Joystick Metrics

Author: Michael A. Covington
Source: PC Tech Journal, vol 3, No 5, May 1985, pages 99+

The IBM PC is put to some of its most interesting applications when it is asked to handle data on and respond to external physical conditions. Microcomputers can serve in many such capacities: as superintelligent thermostat for heating systems, data collection devices for the laboratory, or automated troubleshooters for electronic equipment. In addition, unlike other kinds of measuring equipment, the computer can store and analyze large amounts of data automatically.

Almost all of these applications require some form of analog-to-digital conversion – that is, a continuously varying quantity, such as a voltage, has to be converted into a computer-readable form. Commercial analog-to-digital converters often cost hundreds of dollars, but the PC’s joystick port (officially called the IBM Game Control Adapter) can be used to do the same job in a simpler way. The circuits discussed in this article can convert the joystick port into a device that will measure resistance, capacitance, or voltage. No modification to the PC or the adapter is required; the circuits all plug in externally in place of the joystick.

These circuits have not been tested with the PCjr, but they should be fully compatible with it. The PCjr BASIC and Technical Reference manuals indicate that, with the exception of the pin connections, the PCjr‘s built-in joystick port is identical to the PC Game Control Adapter. Performance with PC compatibles is hard to predict, but if IBM’s Game Control Adapter card can be used in the computer, chances are good for full compatibility in this area.

An IBM PC joystick consists of two potentiometers at right angles to each other, plus two fire buttons; for simplicity, only the potentiometers will be considered here. The computer senses the position of the joystick by measuring the resistance of each potentiometer, which varies from 0 to 100,000 ohms. This is done by timing how long it takes a capacitor of known value to charge through the unknown resistance.

FIGURE l: Joystick Port Circuit (Simplified)

The joystick position is read by moving the short circuit across the capacitor and measuring the time required for it to charge to about 3.3 volts.

FIGURE 2: Joystick Port Pin Connections

On the PC and PCjr, pins are numbered as seen by looking at the joystick socket from outside the computer. On the PC, pin numbers are usually shown on the connector. On the PCjr (at right), AA0l and BA0l are the missing pins.

FIGURE 1 shows a simplified version of the circuitry inside the joystick port; FIGURE 2 shows the pin connections. One end of the resistance to be measured is connected to the positive 5-volt supply; the other end is connected, through a 2200-ohm resistor, to the capacitor. Normally, a switching transistor keeps a short circuit across the capacitor to prevent it from charging; the PC reads the joystick position by removing the short circuit and timing how long it takes the capacitor to charge to approximately 5.3 volts. The result is a number between 0 and 255, roughly equal to the resistance in Kilohms.

This result can be accessed in BASIC using the functions STICK(0), STICK(1), STICK(2), and STICK(3), one for each of the four potentiometers (there are two in each of the two joysticks). When the value of STICK(0) is requested, the PC reads the positions of all four potentiometers; when STICK(1), STICK(2), or STICK(3) is requested, the results of readings that were taken simultaneously with the most recent call to STICK(0) are returned.

Listing 1: ANALOG BAS 

The program in LISTING 1 outputs a continuous display of the value of STICK(0) that can be used while experimenting. Only STICK(0) is used in this article, for the sake of simplicity, but remember that the PC can accommodate four copies of each circuit given here, all operating simultaneously.

FIGURE 3: Measuring Resistance  

The unknown resistance need not be a joystick potentiometer; it can be anything with a resistance in the appropriate range. Because resistor and capacitor values inside the joystick port vary among PCs, some preliminary calibration is required.

Naturally, the unknown resistance need not be a joystick potentiometer. It can be a resistor or anything else with a resistance in the appropriate range (see FIGURE 3). To read out the actual resistance in ohms, some calibration is required, since the values of the resistors and capacitors inside the joystick port can vary from one PC to another.


The program in LISTING 2 will perform the calibration. At the beginning of each session, readings are taken of two known resistances, one of which is 0 ohms (obtained by shorting across the terminals where the unknown resistor should go); the other should be a resistor with an accurately known value between 100 and 200 kilohms.

FIGURE 4: Measuring Capacitance  

This circuit is slightly more complicated than that for measuring resistance. Capacitance is measured by assessing the time taken to charge the unknown capacitor in addition to the
capacitor inside the PC. Two calibration values are needed.

The PC can also measure capacitance, but the circuit required is slightly more complicated: it has two components instead of one (FIGURE 4). The idea is to measure the time taken to charge the unknown capacitor in addition to the capacitor inside the PC, The requisite program is shown in LISTING 3.


 Again, two calibration values are required; one is 0, obtained by leaving an open circuit in place of the unknown capacitor, and one should be a capacitor with an accurately known value between about 0.2 and 0.5 microfarads. The program operates over a range of about 0.01 to 0.5 microfarads, with a resolution of about 0.01 microfarads.

To make the PC respond to light level, just substitute a cadmium sulfide photocell for the resistor in the PC ohmmeter circuit. Depending on the characteristics of the photocell, calibration may turn out to be quite complicated. There is probably no one program to suit all cases, but any mathematical formula can be implemented easily on the PC.

FIGURE 5: A Simple Temperature Sensor

The same is true for measuring temperature, which can be done in many different ways. A thermistor—a temperature-sensitive resistor—can be used, but they are often difficult to obtain. An easier alternative is to use a silicon transistor (see FIGURE 5). All transistors are temperature-sensitive to an extent and this sensitivity can be exploited to take measurements.

The fixed resistor shown in the diagram should be chosen by trial and error so that the value of STICK(0) comes out near 100 at room temperature. The potentiometer can then be used for fine adjustment. In most control applications, the PC will be used to detect that the temperature has reached a certain level rather than to make quantitative measurements. Note that the value of STICK(0) decreases with increasing temperature. A resolution of 1-degree Centigrade or better can be expected, depending on the type of transistor.

The most important kind of analog to digital conversion involves voltage, since other analog quantities can be converted into voltages relatively easily. To make the PC measure voltage, it is necessary to convert the unknown voltage into a constant or nearly constant current and use this to charge the internal capacitor; the higher the voltage, the shorter the charging time will be. The voltage will then be proportional to the reciprocal of STICK(0), rather than to STICK(0) itself, and resolution will be best at lower voltages—which is as it should be, since small voltage differences matter most when the total voltage is small.

LISTING 4 is a program that can be used with any of the PC voltmeter circuits given here. It requires calibration from two known voltages, both in the range of the circuit used.


FIGURE 6: PC Voltmeter - Economy Style

This circuit can be used only to measure voltages that are high enough to charge the internal capacitor to the threshold within the allowed time. The useful range is between 5 and 18 volts. It is accurate to within .003 volts.

The simplest way to convert voltage to current is to run it through a resistor; this is the method used in the circuit in FIGURE 6. But this circuit can measure only voltages that are high enough to charge the capacitor to the threshold within the time allowed, which means that the circuits useful range is about 5 to 18 volts. Within this range, it is quite accurate. I assembled a test version and calibrated it at 5 and 12 volts using VOLT.BAS; it was accurate to within 0.03 volts for all values between.

FIGURE 7: PC Voltmeter - Deluxe Style

FIGURE 7 is the deluxe model PC voltmeter. Its useful range covers 0 to about 6 volts; more importantly, no adjustments are necessary—the software can do all calibration because there is so little variation among units.

FIG 7a - Type 324 Quad Operational Amplifier

The circuit, a voltage—controlled current source, uses a type 324 quad Operational amplifier integrated circuit (op amp IC). The output of the first stage is the same as the input voltage. Resistors R1, R2, and R3 form a summing network. A small, constant bias is added to the input voltage so that 0 volts input will not give 0 output current (if it did, the PC’s capacitor would not charge and it would be impossible to take a reading). The second and third op amps, the two transistors, and resistors R4, R5, and R6 constitute the current source itself. The voltages across R4, R5, and R6 are held equal; the output current is equal to this voltage divided by the value of R6.

The input and output voltages of the type 524 op amp can swing down all the way to 0 volts, making a negative power supply unnecessary. (For this reason, only the 524 or an exact equivalent can be used in this circuit.) However, a problem is posed by the upper limit of the 324‘s output voltage, which is 1.5 volts below the positive supply. If the 324 were powered from the 5—volt supply, its output could not go above 5.5 volts, and the capacitor in the PC would not charge within the time allowed. The solution is to power the 324 (and obtain R2’s bias voltage) from a higher-voltage power supply. In FIGURE 7, this is shown as a 9-volt battery, but any source of between 7.5 and 30 volts DC will do; the current required is less than 1 mA If this circuit is built on the prototyping area of the PC Game Control Adapter Card, the PC‘s positive 12-volt supply can be tapped.

FIGURE 8: Substitute For 9-Volt Battery In Figure 7

This circuit allows the voltmeter to be run from the 5-volt power available at the joystick pins. The ICL 7660, operating as an oscillator at about 8kHz, feeds a voltage doubler and
outputs about 9 volts from the input.

FIGURE 8 shows a more subtle solution that enables the whole circuit to run off the 5-volt power available at the joystick port pins. An ICL 7660 voltage converter IC, operating as an oscillator at about 8 kHz, feeds a voltage doubler to give about 9 volts out from 5 volts in. The output voltage is slightly higher if a type 1N54 or equivalent germanium diodes are used, although silicon signal diodes (1N9l4, 1N4148) are suitable.

The component values in the deluxe PC voltmeter are not highly critical, but it is good to use 5—percent resistors of the values specified. R4 and R5 should be well matched. R6 controls the range of readings obtained; adjust it to get STICK(0) to equal about 200 or 220 for 0 volts input, and below 50 for the highest voltage to be measured. The transistors can be any small signal silicon type (one NPN, one PNP), preferably with gain (beta) above 50. The circuit is designed to virtually cancel out variations in transistor gain.

Used with VOLT.BAS, the deluxe model PC voltmeter performs impressively. Using a breadboarded version of this circuit, 1 obtained 0.01-volt resolution for readings under 1 volt, 0.03-volt resolution up to 2 volts, and 0.1-volt resolution up to 4 volts. After calibration at 0 and 5.2 volts, the deluxe PC voltmeter agreed with my Micronta digital multimeter to within 002 volts, resolution permitting, over the entire 0-to-5-volt range—I am now not certain which of the two is the more accurate!

One precaution: the input of the 324 sources a small current that can be as high as 500 nanoamperes, though 20 nanoamperes is more typical. This can distort readings of extremely high impedance voltage sources, such as the charge on a capacitor. A 1-megaohm resistor across the input greatly reduces this effect and makes possible readings through a standard oscilloscope probe in either the 1:1 or 1:10 configuration. Alternatively, the input voltage range can be extended with a voltage divider.


OSCILL.BAS (LISTING 5) plots a graph of voltage against time, functioning like a chart recorder or slow, single-sweep oscilloscope. The front end is exactly like the PC voltmeter and can be used with any of the same circuits, but the results are displayed as a graph of voltage against time. A sweep takes about nine seconds and could be slowed down by inserting delay loops. (The PC takes the same time to evaluate STICK(0) whether the capacitor charges slowly or quickly; if this were not so, the joystick would be more responsive at one end of its range).

FIGURE 9a: PC Oscilloscope Circuit, 1Hz Square Wave

FIGURE 9b: PC Oscilloscope Circuit, Capacitor Charging Curve

These figures display the output from OSCILL.BAS (LISTING 5) taken with the
deluxe model PC voltmeter circuit, illustrated in FIGURE 7. The top figure shows
a l-Hz square wave; the lower figure shows the charging curve of a capacitor.

The drawings in FIGURE 9 show the PC oscilloscope‘s interpretations of, respectively, a 1-Hz square wave and the charging curve of a capacitor, taken with the deluxe model PC voltmeter circuit. The bottom figure shows how the resolution is highest at low voltages.

This PC oscilloscope program is only the beginning. With proper programming, the PC can do far more than any ordinary oscilloscope. It could generate logarithmic scales on one or both axes, mathematically transform the data before displaying it, automatically cut out long, monotonous stretches from the waveform, store the readings, or even plot calculated values against measured ones on the screen. The PC costs little more than a good chart recorder and is a lot more versatile. It can be the most powerful piece of equipment in the laboratory.

Michael A Covington conducts research in artificial intelligence and supercomputer applications at the University of Georgia.

Content created and/or collected by:
Louis F. Ohland, Peter H. Wendt, David L. Beem, William R. Walsh, Tatsuo Sunagawa, Tomáš Slavotínek, Jim Shorney, Tim N. Clarke, Kevin Bowling, and many others.

Ardent Tool of Capitalism is maintained by Tomáš Slavotínek.
Last update: 08 May 2024 - Changelog | About | Legal & Contact