This text is based on the article that Wouter Brok (a fellow Dutchmen) wrote, with some additions by myself. If any of the following seems not correst to you, please contact me. The full article can be viewed on the Beam-online site. If any of the following is not entirely clear to you, don’t hesitate to contact me, and I’ll try to explain.

THE SUSPENDED BICORE

x(For dummies)

The suspended bicore is the core of most BEAM devices. It’s unique structure makes it very easy to customize for a wide range of uses. These uses vary from blinking LED’s to servo-drivers to walkers.

NEURONS

The heart of core technology is the use of neurons. Neurons are most common in living creatures. It supplies the creature with information processing abilities, caused by the loads of neurons connected together.

The difference between a neuron and a transistor is that the transistor has only a “yes” or “no” if used in logic circuits. It it’s input is directly related to it’s output, and therefore it is not very suitable for circuitry with a certain “intelligence” of its own.

The bicore is built around Nv-neurons, as introduced by M.W.Tilden.The circuit is a pulse-delay-circuit, witch is further described in “Controller for a four legged walking machine”

The triangular thing is an inverter. An inverter “inverts” a logic value ( 1 -> 0 , 0 -> 1 ).

In this case we use an “ideal” inverter, meaning that it’s switching point (threshold-value) is at exactly one half of the voltage difference between “1” and “0”. For example, if ground is 0V and Vcc is 5V, the inverter will switch at 2,5V.

Now imagine, a positive (5V) pulse is given to the “in”, causing the area between the cap and the inverter to be 5V too. The capacitor starts decharging through the resistor. If the voltage between the cap and the resistor passes the 2,5V mark, the inverter will “switch”.

The time between the pulse and the changing state of the inverter is the delay time. It is easy to see that this time is determined by the capacitance of the cap and the resistance value of the resister.

BICORE

Stick two of those neurons together, and voilá, the bicore. It is quite easy to imagine two of these neurons passing the pulse over to one another. If both resistors and both capacitors are equal in value, the bicore will have a duty-cycle of 50% and thus be symmetric.

From now on, all capacitors are equal in value, unless specified otherwise.

SUSPENDED BICORE

To increase the symmetry of the bicore, it is possible to change the resistors for one resistor, connecting the inverters together as shown. It’s easy to imagine that in the original bicore the voltage at the resistors is the opposite of each other. The suspended bicore can be imagined as a normal bicore in witch the one corner functions as a “ground” for the other corner.

To understand the suspended bicore better we are going to look at wat happens with the voltages in the core. The beginning situation is when the sides of the resistor have opposite voltages (Vcc/ground). For this example we take Vcc = 5V and ground = 0V.

Because of the inverters, witch invert the voltage of the opposite side of the resistor, the voltage over the capacitors is 0V. Because the voltage difference over the resistor is 5V, current starts to flow from one end to the other, causing the capacitors to gain a voltage difference. If the voltages reach the threshold-value of the inverter, the inverters start to change state, creating a new voltage difference over the resistor, only with the voltage levels reversed. This is half an oscillation.

This animation might give you a better understanding, blue is 0V, red is 5V, purple is 2,5V. Watch the upper right and lower left corner changing voltage levels.

If we look at the voltage of the outputs of the inverters, they have an almost perfect "block"-oscillation, witch graphically would look like this:

This is the signal used as output of the bicore, wich can drive legs, blink LED's and has many other uses

The replacement of the two resistors of a normal bicore with one as used in the suspended bicore brings up two aspects that need further explanation. At first, if the one corner is used as ground, the voltage will rise at that point till about half the voltage difference between ground and Vcc. If the inverters change state, the voltage will rise with a value op the difference between ground and Vcc, causing the voltage to increase till 1,5 times as high as Vcc (or, if the other corner is viewed, -0,5Vcc).

In designing a suspended bicore, one needs to prevent this effect. 74xxx240 Chips have a system that protects the inverters for over/under-voltage.

The second aspect makes the suspended bicore so incredible. One side of the resistor has the ground voltage, the other Vcc. If the current flows through the resistor, both sides of it will approach 0,5Vcc BUT WILL MATHEMATICALLY NEVER REACH IT, because if the voltage difference over the resistor decreases, the current will decrease as well, causing the change in voltage difference to drop as well.

BUT HOW CAN IT WORK THEN ?

Every circuit has some noise. This noise is a slight fluctuation in the voltages. So if a voltage nears the switching value of the inverter, the noise will cause it to go periodically over it, letting the inverter change state.

This is also the reason why suspended bicores are so often used in BEAM: it responds to the change of noise the circuit produces.

The amount of influence of the noise can be regulated. For this I have to show you some graphics. These graphics show how the voltage oscillates at the input of an inverter.

In the first graphic the noise is drawn with an enlargement to point out how the noise causes the inverter to change state.

It is easy to see that the angle marked with the dot will have a great influence on the effect of noise: a bigger angle will reduce it. In a normal suspended bicore the voltage value will NEAR the threshold-value, so the angle is infinitely small. If you want to increase this angle, the value has to PASS the threshold-value.

Imagine what happens if the capacitors are not equal: Voltage-values will not be equally divided between the two sides of the resistor, instead one side will near a value lower than the threshold-value, and the other will near a value higher than it. To near this higher value it has to PASS the threshold-value, increasing the angle . So creating a difference between the capacitors will make the bicore less sensitive to noise.

The limit of such a bicore is a bicore with only one capacitor. This is the least sensitive variation.

THE MASTER / SLAVE COMBINATION

The master / slave combination is a very good driver for four legged walkers, connecting each bicore to two legs. The master / slave combination is nothing but a suspended bicore with connected to it's input the -normally grounded- resistors of a normal bicore.

The master oscillates unaffected by the slave, while the oscillations of the slave are regulated by the master. The most common master / slave is one in witch the slave oscillates at the same rate as the master. Therefore the "natural" oscillation of the slave has to be around the value of oscillation of the master. The combination can also be designed to let the slave two or more times faster or vice versa. The behaviour of the master / slave is very complex, so I'll only give some characteristics of it.

- The connecting resistors affect the delay between the master and it's slave

- If you want the delay to be longer than half an oscillation time, you'll have to change the value of the capacitors on the slave, because it'll otherwise be to "slow" for the master.

- The resistor in the master bicore determines the oscillation time of the entire bicore, together with the capacitors, but it is most common to change the resistor to adapt the speed of oscillation.

Good luck in BEAMing, and don't let murphy spoil your temper!