CHART recorders capable of registering signal frequencies of more than five cycles per second are difficult for amateurs to build. Such instruments are found in electrocardiographs, electroencephalographs and various kinds of machines that record telemetry signals. The response of recorders to frequencies of more than a few cycles per second has traditionally been achieved by minimizing both the mass and the friction of the moving parts. The parts of the motor that drive the pen weigh only a few grams, and the supporting shafts turn on accurately wrought pivots or on jewel bearings. Making such mechanisms requires the skills of a watchmaker.

J. Barry Shackleford, a simulation analyst with the Computer Sciences Corporation in Huntsville, Ala., has turned to an alternative scheme for achieving the desired high performance. He built a relatively crude pen motor that can be made inexpensively at home, and he applied power to it with transistor amplifiers that include feedback circuits to compensate for the mechanical deficiencies of the motor. Although the performance of Shackleford's recorder does not equal that of most commercial instruments, it is adequate for experiments of many kinds. Shackleford describes the construction as follows:

"Chart recorders consist of three principal subassemblies: a motor that operates the pen, a transport mechanism that advances the paper chart under the pen at a known rate and an electrical circuit that applies signal current to the pen motor. I made two motors that are structurally similar but employ different measuring devices: a torsion spring in one case and an electronic feedback mechanism in the other.

"Both motors have a coil of fine copper wire that is free to rotate about its vertical axis between the poles of a permanent magnet. The coil of one of the motors is supported at its top and bottom by a fine music wire in tension; the wire acts as a torsion spring [see illustration at left]. An electron current applied to the coil generates a magnetic field.

"The torsion fiber is adjusted initially to suspend the coil in the position in which the induced magnetic field makes a right angle with the field of the magnet. The induced field interacts with the permanent field to create a mechanical force that tends to rotate the coil to the position where the directions of the two fields coincide. The torque increases in proportion to the electron current in the coil. The torsion fiber exerts a counter-torque that increases in proportion to the excursion of the coil from its zero, or initial, position. The motor is essentially a galvanometer of the type developed in the l9th century by the French physicist Arsène d'Arsonval. It is similar in principle to the motors in most voltmeters and ammeters except that a torsion fiber has been substituted for the pivots, shaft and hairspring of conventional meters.

"No advanced skills are needed for making the motor. The form on which the coil is wound consists of a framework of 1/16 inch brass tubing. The arm that supports the pen is also made of this tubing, which weighs about a third of a gram per inch. A two-inch length of tubing forms the spine of the framework.

"To the spine I solder a pair of U-shaped brackets that face in opposite directions. The coil is wound over the brackets. The brackets are an inch wide and 3/8 inch deep. Brass tubing of this diameter is available in hobby shops, particularly those that stock model-airplane supplies.

"Cut the tubing by placing it on a flat piece of cardboard and sawing it with the sharpest knife in the house. Let the tubing roll freely under the knife so that the resulting groove encircles the brass. The wall of the tubing is thin. Use moderate pressure to avoid collapsing the tube. You will have to resharpen the knife afterward. Burrs can be removed from the open ends of the tubing with a small twist drill.

"I join the parts with solder. To make the joint with a minimum of solder I firs clean the brass at the joint and tin each part at this point with a small soldering iron. After the parts are assembled and clamped in the desired position I place a small bead of solder on the tinned parts and heat the joint with a propane torch adjusted to produce the smallest possible flame.

"Having assembled the framework, I add the torsion fiber, which is a straight length of music wire about .030 inch in diameter (Brown and Sharp gauge No. 20). The E string of a violin would do. Thread the wire through the spine of the framework and solder it in place at the ends of the spine. Use an acid flux and little solder to make the joint. Heat with a small soldering iron to avoid spoiling the temper of the wire. Remove excess acid flux with rubbing alcohol and apply a thin coat of plastic cement to the brass.

"The coil is wound with 250 feet of 28-gauge enameled copper magnet wire. I wrap the wire on the framework randomly, making no attempt to apply uniform layers. The resistance of the resulting coil is about 15 ohms. The ends of the coil terminate in flexible leads of 24-gauge wire. Solder the inner flexible lead to the end of the magnet wire, insulate the joint with a dab of plastic cement and tie it to the spine with silk thread before winding the coil. Terminate the outer end of the coil similarly and with thread lash both leads to the spine about 1/8 inch from the winding. The coil assembly will weigh some 25 grams.

"The arm that supports the pen is a length of 1/8-inch brass tubing. Solder the arm to the spine parallel to the axis of the coil. Add a small brace of the same tubing to stiffen the joint.

"Incidentally, the frequency response of the pen motor varies inversely with the length of the arm. Arms up to eight inches long can be used to record frequencies of up to two or three cycles per second, such as earthquake waves and microseisms. For recording heartbeats or brain waves the length of the arm from the center of the spine to the nib of the pen should not exceed three inches, including the hinge assembly that supports the pen.

"The hinge is made of brass tubing of two diameters: 1/8 inch and 3/32 inch [see illustration at right]. These sizes telescope to make a sliding fit. A 1/4-inch length of the 1/8-inch tubing is soldered at a right angle to the outer end of the pen arm. This tubing serves as the bearing of the hinge. A 5/8 inch length of the 3/32-inch tubing is slipped through the bearing and centered to form a shaft. Two collars, each 3/16 inch long, are cut from the 1/8 inch tubing and slid over the protruding ends of the shaft. A yoke for supporting the pen is bent from 1/16-inch tubing.

"The frequency response varies inversely with the inertia of the writing system. I prefer a light pen of the capillary type, such as the No. 0 or No. 1 Rapidograph. These pen tips are available from dealers in drafting supplies. The tip has a cylindrical ink reservoir that terminates in a cone from which a capillary tube extends to the paper. The flow of ink is regulated by a fine wire that extends through the capillary to the paper.

"The top of the wire terminates in a cylindrical weight that adds inertia to the system. I cut off and discard the weight, slip the wire into the end of a length of 1/16-inch brass tubing and fasten it by crimping the tubing. The upper end of the tubing is bent at a right angle to rest on the upper edge of the ink reservoir. The bend is made at a point such that the lower end of the wire extends to the tip of the capillary. The substitution of brass tubing for the cylindrical weight not only reduces the inertia of the system but also increases the ink capacity of the reservoir.

"The yoke that supports the pen is made by wrapping two turns of brass tubing around the reservoir. The ends of the tubing extend at an acute angle from the loop thus formed. At the point where they are separated about 1/2 inch bend the ends parallel and slip them into the sockets of the hinge as shown in the illustration. Roger Hayward, who illustrates this department, suggests the substitution of a short strip of thin sheet metal, such as brass shim stock, for the hinge. The flexible strip would be lighter than the hinge and easier to make. I intend to try it.

"The coil and the pen arm are supported by a vertical bracket that has a facing pair of adjustment screws to which the torsion wire is attached. The wire passes through a small hole near the end of each screw and is soldered in place. The screws are tightened to exert a pull of about 10 pounds on the wire. The permanent magnet is clamped to a block of wood fixed to a wood base.

"The coil is centered between the poles at the most intense region of the magnetic field. The field intensity can be examined visually by covering the magnet with a sheet of cardboard and dusting the surface with iron filings. With the magnet clamped in position, rotate one of the screws attached to the torsion wire a half-turn. Rotate the other screw in the opposite direction until the axis of the coil makes a right angle with the magnetic field. Tighten the nuts. This adjustment stresses the wire in both torsion and tension.

"My instrument has a surplus magnetron magnet that is available from the Edmund Scientific Co., Barrington, N.J. 08007 (the catalogue number is 70,571). The gap between the poles is 1-1/4 inches and the width is 1-3/4 inches. The strength of the field is rated at 2,075 gauss. The pen motor can be operated from any amplifier that develops 15 watts and has an output impedance of approximately 15 ohms.

"My second motor is similar to the d'Arsonval type but develops much greater torque. It is of the servo type: the coil is mounted on the shaft of a potentiometer that develops a reference voltage. The reference voltage is fed back to the input of a differential amplifier. The feedback system replaces the torsion springs of the d'Arsonval instrument and stabilizes the performance. For example, variations in friction between the pen and the paper have little effect on the response of the system.

"The potentiometer must be of the low-friction type, preferably made with ball bearings. A suitable one is the Helipot 6502, which is rated at 10,000 ohms. The resistance of the potentiometer must not exceed 20,000 ohms. Precision single-turn potentiometers of the required kind are available from American Design Components, 39 Lispenard Street, New York, N.Y. 10013.

"A two-inch length of brass tubing that makes a tight fit with the shaft of the potentiometer becomes the spine of the framework on which the coil is wound [see illustration at left]. U-shaped brackets of 1/16-inch brass tubing are soldered to the spine, as in the framework of the d'Arsonval pen motor. The framework is insulated with plastic cement, wound with an identical coil and supplied with a pen arm. The potentiometer is mounted shaft end up to a wood base. The permanent magnet can be clamped to a block of wood attached to the base.

"Connect the potentiometer to a power source of about 15 volts. Connect a voltmeter between a terminal of the source and the wiper arm of the potentiometer. Rotate the shaft of the potentiometer to the position where the meter indicates exactly half the source voltage. Without disturbing the position of the shaft, slide the spine of the coil assembly over the shaft so that the axis of the coil is at a right angle to the magnetic field.

"Although the pen servomotor is more powerful than the d'Arsonval type, it cannot drive a high-friction writing device such as a pencil or a crayon. Pen tips of hard felt or porous plastic work well. It pays in terms of frequency response, however, to minimize the mass of the assembly, particularly at the outer end of the pen arm. When I use a felt pen, I cut away most of the body and save only the tip and the ink reservoir. A counterweight is not ordinarily required in the pen servomotor.

"Functionally the electronic circuit of the recorder includes three amplifiers that are interconnected to form a close loop. The signal to be recorded is applied to one input terminal of a summing amplifier [see illustration at right] The summing amplifier can accept two input signals. It multiplies their algebraic sum. The output of the summing amplifier is fed to the input of the power amplifier. The power amplifier drives the pen motor. The potentiometer, which is linked mechanically to the shaft of the motor, develops a voltage that varies in amplitude and polarity with the position of the shaft.

"The potentiometer voltage, which is known as the feedback voltage, is applied to one input terminal of a differential amplifier. The differential amplifier can accept two input signals. It multiplies their algebraic difference. The output of the differential amplifier is applied to the second input terminal of the summing amplifier, thereby closing the loop.

"An adjustable reference voltage developed by a potentiometer connected to the power supply is fed to the second terminal of the differential amplifier. In the absence of a signal to be recorded the amplified difference between the feedback voltage and the reference voltage is applied through the summing amplifier to the power amplifier. Energized in this way, the pen motor rotates in a direction that reduces the difference between the feedback voltage and the reference voltage to zero, thereby stopping the motor. If the reference voltage is altered manually, the resulting difference voltage again causes the motor to seek a position at which the difference is reduced to zero. Thus the pen of the recorder can be positioned at any desired point on the chart by adjusting the reference voltage.

"A signal applied to the summing amplifier similarly energizes the power amplifier. The motor then seeks a position where the sum of the signal voltage and the difference voltage falls to zero. Voltages in the feedback loop are automatically amplified, with the result that the system tends to apply as much power as the motor requires to move the pen, regardless of variations of friction between the nib and the paper.

"Although the system has 82 transistors, the construction is fairly simple. All but two of the transistors are included in four operational amplifiers of the integrated-circuit type. The amplifiers are only slightly larger than the eraser of a pencil. They are available on the surplus market for about $1 each from suppliers such as Poly Paks, P.O. Box 942, Lynnfield, Mass. 01940.

"The electronic circuits should be made and tested one at a time. Make the regulated power supply first. Do not omit the fuses. The voltage of the power supply should not vary more than about 1 or 2 percent when a 30-ohm resistor is connected to the output. The resistor should be capable of dissipating about 30 watts.

"Next, build the power amplifier. Check the completed wiring against the schematic diagram [right] at least twice. Remove one of the fuses, apply power and, with a voltmeter, adjust the corresponding bias potentiometer to deliver a potential of .5 volt with respect to ground. Replace the fuse and similarly adjust the bias of the complementary circuit.

"Make an operational test of the amplifier by connecting the output to a dummy load and applying power. I use for the load a small direct-current motor that is rated at 30 volts. Connect a 5,000 ohm potentiometer to the power supply and wire the arm of the potentiometer to the input of the amplifier. If the amplifier is operating properly, the speed of the motor and the direction of its rotation can be controlled by twisting the knob of the potentiometer. Incidentally, all power transistors in both the power supply and-the power amplifier must be mounted on heat sinks that have large cooling fins. Without adequate heat sinks the power transistors will quickly burn out.

"Finally, assemble the differential-amplifier circuit and the summing-amplifier circuit. Do not apply power to this unit until you have triple-checked the completed wiring against the schematic diagram. Failure to make the check can cost you amplifiers, as I have learned from sad experience.

"Test the amplifiers by connecting a voltmeter to the output and applying input voltage by means of a potentiometer wired as a voltage divider. After completing the checks connect the output of the summing amplifier to the input of the power amplifier, connect the pen motor to the output of the power amplifier and connect the arm of the feedback potentiometer to the lead adjustment of the differential amplifier. Center all potentiometers except the bias adjustments. All ground terminals of the system must be interconnected, including the power supply.

"Apply power to the system. Operate the zero-adjustment potentiometer that provides the reference voltage. If the pen cannot be centered on the chart by operating this control, it is probable that the polarity of the potentiometer is reversed. Interchange the connections of the potentiometer and the power supply. Center the pen. With the power on, push the pen from the zero position with your finger and release it. The pen should promptly return to the zero position and stop without overshooting. If it does not return promptly, increase the loop gain to the point at which the pen oscillates and then reduce the gam slightly. Flick the pen. It should overshoot the zero position and oscillate a few times.

"Next, operate the lead adjustment in the direction that arrests the oscillation. In effect, the lead adjustment gives the system the ability to second-guess what is going to happen next. Actually it has the effect of increasing the loop gain as the frequency of the feedback voltage increases and of altering the phase relations of currents in the network.

"The proper relative setting of the lead adjustment and the loop-gain adjustment must be found by experiment. The optimal settings may require considerable tinkering. When the adjustments have been made properly, the pen will snap back to zero and stop without overshooting when it is pushed from the zero position and released.

"This circuit was assembled from components that I happened to have at hand. The values of the resistors were determined experimentally. It is quite possible that the selected values are not optimal in every case. Doubtless the system can be improved by further experimentation.

"Paper-chart transport mechanisms of two types have been built. One of the two carries a chart in the form of a paper cylinder. I use this unit for recording seismograms. The second mechanism pulls a strip of adding-machine tape under the pen. It is used primarily for recording transient signals of higher frequency such as bioelectric potentials. Both units can be made with ordinary hand tools.

"The drum that supports the cylindrical chart was made from a length of phenolic tubing of the kind used for telescope mountings. The material is 1/4 inch thick. It is available from the Edmund Scientific Co. (the catalogue number is 85,146). The ends were closed with disks of plywood. The axle on which the drum turns should be centered as accurately as possible, even though the hinged pen will follow departures from a true circle. Strips of chart paper are attached to the drum with adhesive tape.

"The lead screw that advances the motor carriage across the drum is a long stud bolt of the kind stocked by most dealers in hardware. The drum and the lead screw can be rotated independently by synchronous motors or by belts or gears according to the requirements of the experimenter. For recording earthquakes I prefer synchronous motors [see illustration upper right].

"The transport mechanism for pulling adding-machine paper is somewhat smaller than the drum arrangement and is easier to make. The writing table consists of sheet steel bent to an acute angle and soldered at one edge to an upright plate attached to the base [see illustration at left]. Paper is pulled across the writing table by a rubber roller. The roller is made of heavy-wall rubber tubing slipped over a shaft. A spring-loaded roller, also made of rubber tubing, increases friction between the paper and the driving roller.

"Usually the experimenter is interested in timing the recorded signal. I accomplish this by driving the paper transport with a synchronous motor, an arrangement that automatically provides the graph with a known time base. The experimenter can substitute a variable-speed motor and record time marks and the signal simultaneously. This task can be done by feeding time signals to one input terminal of a summing amplifier and the data signal to the second input terminal. The combined signals would be fed to the input of the recorder."

Bibliography

HANDBOOK OF OPERATIONAL AMPLIFIER APPLICATIONS. Burr-Brown Research Corporation, Tucson, Ariz.

PRINCIPLES OF SERVOMECHANISMS: DYNAMICS AND SYNTHESIS OF CLOSED-LOOP CONTROL SYSTEMS. Gordon S. Brown and Donald P. Campbell. John Wiley & Sons, Inc., 1948.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.