From BEAM Robotics Wiki
AKA: Grey's Turtles
Over fifty years ago W. Grey Walter started building three wheeled, turtle like, mobile robotic vehicles. These vehicles had a light sensor, touch sensor, propulsion motor, steering motor, and a two vacuum tube analog computer. Even with this simple design, Grey demonstrated that his turtles exhibited complex behaviors. He called his turtles Machina Speculatrix after their speculative tendency to explore their environment.
The links at the bottom of this page will provide more info about the history of the robots as well as some pictures.
 Wilf Rigter's Machina Speculatrix Analysis
A brief circuit analysis by Wilf Rigter - 29 May 2004
 Foreword to the Current Revision
For an introduction to Grey Walter's Turtles read the articles in the links at the end of the article
Thanks to Domenico Mancini for adding the excerpt from the book ""The Living Brain", which explains Speculatrix in detail. This required some small revisions of my own description to correct assumptions such as the HT supply voltage used being 45V not 90V and component values. I have also added an explanation of details like the battery charger interface which is shown in working principles diagram (See Figure 3 below). In addition I have changed some of the terms and names I used to make the language of article more consistent with Grey Walter's original description.
On the whole my initial analysis of Speculatrix operation based on a fuzzy schematic with no component values and a short description of the operating modes (Gasperi) was satisfyingly close to Grey Walter's own explanation.
As a compliment to Grey Walter's own description in the excerpt from "The Living Brain", I hope in the following few pages to shed some additional light on the circuit with a description of operation Machina Speculatrix circuit shown in the partially annotated Figure 1.
The deceptively simple circuit of two sensors, two vacuum tube "neurons" and some RC components and two relays is responsible for the turtle's complex physical behavior. Actually the complexity is in part due to the recursive mechanical feedback loop between the turtle motors, the physical environment (light and obstacles) and the circuit. The two motors control the front wheel assembly. One is used for driving the front wheel in one direction of rotation and the other motor is used for steering the front wheel assembly around its vertical spindle in one direction of rotation. When rotating the Steering motor has the effect of constantly changing the angle and direction of the turtle motion.
The two form C relays alternately connect the motors either directly to the battery or through a current limiting resistor in parallel with an incandescent lamp to the battery. This provides two possible speeds for motor rotation. In addition, the lamp indicates when the steering motor is rotating.
The SEARCH mode is the normal operating condition of the circuit with Tube 1 biased ON and Tube 2 biased OFF. As a result Relay K1 contacts connect half current to the Drive motor and Relay K2 contacts connects full current to the Turn motor. The Steering motor continuously turns the vector of the front drive wheel and the photo cell. The result is a spiral motion of the turtle with the aperture of the photo cell scanning the horizon like rotating radar dish.
The MOVE mode occurs when, exposed to a moderate light level, the photo cell partially discharges C1 which lowers the Tube 1 grid voltage. This dip in grid voltage is amplified by Tube1 and appears as a positive output signal on Tube1 anode. Note also that the input signal is small and Tube 1 anode current remains above the Relay K1 drop out current.
This positive output signal is capacitively coupled through C2 to the grid of Tube 2. If this signal exceeds the turn-on threshold of Tube2, it is further amplified and energizes Relay K2 coil.
Relay K2 contacts change state and the Steering motor stops while the Drive motor is connected to full current. With the photocell rotation stopped pointing to the light, the turtle now moves in a straight line in the direction of the light source.
This MOVE motion continues for the duration of the R4/C2 time constant until the voltage at the Tube 2 grid drops the anode current below the Relay K2 holding current. Relay K2 contacts reconnect the Steering motor to full current and the Drive motor to half current.
This returns the turtle to the SEARCH mode. As photo cell turns away from the light source, the Tube 1 control grid voltage rises and turns Tube 1 further on. The negative pulse on the Tube 1 anode coupled through C2 further turns off Tube 2. As the photocell turns the same light source, now closer, comes into view again and the process repeats.
The DAZZLE mode occurs when the light level on the photocell is high enough to deeply discharge C1. Then Tube 2 anode current drops below the Relay K1 holding current and Relay K1 contacts open to connect half current to the Steering motor. At the same time the positive transition on the Tube 1 anode turns on Tube 2 which causes Relay K2 contacts to connect the full current to the drive motor. This is the first part of the DAZZLE mode in which the Drive motor rotates full speed and the Steering motor rotates half speed which causes the turtle to veer away from the light. The turtle moves rapidly away from the light in a slow curve. The second part of the DAZZLE mode occurs when the photo cell turns away from the light and C1 charges up again. When the Tube1 grid voltage rises it turns on Tube 1 and Relay K1 changes state. Importantly, the large negative transition on Tube 1 anode, coupled through C2, turns off Tube 2 which drops out Relay K2. This large negative signal drives Tube 2 deep into cutoff and ensures that during the following few seconds Tube2 will not respond to small positive pulses on the Tube1 anode resulting from periodic illumination of the spinning photocell. In this second part of the DAZZLE mode the Steering motor rotates full speed and the Drive motor turns half speed making the bedazzled turtle do small random spirals unaffected by the light for a short period until it again responds to light and enters the normal SEARCH mode.
The TOUCH mode occurs when the collision contact closes, which connects Tube2 output to Tube1 input through C3, R7 and R8 forming a cross coupled oscillator. In BEAM robotics this network connection would be called a two Nv neuron grounded bicore. The two inverters are connected in a loop through differentiating R/C networks and this positive feedback causing the circuit to oscillate between SEARCH and DAZZLE modes. The duty cycle of oscillation is 1/3 but is affected by the light level.
The LOW BATTERY MODE can occur when circuit bias conditions are chosen so that a low motor battery voltage will result in the MOVE mode without the DAZZLE termination and hopefully results in the turtle entering its brightly lit hutch for a battery charge. The battery charger interface cleverly uses two additional relays to disconnect the motor battery from the rest of the circuit when to protect the battery from over discharge, to turn of the motors and rest of the circuit during charging and to reconnect the battery to the circuit when fully charged.
 THE TECHNOLOGY
Although bulky and not very efficient by today's standard, vacuum tubes are similar to N-channel JFET transistors. Like their solid state counter part, they are normally conducting and require a negative control grid (JFET gate) voltage with respect to the cathode (JFET source) in order to turn off the anode (JFET drain) current.
The vacuum tubes operate as voltage to current converters. The current flows from the hot cathode through a control grid to the positive anode. The control grid is used to regulate the anode current. Note that the vacuum tubes used in the turtle circuit are tetrodes which have an extra screen grid that can be used to linearize control voltage curve.
With the screen grid connected to the anode, the tube behaves like a triode. Tube1 is operated as a triode while Tube2 uses the screen grid in an unusual way probably to increase amplification gain and increasing the non-linearity of the Tube2 operating curve.
The vacuum tubes require a low current high voltage (High Tension) supply which in this case is a +45V dry cell connected to the HT+ and HT- terminals. In addition a low voltage supply is used to heat the cathode filament which, in this case, is supplied through a series dropping resistor (R3 and R6) from the rechargeable low voltage (Low Tension) motor battery (probably 6V) connected with the LT+ and LT- terminals.
The vacuum tube anodes drive two relay coils which are connected between the anodes and the HT+ supply battery. The HT- terminal is connected to the cathodes. The motor battery is connected with the LT+ to the vacuum tube cathodes and LT- to GND. This last connection is important to explain the low battery sensing circuit covered later on.
In portable tube circuits, a direct heated cathode filament is used which requires a very low heater voltage. The heater current is supplied by the motor battery and a resistor in series with the cathode filament which drops the heater voltage to 1.5V.
The photo cell is a light sensitive cold cathode vacuum tube that conducts micro amps of current when photons impinge on the cathode, knocking off free electrons which are attracted to the positive anode.
RELAY K1 is the controlled by TUBE1 and RELAY K2 is controlled by TUBE2
The 6V battery is switched directly to motors by RELAY K2 for full speed. RELAY K1 contacts connect the battery through a resistor with a parallel lamp to the motors for half speed operation.
 MODES OF OPERATION
From Gasperi's link below:
"The turtles had four modes of operation: Search, Move, Dazzle and Touch. The first three modes are determined by light level. Dark causes Search mode, moderate light cause Move mode and bright light is Dazzle Mode. Touch causes an oscillation between Search and Dazzle. In Search mode the steering motor is on full and the propulsion motor is on half speed. In Move mode the steering motor is off and the propulsion motor is on full speed. And in Dazzle mode the steering motor is on half and the propulsion motor is on full speed.
At first the Dazzle mode may seem superfluous, but it really helps to create unique behaviors. For one, it prevents the moth drawn to the flame phenomenon. When there are two light sources, the turtle is drawn first to one light and then the other rather than being stuck between the two. When lights are attached to the turtles themselves, they are attracted and then repelled from each other creating a dance something like mating and territorial aggression".
 THE SEARCH MODE
The cathodes of both tubes are at +6V and the control grid of TUBE 1 is connected through a high value resistor (R1) to +6V. This means TUBE 1 is normally ON and RELAY K1 is normally ON.
TUBE 2 control grid is connected through a high value resistor (R4) to 0V which means TUBE 2 is normally OFF and RELAY K2 is normally OFF.
By convention, in schematics relays are always shown in the de-energized state but note that in the search mode, RELAY K1 is normally picked up and its contact causes the drive motor to receive half current. RELAY K2 is normally OFF and its contact causes the turning motor to receive full current.
RELAY K1=ON, RELAY K2=OFF
This mode results in slow small spiral movements of the turtle
 THE MOVE MODE
When the turning motor rotates it also turns the hooded light sensor to scan 360 degrees and as it passes a light source, the photons cause the photo sensor to conduct, discharging the 0.01uf cap, so that TUBE1 control grid voltage pulses toward GND and at TUBE1 anode, a positive voltage transition is generated but not enough for RELAY K1 to drop out.
This positive transition is connected via a 0.5uF cap (C2) to the control grid of TUBE2 which turns on resulting in:
RELAY K1=ON , RELAY K2=ON
This causes a straight motion toward the light.
After a short time, C2 has charged up and TUBE2 input R4/C2 has timed out resulting in a return to the SEARCH MODE
RELAY K1=ON, RELAY K2=OFF
This lasts for one rotation of the sensor until the light comes in view again and the MOVE MODE is resumed.
As the un-turning photo sensor approaches the light, the increasing brightness finally cuts off TUBE1 current which results in the DAZZLE MODE.
 THE DAZZLE MODE
RELAY K1=OFF, RELAY K2=ON
This mode does not last long since the photo sensor is slowly turned away from the light and TUBE1 turns on again to to part two of the DAZZLE MODE which might be called the BLIND SEARCH MODE. This is the same relay logic state as the SEARCH MODE but with one difference:
The large (45V) negative transition on TUBE 1 anode as it turns back ON now turns TUBE2 OFF very deep and holds it off despite the normal small positive transitions pulses on TUBE1 anode as the turning motor rotates the sensor and causes intermittent light exposures. This persists for several seconds and during this time the light is ignored. The result is a fast large spiral away from the light source followed by a period of random small spirals before the negative transition on TUBE2 input decays and the normal photo tropic SEARCH/MOVE MODE can resume.
 THE CAPTURED DAZZLE MODE
If there is a possibility that the photocell is continuously exposed to bright light no matter what direction, the turtle can hang up with the steering motor ON and the drive motor OFF.
 THE COLLISION MODE
The normally open ring shaped collision contact closes when the turtle shell hits an obstacle. This connects the R7/R8/C3 network at the anode of TUBE2 back the control grid of TUBE1. Resistor R8 is referenced to GND and when connected to TUBE2 grid in parallel with the photo cell, R8 lowers the grid voltage and forces TUBE1 OFF. The transition on the TUBE1 anode turns TUBE2 ON which gives the DAZZLE mode and this is followed by the BLIND SEARCH mode as previously discussed. As long as the contact is closed, the turtle oscillates between the BLIND SEARCH and DAZZLE modes. The random motion this generates will eventually extract the turtle from any obstacle or corner.
 THE DISCHARGE MODE
As the battery voltage drops this reduces the bias voltage which normally holds TUBE 2 OFF
At the same time a lower motor battery voltage it makes the photo sensor less sensitive. As a result TUBE1 does not easily turn OFF in very bright light while TUBE2 is easier to turn ON by small pulses on the TUBE1 anode. The DAZZLE mode will not result in the BLIND SEARCH mode and the turtle will dance near the light for a while and eventually enters the MOVE mode but without being terminated by the DAZZLE mode and drives straight into the light and hopefully contacts the charge terminals.
If Speculatrix fails to find the battery charger before the motor battery voltage becomes dangerously low, then the voltage detecting relay (K3) drops out and the negative side of the 6V battery is disconnected from the Speculatrix main circuit. The now dormant turtle will need to be manually pushed into the charging dock in order to charge the battery and reconnect the negative 6V line to the main circuit.
With the 6V motor battery low Speculatrix searches for the light beacon associated with the charger. If it successfully engages the battery to the charger terminals, then the charging current passing through the current sensing relay coil (K4) disconnects the positive 6V battery terminal from the main circuit and motors and stops Speculatrix. As long as the charging current is high enough, the K4 contact remains open and charging continues. When the battery is fully charged, the charging current tapers off and K4 drops out. This reconnects the positive 6V battery terminal to the main circuit which powers up Speculatrix to drive forward and immediately enters the Touch mode to extracts itself from the charging dock and resume normal mode operation.
 The Last Word
This is the third draft analysis of Machina Speculatrix and needs your comments for corrections and missing details. I especially would like to confirm more component values and the vacuum tube types.
Shown in Figure 2, Speculatrix Redux is an as yet untested JFET + PD circuit, true to the original low complexity "analogic" design which should operate nearly identical to the original circuit. The component values will be chosen based on the original circuit time constants, etc.
 Original Schematics of Machina Speculatrix
 External References:
- The Grey Walter Online Archive
- Michael Gasperi’s Grey Walter Machina Speculatrix Page
- Robot Critters
- The Tortoise Against Modernity: Cybernetics as Science and Techonology, Art and Entertainment By: Andy Pickering
- Exploration and high adventure: the legacy of Grey Walter By: Owen Holland
- Instructions For Constructing Grey Walter’s Tortoise Robot From Lego Mindstorms Components
- A simulator program
- Pages 287 to 292 of Grey Walters book "The Living Brain"
- A Video about Grey Walter's tortoises