D simply act as =
motor drivers. With R1 adjusted to the maximum value, and exactly equal lig=
ht falling on the photo sensors, the motor receives equal duration left/rig=
ht pulses, it vibrates rapidly but does not move. But when the photo-bridge=
is even slightly unbalanced, the effect of the offset current causes the p=
ulse width to change asymmetrically producing a net rotation in the directi=
on of the longer motor on time. By adjusting R1, the photo bridge sensitivi=
ty is changed and the motor will respond more slowly to light imbalance. As=
a rule, the value of R1 between inverter A input and output is adjusted be=
tween 90 - 50% of its full value before the sensitivity of the photo bridge=
becomes zero. Adjusting R1 to a smaller value yet is not recommended as it=
will only increase the frequency of the square wave, which will become aud=
ible but motion will drop to zero and the supply current will increase. As =
a small improvement, a 5M resistor can be added between the output of inver=
ter A and R1. With R1 adjusted between 90 and 80%, the optimum (medium) sen=
sitivity is achieved. At that point, the motor is stopped when the light is=
nearly balanced and the motor moves slowly left or right when the light is=
slightly unbalanced. When the photo-bridge is more unbalanced, the oscilla=
tor outputs are forced to a continuous high and low state and the motor rot=
ates at full speed in the desired direction to balance the photo-bridge. P>

THE Nv/Nu OUTPUT STAGE

Let's have a look now at Fig 9, the Nv/Nu output stage. With the value o=
f R2 reduced to 0 ohms during the adjustment procedure of R1, the two outpu=
t inverters C and D have no effect on the complementary square waves at inp=
uts IN and /IN other than inverting and isolating them from the motor load.=
If potentiometer R2 is 0 ohms and the photo-bridge output is balanced, the=
left right pulse width is symmetrical, and the head stops rotating but the=
power continues to flow through the motor while stopped and this is wasted=
energy. A solution to this problem is a circuit, which detects that condit=
ion and shuts off the power to the motor. . Increasing the R2 pot resistanc=
e value to mid point, it forms, in combination with C2 and inverter C, a "N=
v/Nu" circuit. When R2 is precisely adjusted, the Nv/Nu time constant will =
match the frequency of the HLO and when symmetrical or balanced as would be=
the case when the head is "locked on", the motor driver output voltages wi=
ll be in phase, completely turning off the motor current and reducing suppl=
y current to about 2 ma.

When the photo bridge is slightly unbalanced, the HLO output waveforms a=
re asymmetrical and the duration of either the left or right pulse widths w=
ill be longer than the Nv/Nu time constant and the motor will turn at a spe=
ed proportional to the difference. When the HLO output is high or low, thos=
e signals will of course quickly pass through the Nv/Nu circuit to drive th=
e motor left or right at full speed.

To recap: when the photo-bridge is greatly unbalanced, the HLO outputs a=
re forced high or low. This results in a steady state condition in which th=
e dc voltage from the two HLO outputs are can pass through the resistor but=
not the capacitor (in the fashion of a Nu). This results in the Nv/Nu stag=
e having one driver on and one off and the motor continuously rotating towa=
rds the light source. Now assume that the light source is near the center o=
f the field of vision and that the astable is oscillating but with a left p=
ulse on time which is longer than the right pulse. Now the Nv/Nu stage acts=
like a Nv monostable but again with a small twist: unlike its cousin, the =
classic Nv, this one can produce positive and negative output pulses. The r=
esult is that for the "right" direction pulses, the two outputs of the Nv/N=
u stage are the same polarity and NO current flows through the motor. But f=
or the "left" direction pulses, the Nv times out and for a short duration t=
he two motor drivers have different output polarity. This provides a propor=
tional output, which turns the head SLOWLY to the left. So when approaching=
the balanced condition the PS head slows down accordingly and the current =
also starts to drop off saving power. When the light on both photocells is =
fully balanced, the lowest power condition exists where the left and right =
pulses are the same duration and the Nv monostable never times out. Therefo=
re the motor driver outputs always have the same polarity, no motor current=
flows and the motor is stopped. This is the low power (2ma) standby mode t=
hat gives the PS head its name. The circuit is now ready to hook up to the =
SE stage.

A SOLAR POWERSMART HEAD EXAMPLE

Fig 9 shows the solarized PS head layout drawing by Darrell Johnson usin=
g a 1381L solar engine.

Note the unused input pins 6 and 8, which have been terminated to reduce=
power.

WHAT'S A BICORE HEAD

Most BEAM heads use the ubiquitous bicore circuit with photodiodes (Fig =
10), designed by Mark Tilden, also used in walkers and photovores. The desi=
gn is elegantly simple but has some shortcomings. One problem is the effect=
of the photo diodes on changing the bicore frequency as well as on/off tim=
e. Additional resistors can limit the influence of the PD's but this effect=
ively limits operation to a narrow range of lighting conditions. The circui=
t also consumes a significant amount of power in proportion to the work of =
rotating the head and that's an important factor for battery or solar opera=
tion. There are reports of an updated bicore head design but at this time i=
t has not been published.

The bicore operation is as follows:

Assume that the voltage on bicore 240 inverter A input pin is V+ and tha=
t inverter B input pin is 0V. Therefore the difference voltage (V+) appears=
across the resistor and the reverse biased PD2. The output pin of inverter=
A will be opposite to its input voltage ie 0V and inverter B output pin wi=
ll be V+. The current through the resistor and PDs starts to rapidly discha=
rge the two capacitors with the voltage on both ends of the resistor changi=
ng towards 1/2V+. The rate of discharge is influenced by increased light le=
vel on PD2, which will effectively lower the total resistance. As the volta=
ge across the resistor and PD2 drops the current decreases and the rate of =
discharge slows down. This is the characteristic exponential discharge curv=
e for a RC network. Theoretically, if the switching threshold of the invert=
ers (where the output start to change) was precisely 1/2V+ then it would ta=
ke a long time for the caps to completely discharge. But unequal capacitor =
values, small threshold offsets, superimposed voltage ripple, thermal noise=
etc all contribute to cause the threshold to be crossed before the theoret=
ical "end" of the discharge curve. When either inverter (ie inverter A) out=
put voltage starts to change, that change is coupled through the capacitor =
to the input of inverter B. This voltage is added to the input voltage of i=
nverter B pushing it across the threshold which in turn switches inverter B=
output voltage. Inverter B changing output voltage is coupled back through=
the capacitor to the input of inverter A and that pushes both inverter A a=
nd B input voltages rapidly to their respective 0V(-0.6V) and V+(+0.6V). Th=
e additional 0.6V is caused by the clamping action of the inverter input pr=
otection diodes. The inverter A and B output voltages will then be V+ and 0=
V respectively. Now the process repeats but with the capacitors discharging=
in the opposite direction and PD1 influencing the discharge rate in propor=
tion to the light level. Photodiodes PD1 and PD2 therefore influence the on=
/off times of the bicore outputs causing it to rotate the head for a longer=
time in one direction compared to the other direction. When the PD light l=
evels are balanced the head rapidly shakes back and forth equal distance ef=
fectively keeping it pointed in the same direction. For some sample bicore=
waveforms check the section on analogical waves in the appendix.

 APPENDIX A

SOME SIMPLE ELECTRONIC CIRCUIT MODELS

I find that I can imagine the signals of a circuit in my mind as I run a=
simulation on my mental circuit model. A good model of a circuit and of it=
s components are the best way to understand the behaviour of a BEAM design.=
Since BEAM circuits are so simple you can follow the action right down to =
the electron level where physics rules. Imaging you could do something like=
that with a PIC? Forget it! The mechanical models I have chosen are not pe=
rfect but they do shed light on what to expect from electronics components =
if you consider how the equivalent mechanical model would behave. Understan=
d the similarity between the model and reality and pretty soon you will hav=
e a good 'feel" for the behaviour of the real thing.

SOME SIMPLE ELECTRONIC COMPONENT MODELS

I find that I can imagine the signals of a circuit in my mind as I run a=
simulation on my mental circuit model. A good model of a circuit and of it=
's components are the best way to understand the behaviour of a BEAM design=
. Since BEAM circuits are so simple you can follow the action right down to=
the electron level where physics rules. Imaging you could do something lik=
e that with a PIC? Forget it! The mechanical models I have chosen are not p=
erfect but they do shed light on what to expect from electronics components=
if you consider how the equivalent mechanical model would behave. Understa=
nd the similarity between the model and reality and pretty soon you will ha=
ve a good 'feel" for the behaviour of the real thing.

RESISTANCE

The most elementary electronic component is the resistor, a device usual=
ly consisting of small tubular body with 2 leads. The body contains a "resi=
stive" conducting material like Carbon which slows the rate at which electr=
ons flow (amps) in direct proportion to the electron force (volts) and inve=
rsely proportional to the resistance (ohms). A useful model is that of curr=
ent as the flow of water and resistance as a constriction opposing the flow=
. eg a faucet would be equivalent to a variable resistor.

There are a number of "active" resistors used for sensors such as light =
dependent resistors (LDRs) which change value when exposed to light or ther=
mistors, which change value when exposed to varying temperatures. There are=
also humidity sensors and pressure sensors which change their resistance v=
alue according to changing conditions. Even two electrodes inserted into pl=
ant soil can be used as a resistive moisture sensor.

CAPACITANCE

Capacitors are also simple two leaded devices but they are quite differe=
nt in the effect they have on electrons. A capacitor consists of two thin c=
onducting layers separated by an insulating layer. A typical capacitor is m=
ade with two strips of plastic separating two strips of metal foil rolled i=
nto a multi-layer tube with a wire terminal connected to each metal foil. T=
he total surface area of the metal foil, the thickness of the insulation an=
d the quality of the insulation determine the capacitance (in farads) of a =
capacitor.

A useful model of a capacitor is a water filled cylinder with a flexible=
membrane separating the cylinder into two sides and two pipes at the ends =
of the cylinder (like the wire terminals). These pipes connect each side of=
the cylinder to other parts of the "circuit" allowing the water to flow ba=
ck and forth into each side. The equivalent capacitance then is proportiona=
l to the membrane area, thickness and elasticity. When the pressure on one =
side of the membrane in increased, the pressure on the second side increase=
s an equal amount. If water is permitted to flow from the second side, the =
pressure of the second side decreases as the membrane is deflected displaci=
ng a volume of water. A voltage applied to one terminal of a capacitor is t=
ransferred to and remains on the other terminal if that terminal is not con=
nected to the rest of the circuit. With the other side terminal connected t=
hrough a resistance to the rest of the circuit, a voltage (pressure) applie=
d rapidly to one side will appear on the other side. But the voltage (press=
ure) will start to drop as the charge (volume) bleeds off through the resis=
tance back to the rest of the circuit. You can try this experiment yourself=
by measuring the voltage on one terminal of a capacitor with a digital vol=
tmeter while applying a small voltage on the other side. Keep in mind that =
the voltmeter is itself a 10M resistor so it will bleed off the charge of a=
capacitor in the same way. There are some sensors that use capacitance usu=
ally in the form of a metal plate ie the capaciflector or E-field sensor. <=
/P>

DIODES - OF ALL KINDS

The diode is the simplest "active" semiconductor component. It is a non-=
linear device, which behaves quite differently depending on the polarity of=
the applied voltage.

When the cathode (stripe) is negative and the anode is positive, most co=
mmon Silicon diodes (1N4148 or 1N4003 etc) have a forward voltage (Vf) drop=
of 0.5-0.6V with the current limited at 10-100ma. At higher currents the v=
oltage can rise to 1V and at much lower currents the voltage can be as low =
as 200mV. Germanium diodes are still used for low frequency signal detector=
applications since they offer a much lower forward voltage of about 100mV =
at 10-100ma and correspondingly even lower Vf at lower currents. The Schott=
ky diode has similarly low forward voltage drop but can be designed for muc=
h higher currents. These are frequently used in high frequency switching po=
wer supplies because of the low losses when switching. Although not often u=
sed for "rectification" or signal diode applications, the light emitting di=
ode (LED) behave quite similar to ordinary diode but with higher forward vo=
ltages of 1.6V-2.0V depending on the color and semiconductor material and c=
urrent. LEDs of course also emit light when passing forwards current. Photo=
Diodes are designed to generate a voltage or current proportional to the l=
ight falling on the large geometry junction of a Si diode chip encased in t=
ransparent (but sometimes filtered) lens. When the voltage is applied in th=
e reverse direction, in theory, no current will flow but, in practice, ther=
e is "leakage" current which varies from 0.1ua or less for a Si diode to 10=
0ua for a Ge or Schottky diode depending on the applied reverse voltage. Gr=
een LEDs, like photodiodes, have a reverse leakage current proportional to =
the external light level and be used for some applications instead of Photo=
diodes. The simple mechanical model is a pressure operated one way valve t=
hat permits the flow of water in one direction but not in the other. When a=
pplying lots of reverse pressure on the closed valve, it may leak a little.=

TRANSISTORS - BASES AND GATES

The transistor is a semiconductor device that amplifies. It has a contro=
l terminal used to control the flow of electrons from input and output term=
inals.

There are two types and two polarities of transistors. The two types are=
the bipolar transistors like the 2N3904/2N3906 and the other type are the =
field effect transistors like the 2N7000. The two polarities are the bipola=
r NPN (2N3904) and PNP (2N3906) and the N channel (2N7000) and P channel mo=
sfets.

Bipolar transistors use a current controlled BASE terminal to control el=
ectron flow from the EMITTER input terminal to the COLLECTOR output termina=
l. Mosfets use a voltage controlled GATE terminal to control the electron f=
low from the SOURCE input terminal to the DRAIN output terminal. Because tr=
ansistors are made with diode-like internal connections they can control th=
e direction of electron flow only one way. Hence the need for polarities: N=
PN or N-channel devices to control the current flow from negative to positi=
ve and PNP or P-channel devices for current flow in the other direction. P>

OTHER SENSORS

The passive infrared (PIR) motion detector uses a body heat sensitive de=
vice that effectively behaves as a changing capacitance. To be sure, it act=
ually a semiconductor device with piezo electric behaviour from differentia=
l thermal stress and which results in a momentary charge transfer at the ou=
tput. The PIR module used inside motion detectors has the body heat focuses=
on the sensor with multiple "freznel" lenses (like bug eyes), which increa=
se the effect and sensitivity of the sensor. There are many other sensors t=
hat have potential BEAM applications ranging from ALPHA brain wave sensors =
to microphones, all subject for future beam designs and discussion.

MAGNETICS

There are other passive components especially electromagnetic devices li=
ke inductors and motors, which are used in BEAM tech. However they are more=
complex in behaviour and will be "modeled" in a future tutorial.

SOME ELECTRONIC TERMS AND PRINCIPLES

AC/DC

The terms "AC" and "DC" will be used in the description. While these ter=
ms stand for "alternating current" and "direct current", in this discussion=
, we will say that AC stands for voltage signals connected via capacitors a=
nd DC stands for voltage signals connected directly or through a resistor t=
o a voltage source. Capacitively coupled AC signals (like the connections b=
etween Nv stages) have a characteristic that fast changing waveforms can ea=
sily pass through capacitors but slow changing or static voltages cannot. R=
esistors are used for connecting dc signals (in combination with capacitors=
to ground (ie Nu neurons) to delay and resist rapid change and instead per=
mit slow or unchanging voltage signals to pass through.

NO GAIN - NO BRAIN

The bicore and PS circuits use a HC240 octal inverting buffer, which as =
you will recall has 8 inverters, organized in two separately enabled groups=
of 4 inverters. Each inverter is an amplifier with one input pin and one o=
utput pin. The logic function of the inverter is to make the output pin +V =
when the input pin is 0V and make the output 0V when the input is +V. In ot=
her words the state of the output pin is opposite to the state of the input=
pin. Each group of 4 inverters is controlled by a tri-state enable pin whi=
ch when held high (+V) will disable (float) the corresponding group of 4 in=
verter outputs by isolating the output pins. The output pin of each HC240 i=
s the midpoint of a pair of series connected N channel and P channel mosfet=
s. When the n channel device is on the output is low and with the P channel=
device on the output is high. When both devices are turned off the output =
floats. Aside from the inverting logic function, the inverting amplifiers a=
lso have "gain" which causes a small voltage change around the switching th=
reshold of the input pin to result in a large voltage change on the output =
pin. This is very important since many of the "passive" components reduce, =
attenuate or slow down voltage waveforms and it is the job of the amplifier=
s to compensate for the losses and speed up the waveforms.

TRUE OR FALSE

We mentioned earlier that the voltage on input and output pins can be +V=
and 0V logic levels. These levels are often referred to by other names suc=
h as "high and low" or "1 and 0" or "on and off" or "true and false". These=
terms are used interchangeably and have the same meaning of identifying on=
e of two binary states.

MAKING LOGICAL WAVES

The term square wave, rectangular wave or pulse will be frequently used =
in the description of these circuits. They graphically describe the two log=
ic levels with horizontal and vertical lines. A squarewave uses a lower hor=
izontal line representing 0V which rapidly changes to and continues along a=
n upper horizontal line representing +5V then drops down again to continue =
the lower horizontal line etc. The vertical position of these lines corresp=
onds to the high or low logic levels and the length of the horizontal lines=
represent the duration of each logic level.

Square waves have upper and lower line segments which are the same lengt=
h, meaning the on and off times are the same duration. Rectangular and puls=
e waveforms have unequal on/off times.

MAKING ANALOGICAL WAVES

In contrast to binary waveforms, which change rapidly from high to low e=
tc, other waveforms change relatively slowly. Such waveforms are described =
by graphical terms like triangular or exponential waves. One feature of ana=
log waveforms is that they are records of analog voltage levels, which can =
be any voltage between 0V and +V.

The figure shows several cases of bicore circuits and the waveforms at t=
he inverter inputs. Analog waveforms generally only occur at the input pins=
of this class of astable oscillators. With voltage gain or amplification t=
hrough the inverter gain stages, the input voltages are amplified and the o=
utput voltage waveforms switch rail to rail (0V or +V) depending on whether=
the input voltage is above or below the input trigger or threshold voltage=
. When a slowly changing input waveform voltage crosses the trigger it caus=
es the output to change much more rapidly. This effect of voltage gain is c=
alled clipping or limiting amplification: changing analog waveforms back to=
digital waveforms.

HASTE MAKES NO WASTE

Moderately slow waveforms are restored by the gain of the amplifiers bac=
k to binary waveforms and for digital circuits, such as the 74HC240, these =
speedy binary waveforms waste little or no power quite unlike their slower =
moving cousins, the "linear" waveforms. Extremely slow waveforms, ie the ri=
sing voltage on a capacitor being charged by a solar cell, cross the trigge=
r level so slowly that the output voltage, while changing more rapidly than=
the input, will still spend a significant time in a state between the +V a=
nd 0V levels. With the input voltage in a narrow range around the threshold=
voltage (about 1/2V+ for a (HC240), the resulting "in between" output stat=
e is said to be in the "linear region". An output voltage in the linear reg=
ion is an undesirable state of affair, which causes high power loss and ele=
ctrical noise upsetting other waveforms in the circuit. There are several s=
olutions to eliminating those linear region output states and they will be =
discussed next.

FEED ME!

Now that we have the inverter basics covered let's look at a more advanc=
ed concept: FEEDBACK.

Feedback is a simple idea with a surprisingly complex range of behaviour=
. The general concept is that if a fraction of the voltage of an output pin=
of an amplifier is connected back to it's own input pin then, since the in=
put pin determines the state of the output, the output pin will literally i=
nfluence itself! Feedback is possibly the most important concept in electro=
nics, neural networks and general control theory. There are two kinds of fe=
edback: negative and positive. Negative feedback is an (inverted) output si=
gnal of an amplifier which, when fed back to it's input, opposes the presen=
t state of the output. An amplifier with negative feedback will be stable o=
nly when the output is equal to the input, which occurs at the switching th=
reshold (1/2V+) Positive feedback is an (non inverted) output signal which,=
when fed back to it's input, promotes or reinforces the present state of t=
he output. Since the output and input levels are the same (not inverted), a=
n amplifier with positive feedback is stable only when the output and input=
levels are at 0V or +V (high or low).

There are numerous examples in nature of positive and negative feedback =
but for simplicity I will use the lowly balance to illustrate DC feedback. =
The figures below graphically shows the three different cases of applying a=
force (a steel ball) on a balance beam which is suspended (at its fulcrum =
of rotation) at the center of gravity (no feedback) or above the center of =
gravity (negative feedback) and below the center of gravity (positive feedb=
ack). The arrows show the direction in which the ball is pushed and the eff=
ect of gravity when it crosses the "threshold". Below that, the equivalent =
circuits are shown with sample input and output voltages showing the effect=
of the input voltage on the output state.

It is quite likely that a circuit will contain several stages of inverte=
rs or amplifiers connected in series, with the output of one connected to t=
he input of the next stage. Feedback occurs when a sample of the output of =
the last stage is fed back to the input of the first or intermediate stages=
. In that case, an even number of inversions in the feedback path results i=
n non-inverting positive feedback while an odd number of inversions causes =
overall inverting or negative feedback.

While we have just scratched the surface layer of a very deep subject, t=
he circuit descriptions should make a little more sense when referring to t=
he various technical terms in the tutorial.