Low Power, High Resolution Digitizing System With Cordless Pen/Mouse

The invention relates to digitizing apparatus, especially digitizing apparatus including a tablet with embedded X and Y conductors and a pen, cursor, or other transducer having a conductive stylus or tip that transmits electric signals that are capactively coupled to the embedded conductors in order to produce signal levels in the embedded conductors indicative of the coordinates of the stylus; the invention also relates to a computer input device of a type commonly referred to as a "mouse", and more particularly to a low cost, low power digitizing system including a cordless pen/mouse or a cordless keyboard.

Electronic digitizers which include a horizontal "tablet" having a grid of spaced X and Y conductors embedded beneath a digitizing surface of the tablet, and also a "pen" with a writing tip or stylus that can be lowered onto the digitizing surface, are well known. Some devices of this type are based upon electrostatic, or capacitive, coupling between the conductive stylus or cursor and the digitizing surface. Other systems rely on inductive, or electromagnetic, coupling between the digitizing surface and the pen or cursor also have been utilized. The two basically different design approaches have both been used, to meet different objectives. Systems which use inductive coupling between the pen and the digitizing surface typically must use ten to one hundred times more power than is the case if capacitive coupling approaches are used, and are not suitable for systems in which very low power dissipation is a major objective. U.S. Pat. Nos. 3,444,465, 3,767,858, 3,851,097, 4,034,155, 4,158,747, 4,227,044, 4,492,819, 4,451,698, and 4,491,688 perhaps are generally illustrative of the state of the art.

Presently known systems typically utilize a high level of power, and are characterized by large amplitude pen or grid signals which couple the digitizing surface to the cursor in order to achieve adequate reliability in the presence of relatively high electrical noise levels. Although a cordless cursor device, such as the one disclosed in U.S. Pat. No. 4,451,698, has been proposed, no cordless digitizer stylus structure has been utilized in a low power, low cost digitizing system, especially one of the type utilizing electrostatic coupling between the stylus and the digitizing surface. This is probably the case because of the unacceptably high weight of a self-contained battery power supply, and also because of the absence of a good electrical ground. A good electrical ground ordinarily would be necessary in a system that utilizes electrostatic coupling. The device disclosed in above-mentioned U.S. Pat. No. 4,451,698 is based upon inductive coupling between the cordless cursor and the digitizing surface, and uses much higher signal levels, and dissipates much more power than would be dissipated in a system using electrostatic or capacitive coupling between the pen and the digitizer surface.

There is an unmet need for a low cost, high resolution digitizing device utilizing a stylus which is cordless, and thereby more convenient to use, than one having an electrical cable connecting it to the digitizing tablet, and also having good noise immunity to normally present levels of ambient noise, and also having the capability to function as a mouse as well as a digitizing device.

It is an object of the invention to provide a digitizing device including a cordless electrostatic stylus or cursor that is capacitively coupled to a digitizer surface.

It is another object of the invention to provide a cordless electrostatic digitizer stylus or cursor that also is capable of transmitting commands to a digitizer surface, which command signals are decoded and presented in a suitable format to a computing device.

It is another object of the invention to provide an inexpensive digitizing system that includes a cordless stylus or cursor, and which is operable at far lower power levels than any prior digitizing system.

It is another object of the invention to provide a computer input device that can function as a keyboard, a digitizer, or a mouse.

Briefly described, and in accordance with one embodiment thereof, the invention provides a digitizing system including a cordless pen containing an oscillator that generates a signal that is applied to a conductive stylus which is surrounded by a relatively massive conductive shield. The shield has a large physical size that enables it to function as a good virtual ground and to minimize capactive coupling with the stylus. The relative massiveness of the shield causes it to have much greater stray capacitance than the stylus. This large stray capacitance, together with the small capacitive coupling to the stylus, results in the effective virtual ground and highly directional capacitive coupling from the stylus tip to X and Y conductors of a digitizer tablet. Several pen switches, one actuated by pressing the pen tip against a surface and the others actuated by depressing buttons on the pen, change the oscillator frequency to produce frequency changes that represent a pen down condition or "mouse" commands. A voltage booster circuit including two integrated circuit CMOS inverters and a pair of output capacitors and clamping diodes alternately boost high and low levels of the output of the first inverter and apply them to the power supply terminals of the second CMOS inverter, which then provides a boosted output signal to the conductive stylus. An automatic time out counter that is reset in response to the pen switches to time out after several minutes and effectively disconnects the batteries if the pen is not being used.

The digitizer tablet includes a plurality of substantially spaced major X conductors and major Y conductors beneath the digitizing surface. Between each adjacent pair of major X conductors there are a number of precisely spaced minor X conductors. The voltage difference between the adjacent major X conductors is linearly distributed among the minor X conductors by a resistive voltage divider. Similarly, the voltage difference between two adjacent major Y conductors is linearly distributed between a number of precisely spaced minor Y conductors therebetween. Each major X and major Y conductor is coupled by a resistor to ground. This arrangement of minor grid conductors improves the capacitive coupling to the pen stylus, distributes the signal between major lines linearly and increases the signal-to-noise ratio. A microprocessor controlled scanning system sequentially monitors or "scans" the major X conductors and the major Y conductors by sequentially multiplexing the signal levels thereon through suitable amplifying and filtering circuitry. The amplified output is integrated by a dual-slope integrating circuit and then is converted to a digital output signal, the pulse widths of which represent the pen signal levels coupled onto the major X conductors and major Y conductors. The output pulses are fed into the serial input of a microcomputer chip which measures and stores the output pulse widths during a data scanning operation. Between data scanning operations the microprocessor measures the frequencies of the signals coupled to one of the Y conductors from the pen stylus to determine the frequencies thereof and then decodes the states of the three pen switches.

The microprocessor executes a routine that evaluates the digital data representing the signal amplitudes on the major X conductors, determines the three successive X conductors having the largest signals and the three successive Y conductors having the largest signals in order to roughly determine the X and Y coordinates of the stylus. The microcomputer then executes a routine that inserts the three maximum X signal amplitudes and the three maximum Y signal amplitudes into empirically obtained formulas that precisely extrapolate the precise values of the X and Y coordinates of the present pen position.

In one embodiment of the invention, a separate cordless keyboard includes a ground plane that is aligned with and placed on the digitizer tablet. A plurality of depressable keys are provided on the ground plane. Conductors are provided for routing an oscillator signal to each of the depressable keys, so that capactive coupling between each key and the oscillator signal is greatly increased when that key is depressed. Each key includes a conductive tip that extends through the ground plane nearly to a point on the digitizer surface when that key is depressed, thereby maximizing coupling of the oscillator signal to the surface of the digitizer tablet. The above-described scanning operations are performed by the circuitry in the digitizer tablet to produce the coordinates of the depressed keys. The resulting data produced by the microprocessor in the digitizer tablet then is output to a separate computer, which determines the identity of the depressed key from its X and Y coordinates. Signal producing indicia on the bottom surface of the ground plate are detected by the scanning operation and are evaluated to determine the alignment of the ground plate with the digitizer tablet and to identify the type of the keyboard so that an external computer can select a "menu" that corresponds to the type of keyboard. The depressed key coordinates then are compared to the selected menu to identify the character represented by the depressed key.

FIG. 1 is a perspective diagram illustrating the digitizer system and cordless pen of the present invention.

FIG. 2 is a section view taken along section line 2--2 of FIG. 1.

FIG. 3 is an enlarged view of part of FIG. 1 usful in describing the major and minor grid conductors.

FIG. 4 is a schematic circuit diagram illustrating the circuitry contained in the cordless pen of FIG. 1.

FIG. 5 is a schematic circuit diagram illustrating the basic structure of the CMOS inverters of FIG. 4.

FIG. 6A is a logic diagram of the circuitry contained in the digitizer tablet shown in FIG. 1.

FIG. 6B is a circuit diagram of the graphic module 69 contained in FIG. 6A.

FIG. 7 is a perspective view of the cordless keyboard used in conjunction with the digitizing tablet of FIG. 1.

FIG. 8 is a partial section view of a cordless keyboard of FIG. 7.

FIG. 9 is a diagram showing certain waveforms that are useful in describing the operation of the circuit of FIG. 6A.

FIG. 10 is a flow chart of a program executed by the microcomputer of FIG. 6A.

FIGS. 11A and 11B constitute a subroutine executed in block 103 in the flow diagram of FIG. 10.

FIG. 12 is a flow chart of an interrupt routine executed in accordance with block 110 of the flow chart of FIG. 10.

FIG. 13 is a flow chart of a subroutine executed to compute the precise pen coordinates in accordance with block 105 of the flow chart of FIG. 10.

Referring now to the drawings, particularly FIG. 1, digitizing system 1 includes a digitizer tablet 2 having a plurality of major X conductors 3 embedded therein very slightly below the upper "digitizing" surface of tablet 2, the upper surface being composed of suitable dielectric material. A plurality of major Y conductors 4 also are embedded slightly beneath the top surface. In the presently described embodiment of the invention, there are nine major X conductors 3, each spaced 1.28 inches apart. Each of the major X conductors 3 is connected to a separate input of an X multiplexer circuit 5, which, in response to scanning circuitry in block 7, sequentially couples the voltage on the major X conductors 3 to the input of amplifying and filtering circuitry 7 that determines the coordinates of the tip 10 of a cordless pen 9 and also decodes frequency modulated command signals sent from pen 9. Between each pair of the major X conductors 3 there are five minor X conductors 3A, spaced about 0.2 inches apart. The minor X conductors are not, however, fed into the multiplexer circuit 5. Each of the minor X conductors 3A is coupled to the adjacent one by a high resistance resistor 11. The minor X conductors closest to the adjacent major X conductors also are connected by a resistor 11 to such adjacent major X conductors. Each major X conductor is coupled to ground by a resistor 11A. Thus, the voltage difference between any two adjacent major X conductors is linearly distributed among the minor X conductors by the above-described resistive voltage divider network. An exemplary value of the resistors 11 is about five kilohms, and an exemplary value of each resistor 11A is about twelve kilohms.

The structure of the major Y conductors 4, the minor Y conductors 4A, and the Y multiplexer circuit 8 is essentially indentical to that described above. In the presently described embodiment of the invention, there are seven major Y conductors. The Y multiplexer circuit 8 sequentially applies signal levels capacitively coupled to the Y conductors from the stylus tip 10 of pen 9, which acts as a transmitter. The seven major Y conductors are spaced 1.28 inches apart. There are five minor Y conductors 4A disposed between each adjacent pair of major Y conductors 4. Again, a resistive voltage division network linearly distributes the voltage difference between two adjacent Y conductors 4 evenly among the five minor Y conductors 4A positioned therebetween. Each major Y conductor is resistively coupled to ground.

The general system of X and Y conductors described above is described in further detail in commonly assigned pending patent application Ser. No. 06/631,146, U.S. Pat. No. 4,616,106, entitled "GRAPHIC TABLET AND METHOD", by Fowler, et al., filed July 16, 1984, and incorporated herein by reference.

As previously mentioned, cordless pen 9 (FIG. 2) has a conductive cartridge tip 10 which is capacitively coupled to the grid conductors. Pen 9 includes an internal pen switch 57 that is actuated when stylus tip 10 is pressed against the digitizer surface. Pen 9 also includes two manually actuated buttons 24 and 25 which can be utilized to send "mouse" or tablet "cursor" commands via digitizer tablet 2 and coordinate generating circuitry 7 to a buffer that can be read by a separate computer, such as an IBM personal computer. Pen stylus 10 is simply an ordinary, commercially available ball point refill cartridge 11 having a barrel that extends through an opening in an end piece 12 of the insulative housing 15 of digitizer pen 9. A compressive spring 13 disposed between an inner shoulder of end piece 12 and an enlarged section 14 of ball point cartridge 11 urges cartridge 11 toward the right in FIG. 2.

A thin walled tubular metal shield 16 surrounds cartridge 11 and functions as a virtual ground. The inside diameter of shield 16 is as large as practical in order to minimize capacitive coupling between the barrel of cartridge 11 and conductive tubular shield 16. An exemplary outside diameter of cartridge 11 is about 0.12 inches and an exemplary inside diameter of tubular shield 16 is about 0.22 inches. The writing tip 10 of cartridge 11 extends beyond the left end of tubular shield 16, and acts to directionally focus the electric field toward the embedded conductors in digitizer tablet 2 when the pen tip 10 is held on the surface of digitizer tablet 2.

The right end of pen cartridge 11 is connected to a pen switch assembly including a compressive spring 18 that electrically connects the pen cartridge 11 to a conductor 19 extending through a printed circuit board 17 to continuously connect the output of an oscillator circuit (described with reference to FIG. 4) to cartridge 11 and hence to tip 10. The right end of cartridge 11 actuates a pen switch 20, 57 (FIG. 2) when the tip 10 is pressed against a surface. Pressing tip 10 into end piece 12 in the direction of arrow 26 causes a plunger 57 to move in the direction of arrow 27, contacting a conductor 20 that extends through printed circuit board 17. This closes switch 20, 57 (FIG. 2). Closing of switch 57 (FIG. 4) boosts the frequency of the oscillator from a low idle frequency of approximately 42.5 kilohertz up to approximately 98 kilohertz, which is the desired frequency for the subsequently described scanning operations to determine the X and Y coordinates of pen tip 10. The lower idle frequency results in lower power dissipation, yet enables the X and Y coordinate generating circuitry 7 to determine roughly the position of the pen tip when it is within an inch or so above the surface of digitizer tablet 2.

Switch buttons 24 and 25 in FIG. 2, when depressed, respectively close switches 58 and 59 in FIG. 4 and change the oscillator frequency to different values, which then are decoded as mouse or cursor commands. Suitable conductors on printed circuit board 17 route the switch conductors back to the appropriate points of circuitry 30 of FIG. 4, some of which appears on printed circuit board 17. Other portions of the circuitry 30 are contained in the space 21 of FIG. 2.

The compression spring 18 urges the pen switch mechanism 57 and the ink cartridge 11 to the left more strongly than spring 13 urges it to the right, so that when the pen is lifted from the surface of digitizer tablet 2, the "pen down" switch 57 is opened, and the oscillator returns to the low idle frequency.

Five 1.5 volt batteries are stacked in a cylindrical compartment at the left end of pen 9. A spring 22A provides electrical connection from the positive battery electrode at the left end of the battery stack to the +5 volt conductor of the circuitry shown in FIG. 4, while a suitable contact on a screw-in plug 23 provides electrical connection to the ground conductor 41 in the circuit of FIG. 4.

Thus, the above-described digitizing pen and mouse device is completely self-contained, requiring no cord. Although this approach (without a "fixed" electrical ground for the pen circuitry) would normally be incompatible with transmission of very low level, low power electrostatic signals from pen tip 10 by capacitive coupling with grid conductors in the digitizer tablet, especially considering the rather high level of noise signals that ordinarily would be coupled to and induced in the embedded grid conductors. The present invention overcomes the lack of a "fixed" electrical ground in cordless pen 9 by provision of the tubular shield 16, by providing a structure in which the total stray capacitance associated with tubular shield 16 is at least ten times greater than the total stray capacitance associated with the pen tip 10.

The way in which tubular conductor 16 acts as a virtual ground is by providing a much larger capacitance than the pen tip capacitance, so that the voltage change produced in the shield capacitance by the oscillator output current is very small. This causes a much larger voltage change to be produced across the much smaller pen tip capacitance by the oscillator output current.

Other features, especially the provision of the above-described minor X conductors and minor Y conductors and the above-described resistive voltage divider networks provided therefor, further contribute to the workability of the cordless electrostatic pen structure of the present invention.

Next, the circuitry contained within cordless pen 9 is described with reference to FIG. 4. Pen circuitry 30 includes a oscillator/counter circuit 46, which contains an oscillator 46A including two CMOS buffer 47 and CMOS inverter 48. FIG. 5 shows a typical CMOS inverter. Oscillator/counter 46 can be implemented by means of a CMOS 4521 oscillator/counter integrated circuit. The output of CMOS buffer 47 is connected by conductor 49 to one electrode of capacitor 50 and to the input of CMOS inverter 48. The output of inverter 48 is connected by conductor 45 to a feedback resistor 53. Feedback resistor 53 is coupled by conductor 52 to the other terminal of capacitor 50. Another feedback resistor 51 is coupled between conductor 52 and the input of CMOS inverter 47. The above mentioned idle frequency of oscillator 46 is determined mainly by the time constant of capacitor 50 and resistor 53.

The output of inverter 48 is connected to the input of 24 stage binary counter 46B, the Q.sub.24 output of which is connected to conductor 223.

Conductor 52 is connected by resistor 54 to one terminal of pen switch 57, the other terminal of which is coupled by a resistor to the cathode of diode 60 and to the barrel of pen cartridge 11 (FIG. 2) and hence to the pen tip 10. The idle frequency produced by oscillator 46A is coupled by conductor 45 to the input of a CMOS inverter 43, the output of which is inverted and boosted by a CMOS inverter 33, the output of which is also connected to pen cartridge 11 by conductor 31. Thus, it can be seen that any time oscillator 46A is oscillating, its output frequency will be applied to pen tip 10, regardless of whether any of the three switches 57, 58, and 59 is open or closed.

If pen switch 57 is closed, as a result of pressing pen tip 10 against the digitizing surface, resistor 54 is effectively connected in parallel with resistor 53, increasing the frequency of oscillator 46A from 42.5 kilohertz to approximately 98 kilohertz.

If button 24 is depressed, switch 58 will be closed, effectively connecting resistor 55 in parallel with resistor 53, and increasing the oscillation frequency from the idle frequency to approximately 77 kilohertz. Similarly, if button 25 of the pen 9 is pressed, switch 59 will be closed, in effect connecting resistor 56 in parallel with resistor 53. Thus, it can be seen that there are a total of eight different possible states of switches 57, 58, and 59 which produce different oscillation frequencies. These different frequencies can be discriminated by the circuitry in block 7 of FIG. 1 to represent different commands to an external computer that can be coupled to the digitizing system 1 of the present invention.

The voltage level provided on conductor 37 by the batteries 22 (FIG. 2) is inadequate to generate the constant level of the low frquency AC signal on pen tip 10 that would be desired, considering the level of noise immunity that is desirable. In order to boost the voltage level of the pen signal, which is "transmitted" or capacitively coupled to the X and Y grid conductors, to a higher level, voltage boosting circuitry including CMOS inverters 33, 40 and 43 is utilized. The output of inverter 43 is coupled by conductor 44 to one input of inverter 40, the two power supply terminals of which are connected to ground conductor 41 and to +5 volt conductor 37. The output of inverter 40 is applied to one terminal of each of capacitors 38 and 39. One terminal of capacitor 38 (conductor 34) is connected to the cathode of a diode 36, the anode of which is connected to the +5 volt conductor 37. The cathode of diode 36 is also connected to the positive power supply terminal of CMOS inverter 33. Similarly, the second terminal of capacitor 39 is connected to the anode of diode 42, the cathode of which is connected to ground conductor 41. The anode of diode 42 is also connected to the ground or negative power supply terminal of CMOS inverter 33. Then, whenever conductor 45 is at a high voltage level, conductor 34 will be held at a minimum voltage of at least one diode drop below the +5 volt level of conductor 37. Whenever the output of inverter 40 goes high, diode 36 will become reverse biased, and the voltage on conductor 34 will be boosted by an amount almost equal to the positive voltage transition of the output of inverter 40. This will occur at the same time that conductor 44 is undergoing a negative transition from +5 to ground, so the positive power supply terminal of CMOS inverter 33 will go to about 12 volts when conductor 31 is undergoing a positive transition. This enables inverter 33 to cause conductor 31 to rise to almost 12 volts.

Diode 42 clamps conductor 35 to one diode drop below ground when the output 40A of inverter 40 is high. Similarly, when conductor 40A is at approximately +5 volts, capacitor 39 charges up, as a result of conductor 35 being clamped to ground by diode 42. Then, when conductor 40A goes from approximately +5 volts to ground, conductor 35 is boosted in a negative-going direction to approximately -5 volts, which is applied to the negative power supply terminal of CMOS inverter 33 at the time that conductor 31 undergoes a negative transition. This allows inverter 33 to cause the negative transition of conductor to 31 go to approximately -5 volts.

Thus, the voltage boosting circuit 32 of FIG. 4 nearly triples the amplitude of the output signal produced on conductor 45 by oscillator 46A.

Returning to the portion of FIG. 4 associated with oscillator 46A and counter 46B, switches 58 and 59, corresponding to push button switches 24 and 25 on pen 9, are connected to conductor 45. The cathodes of diodes 61 and 62 also are connected to conductor 45. The Q24 output of 24 bit counter 46B is connected to the set input of an RS flip-flop 219 and also to an inverter 220. The output of inverter 220 is coupled by means of a capacitor to the Q output of flip-flop 219 by means of resistor 221 and also to the reset input of oscillator/counter 46 by conductor 222.

The counter portion 46B of oscillator/counter circuit 46 is a 24 stage binary counter that divides by factor of approximately 16,000,000. At the oscillator idle frequency of 42.5 kilohertz it takes counter 46B approximately three minutes to "time out", at which point the set input of flip-flop 219 is actuated in response to the Q24 output, which is applied to conductor 223, and inverted by means of inverter 224, the output of which is connected to the set input of RS flip-flop 219. This causes conductor 225 to go positive, completely shutting off low current zener diode 226, which causes the regular circuit 227 to produce a precise regulated voltage level of about +5 volts on conductor 37.

Regulator circuit 227 was found to be necessary because the battery voltages can vary by 20% over their lifetime. This is too large a variation to be acceptable for the accuracy of frequencies that need to be generated in the pen circuitry 30. Transistor 228 provides a regulated voltage level of about +5 volts on conductor 37. The resistive feedback to the base of the transistor 228 limits the current flowing through zener diode 226 to about 10% of the total current supplied to the pen circuitry. However, when flip-flop 219 is set as a result of the counter 46B timing out, the resulting low level on the anode of zener diode 226 would cause a continuous reverse current to flow through zener diode 226, greatly decreasing the battery life. Therefore, the flip-flop circuit 219 is necessary to eliminate this reverse current loss when the timer 46B runs out. This causes the voltage to rise in the battery voltage level. The current drain then is extremely small. When counter 46B "times out", it also, stops oscillator 46A, which further reduces the "standby" current drain.

In operation, actuation of any of the switches 57, 58, and 59 resets the oscillator/counter circuit 46. If, during use, the pen circuitry is turned off, the user merely has to tap the pen tip 10 against the digitizer surface or push either of the buttons 24 or 25 to again reset the oscillator counter circuit 46.

The above-described circuitry in FIG. 4 draws a battery current of only roughly 170 microamperes in the standby condition, wherein all of the pen switches are open, 450 microamperes when only pen down switch 57 is closed, and about 600 microamperes when all of the pen switches are closed. This leads to very long battery life, of the order of roughly 500 hours of average pen use.

Referring now to FIG. 6A, the major components of X, Y coordinate generating circuitry 7 of FIG. 1, X multiplexer circuit 5, and Y multiplexer circuit 8 are shown. Multiplexer circuits 5 and 8 are implemented by means of CMOS 4051 eight line analog multiplexer integrated circuits, which are widely available. Its analog inputs are connected to eight of the major X grid conductors 3. Similarly, seven major Y grid conductors 4 and the last major X grid conductor are connected to the analog inputs of multiplexer 8. The line selection inputs of multiplexers 5 and 8 are connected to the Q11, Q12, and Q13 outputs of a counter 71, which is implemented by means of an HC4020 integrated circuit CMOS counter. The selected analog input 3 or 4 is coupled to conductor 70, which produces the SIGNAL IN voltage that is applied to a "graphic module" circuit designated by reference numeral 69, and described in detail with reference to FIG. 6B. The X return and Y return lines are connected to another input of graphic module 69.

Briefly, graphic module 69 amplifies, deletes, and filters the very small SIGNAL IN voltage, integrates it by means of a dual slope integrater, and finally produces a digital DATA OUT signal on conductor 78. The pulse width of the DATA OUT pulse is inversely proportional to the amplitude of the SIGNAL IN voltage (as also explained subsequently with reference to the TIMER DATA IN waveform of FIG. 9).

If the state of an UP/DOWN IN signal applied to graphic module 69 by the Q10 output of counter 71 is a high level, FREQ OUT voltage is produced which is an amplified version of the SIGNAL IN voltage. FREQ OUT is used to enable the circuitry 7 to discriminate the frequency of the SIGNAL IN voltage produced in response to the pen signal during the "frequency scan" part of the system operation, subsequently described.

Coordinate generating circuitry 7 includes a microprocessor 72, which can be a Motorola 6805 microprocessor.

The input select inputs of analog multiplexers 5 and 8 are connected to Q11, Q12, and Q13 of counter 71. The PA6 port of microprocessor 72 is connected to the reset input of down counter 71.

The PB0, PB1, and PB2 ports of microprocessor 72 are programmed as inputs, and are connected to the Q9, Q10, and Q11 outputs, respectively, of counter 71. The signals on these ports are utilized to assure synchronization between counter 71 and microprocessor 72. This synchronization is checked and established in conjunction with block 152 of the interrupt subroutine subsequently explained with reference to FIG. 12.

An external crystal is connected to the crystal control inputs of microprocessor 72 in order to produce a 1,228,800 hertz clock signal. As those skilled in the art know, a Motorola 6805 microprocessor can receive serial data input pulses, and can automatically measure the pulse widths of such incoming pulses by internally ANDing the incoming data stream with an internal timer. The 6805 microprocessor can also count the number of pulses in the incoming data stream, and thereby determine the frequency of the incoming data stream. Both of these capabilities are used in the present system.

The PA0, PA1, PA2, PA3, and PA4 ports of microprocessor 72 are programmed as outputs, and are connected to various resistors as shown to allow microprocessor 72 to automatically adjust the offset control signal on conductor 80, and thereby alter the offset of the RAMP IN signal applied to graphic module 69, in accordance with the subsequently described subroutines executed by microprocessor 72. The PA5 port is programmed as an output and is connected to the GAIN SELECT input of graphic module 69 to allow microprocessor 72 to automatically adjust the gain (range) thereof, as subsequently explained.

Microprocessor 72 has its PA7 port programmed as a data output which, by means of conductor 75 produces a serial data output on conductor 75 that (1) represents the present coordinates of the tip 10 of cordless pen 9 on the surface of digitizer tablet 2, as a result of a data scan operation, or (2) produces data on conductor 75 that represents a "mouse" command decoded from the frequency of the pen as modulated in response to the various combinations of the pen switches 57, 58, and 59. The data on conductor 75 is loaded in serial format into an external parallel buffer (not shown), which then can be read by an external computer (not shown), such as an IBM personal computer.

The DATA OUT signal on conductor 78 is applied to one input of a two input NAND gate 230. The other input of NAND gate 230 is connected to the DATA SELECT output signal produced by the PB6 port of microprocessor 72, which is high during a data scan operation to "enable" the output of gate 230. During a data scan, the DATA OUT signal is inverted by NAND gate 230, and fed into one input of two input NAND gate 231. The other input of NAND gate 231 normally will be high during a data scan operation. Therefore, the signal on conductor 232 will be inverted again, reproducing the DATA OUT pulse on conductor 81, which is applied to the TIMER IN input of microprocessor 72.

However, if a frequency scan operation is occurring, the frequency select signal produced by the PB7 port of microprocessor 72 will be high, and the data select output will be low. This causes NAND gate 230 to be disabled, and two input NAND gate 234 to be enabled. Consequently, the FREQ OUT signal on conductor 235 is fed through two input NAND gate 236, which is enabled by the UP/DN IN signal on conductor 237 and is applied to the other input of NAND gate 234. Consequently, the FREQ OUT signal is reproduced on conductor 238, is fed through NAND gate 231, which is now enabled by a logical "one" level on conductor 232, and appears on conductor 81, and hence is applied to the TIMER/IN input of microprocessor 72, which then counts the number of pulses for a predetermined period of time to determine the present pen frequency.

When the DATA OUT pulse is being fed into the TIMER IN input of microprocessor 72, its pulse width is measured and stored, and represents the amplitude of the X or Y grid conductor presently being multiplexed into graphic module 69.

Although some of the various filters, bypass capacitors, and resistors shown in FIGS. 6A and 6B are not described in detail, those skilled in the art will easily be able to understand their functions and select appropriate values and/or select alternate conventional circuitry to perform the same amplifying, attenuating, by-passing, pull-up functions, etc.

Next, the graphic module circuit 69 of FIGS. 6A and 6B will be described in detail with reference to FIG. 6B. In FIG. 6B, SIGNAL IN (the scanned grid conductor analog signal multiplexed by multiplexer 5 or 8 to the input of graphic module 69) is capacitively coupled from conductor 70 to the input of CMOS inverter amplifier 85 to provide a first stage of amplification. The output of amplifier 85 is capacitively coupled to the input of a similar CMOS inverter amplifier 86, the output of which is connected to one input of an analog switch, which can be implemented by means of a CMOS 4053 integrated circuit. Depending on whether a GAIN SELBCT IN signal on conductor 90, produced by the PA output port of microprocessor 72, is a "1" or a "0", the gain by which the SIGNAL IN level is amplified can be selected to either include or omit CMOS inverter amplifier 87. The resulting amplified and filtered input signal then is applied to the base of a PNP emitter follower 89. The output of detector 89 is coupled to the input of an NPN detector 91, the output of which produces a DATA INPUT signal that is applied to an analog switch 92, which also can be implemented by means of a CMOS 4053 analog switch.

One output of analog switch 92 is inverted by CMOS inverter 93. The output of inverter 93 is applied to the emitter of PNP transistor 94. The base of PNP transistor 94 is connected to the RAMP IN signal conductor. The collector of PNP transistor 94 is connected to a resistive divider circuit including resistors 95 and 96, the junction between which is connected to the base of NPN output transistor 97. The emitter of NPN transistor 97 is connected to ground. The collector of NPN transistor 97 is connected by a pull-up resistor to the +5 volt power supply line, and is also connected to DATA OUT conductor 78.

Analog switch 99 is responsive to the UP/DN signal to "square up" or buffer the RAMP IN signal. The AMP. OUT and DATA INPUT signals of FIG. 6B are connected to a gain adjustment resistor.

Referring now to FIG. 9, the operation of the circuit of FIGS. 6A and 6B will be better understood by those skilled in the art by referring to the various waveforms shown in FIG. 9. The RAMP OUT waveform appears at the output of CMOS inverter 93 in FIG. 6B and on Pin 15 of graphic module 69 as shown in FIG. 6A. As previously mentioned, the circuitry including CMOS amplifier 93 functions as a dual-slope integrater. The positive and negative slopes of most of the peaks of the RAMP OUT signal in FIG. 9 correspond to zero values of SIGNAL IN. The slope indicated by reference numeral 247 corresponds to a medium amplitude of SIGNAL IN. Reference numeral 248 designates a shallower slope of the RAMP OUT signal corresponding to a relatively large amplitude of SIGNAL IN, indicating that the cordless pen tip 10 is close to the grid conductor. More detail on the structure and operation of a dual slope integrator, and conversion of analog grid conductor signals to digital numbers is set forth in commonly assigned U.S. Pat. No. 4,492,819 issued Jan. 8, 1985, entitled "GRAPHIC TABLET AND METHOD", by Rodgers et al., incorporated herein by reference.

The circuitry shown in FIG. 6B converts the RAMP OUT signal to the DATA OUT signal, which is combined with the FREQ OUT signal on pin 10 of graphic module 69 to produce the TIMER DATA IN signal on conductor 81, which then is applied to the TIMER IN input on pin 37 of microprocessor 72. During the grid conductor data scanning operations, the signal level information contained in the slopes of the RAMP OUT signal are converted to the signal level information in the pulse widths of the TIMER DATA IN waveform in FIG. 9. In FIG. 9, most of the pulse widths, such as 121, of the TIMER DATA IN waveform correspond to grid conductor signal levels of zero. The widths of the TIMER DATA IN pulses decrease as the grid conductor SIGNAL IN increases. Thus, the pulse width 122, which corresponds to the slope 247 of the RAMP OUT signal, is narrower than pulse width 121. Similarly, the pulse width 123, produced in response to slope 248 of the RAMP OUT signal, which in turn is produced by a relatively large amplitude of SIGNAL IN, is even narrower than pulse width 122, because it corresponds to a larger grid conductor signal level than pulse width 122.

The CONTROL B signal in FIG. 9 corresponds to the Q10 output of timer counter 71. This signal also produces the UP/DN IN signal applied to graphic module 69 to establish the leading edges of the DATA OUT signal on conductor 78 of FIG. 6B.

The Q14 waveform of FIG. 9 is the signal produced on the Q14 output of down counter 71. This output selects multiplexer 5 when at a low level, and causes multiplexer 8 to be selected when at a high level.

The DATA SELECT signal appears on the PB6 output port on pin 18 of microprocessor 72. When this signal is at a "1" it means that the scanning operation now being conducted is a data scan operation in which the signal levels of the 16 major grid conductors are multiplexed to the input of the graphic module 69 to produce the SIGNAL IN signal levels which are used to obtain the X and Y coordinates of the pen tip 10. When the DATA SELECT signal is a "zero," the graphic module and the circuitry in the upper right hand corner of FIG. 6A operates so that the SIGNAL IN signal is reproduced as the FREQ OUT signal on pin 10 of graphic module 69. This signal is gated onto the TIMER DATA IN conductor 81 and applied to the TIMER IN input of microprocessor 72, which then counts the number of such pulses to determine the pen frequency. Reference numeral 124 in FIG. 9 designates bursts of the FREQ OUT signal that appear on the TIMER DATA IN conductor 81 when the DATA SELECT signal is at a "zero", during a frequency scan operation.

The READ TIMER waveform of FIG. 9 simply indicates when the internal timer that measures the widths of the pulses such as 121, 122, and 123 is internally read in microprocessor 72. Each of the READ TIMER pulses occurs after the incoming data pulse has elapsed.

The RESET TIMER pulses in FIG. 9 are internally produced in 6805 microprocessor 72, and simply reset the internal timer thereof that automatically measures the widths of the pulses applied to the TIMER IN input of microprocessor 72.

In FIG. 9 the time interval designated by reference numeral 249 corresponds to the interval during which the data scan operations are being performed, wherein each of the 16 major grid conductors is multiplexed into the graphic module 69, a DATA OUT pulse is gated to the TIMER IN input of microprocessor 72, and microprocessor 72 is set to automatically measure the width of each of the data pulses, such as 121, 122, and 123. Reference numeral 250 designates the interval during which the first eight lines scanned during a frequency scanning operation are ignored. Only the signal from the Y line previously determined to have the greatest signal strength is utilized to determine the present pen frequency.

Reference numeral 251 designates the interval during which the second group of eight major grid conductors are scanned during the frequency scan operation. The internal timer of the 6805 microprocessor 72 is set to count the number of pulses received at the TIMER IN input of microprocessor 72 during a predetermined time interval, to thereby determine the frequency of the signal on each of the second group of eight conductors. The frequency of the particular Y conductor having the previous maximum signal level is the only one saved internally.

The specific operation of the above-described circuitry of FIGS. 6A-6B can perhaps be best explained by describing in detail the program executed by microprocessor 72. The flow charts of the program and subroutine executed by microprocessor are set forth in detail in FIGS. 10, 11A, 11B, 12, and 13, and in Appendix A attached hereto.

Referring first to FIG. 10, this flow chart illustrates the general operation of microprocessor 72 to effectuate the scanning, X and Y coordinate generation, and frequency discrimination decoding functions referred to above. The program is entered via label 100, in response to a power on reset signal which is applied to microprocessor 72. The program then enters block 101 and performs various conventional "initialization" operations, such as setting predetermined initial conditions that establish whether the various microprocessor ports are to be programmed as inputs or outputs during operation, and establishes various initial values in predetermined locations of the internal random access memory (RAM) of microprocessor 72. Part of the initialization routine also causes the microprocessor to set up initial values for certain parameters of the graphic module 69, including initial values of gain and offset, previously described.

PB5 of processor 72 is set "high" during initialization to inhibit all inputs from the first eight X grid lines. These eight lines are then "scanned" for a predetermined "nominal" value. The port bits PA0 through PA4 are used to adjust the DC offset to graphic module 69 until the predetermined nominal data value is obtained. This procedure is repeated for both gain ranges (as controlled by PA5 of 72). In this manner, the circuitry is always "auto adjusted" to the same initial reference point, so that slight differences in the analog gains or other circuitry can be discounted. It also eliminates the time consuming and costly procedure of "manual" adjustment that might otherwise have to be done to each printed circuit board during the manufacturing operation.

After the initialization operations are complete, the program goes to block 102, which represents a "spare time loop" in which the microprocessor 72 operates in a "wait" state, in a low power mode, waiting for an interrupt from the Q8 output of counter 71, which interrupt occurs at a 4800 kilohertz rate.

If an interrupt occurs, the program enters label 110, and executes the interrupt routine of FIG. 12, described in detail subsequently. However, blocks 111-115 of FIG. 10 provide an overview of the interrupt subroutine of FIG. 12. After entering block 110, the interrupt routine raises or lowers the logic level of the serial output data on conductor 75, from the PA7 port of microprocessor 72, if the present time is the proper time to do this, as indicated in block 111. The interrupt subroutine then determines if it is the proper time to read (input) the timer data register, which contains the "data" or "frequency" information as loaded into microprocessor 72 by the "timer in" line 81. The external timing is designed such that the information from one of each of the major sixteen grid lines is available every four interrupt times. If it is not the proper time to read the timer data register, then the IRQ routine is exited, returning program execution to the point at which it was interrupted. If it is time to read valid timer data as determined in block 11A, the interrupt subroutine then determines in block 112 if a "data" or "frequency" scan of the X and Y major grid conductors is currently in process. If currently a data scan is being performed, then the timer register contains the measured pulse width of one of the sixteen major grid lines, which is now input and saved dequentially in an internal RAM location, for evaluation later during the "EVAL" routine of "spare time", as indicated in block 113. After some "housekeeping" activities, the program execution is then returned via label 115 to the point at which it was interrupted.

If the determination of decision block 112 is that a frequency scan is currently in progress, the timer register contains "frequency" information. If the frequency scan is at the Y grid conductor having the maximum signal level, as determined by the previous "data" scan, then this frequency information is input to a RAM location and saved, as indicated in block 114.

The data scan operation of block 113 or the frequency scan operation of 114 are performed only every fourth interrupt, i.e., at a 1200 hertz rate. The rest of the time, the microprocessor 72 is available for "spare time" processing, which includes going through the blocks 103, 104, 105, etc. to evaluate the data obtained and stored during execution of the interrupt routine of blocks 110-115 and to compute the pen position or decode pen switch commands resulting from operation of pen 9.

After all sixteen major X and Y lines have been scanned in accordance with block 113, the determination of decision block 112 is that the next scanning operation will be a frequency scanning operation. The scanning of the first eight major grid lines is ignored, and the FREQ OUT signals produced in response to scanning of the next eight major grid lines also is ignored until the major Y conductor on which the last maximum Y signal was obtained is scanned. Microprocessor 71 reads the timer data register, which has just been loaded with the count of the number of pulses of that major Y grid conductor within a predetermined amount of time, compute the present pen frequency, and save it in accordance with block 114. This pulse count is decoded to determine its frequency during spare time processing, as indicated by block 121.

Most of the time the program executed by microprocessor 72 waits in the spare time processing loop 102 of FIG. 10, waiting until a new major grid conductor signal strength is saved in accordance with block 113. As soon as the program receives such a piece of information, it goes from block 102 to block 103 and executes the EVAL subroutine of FIGS. 11A and 11B, wherein the program evaluates and sorts the stored data representing the signal strengths of the major grid lines scanned until the present time.

After execution of the EVAL subroutine, subsequently described, the program goes to decision block 104 and determines whether the X and Y major grid conductors having maximum signal strengths have yet been found. If this determination is negative, the program returns to the spare time processing loop 102 and waits for the present (or next) data scanning operation to provide additional signal strength data. If both the maximum X and Y signal strengths have been found, the program goes to block 105 and executes the CALC subroutine of FIG. 13, wherein the precise pen position is computed on the basis of the three maximum X line signal strengths and the three maximum Y line signal strengths, in accordance with certain predetermined formulas. In the course of executing the CALC subroutine, several additional subroutines are executed, including one for decoding the status of the pen switches 57, 58, and 59. After the CALC subroutine of block 105 has been executed, the binary pen position is known, with a resolution of 200 points per inch. If the determination of decision block 104 is negative, this means that the pen is "out of range"; what this means is subsequently described.

Next, the program goes to decision block 106 and determines if the digitizer tablet 2 is in a "tablet" or "mouse" mode. If it is in a "tablet" mode, in which the objective to provide precise X and Y pen position coordinates, the program goes to block 107 and executes a subroutine that filters or averages out "jitter", i.e., signal variations due to slight variations in the pen as it is being held by a slightly moving human hand, and then enters block 108 and formats the binary pen X and Y coordinates for transmission in the tablet mode to an external parallel buffer. The program then enters block 120, loads an internal software transmit buffer, and sets a transmit flag in block 111, enabling microprocessor 72 to serially transmit the binary pen coordinate data to a separate computer or peripheral device. The program then returns to the spare time processing loop 102.

If the determination of decision block 106 is that the digitizer tablet 2 is in a "mouse" mode, wherein the objective is to detect amounts of change in the pen position, rather the precise coordinates of the pen, the program goes to block 116 and rounds off the binary coordinate data to an accuracy of 50 or 100 points per inch (as preselected by the user), goes to block 117 and filters out jitter, and then goes to block 18 and computes the amount of change in the present X and Y pen coordinates with respect to prior values of those coordinates obtained in the previous operation. The program then goes to block 119, formats the computed changes for transmission to the external parallel buffer or an external peripheral device, enters block 120, and transmits the "mouse mode" data to the external device.

It should be appreciated that those skilled in the art know that the nature of "mouse mode" data is different than coordinate data obtained by digitizing the position of a pen or cursor. In the latter case, X and Y coordinates are measured with respect to a fixed reference point. "Mouse mode" data does not consist of X and Y coordinates, but of positive or negative amounts of change from the last location of the "mouse".

It should be noted that the unused information resulting from the above described frequency scanning of the first eight major grid lines gives the microprocessor 72 plenty of time to perform all of the calculations of the EVAL and CALC routines.

Referring now to FIG. 11A, the EVAL subroutine is entered at label 160. The purpose of this subroutine is to "sort out" the signal strengths from the sixteen major grid conductors and determine the conductors having the three maximum X line signal strengths and the conductors having the three maximum Y line signal strengths. First, the program goes to decision block 161 and determines if the signal from the line (i.e., grid conductor) presently being scanned is large enough to be significant, i.e., if it has at least a predetermined "proximity" level. If this determination is negative, it may mean that the scan is approaching the location of the pen tip, or retreating from the location of the pen tip. In any case, the program goes to decision block 171, and determines if the signal strength of the previously scanned line is within a predetermined range. If this determination is negative, the program goes to label 170.

If the determination of decision block 171 is affirmative, then it means that the previously scanned line is the one having a maximum value. In this case, the program then goes to block 167, sets the previously scanned line as the one having maximum signal strength, and goes to decision block 168. (Note that all of the operations in FIG. 11A can apply to evaluation of signal level data from either X or Y lines; the flow chart of FIG. 11B determines which.)

In decision block 168 the program determines if the present line signal is greater than the next to last line scanned. If this determination is affirmative, the program goes to decision block 169 and sets a "positive" pen position flag that indicates that the pen position is past or beyond the grid conductor having the maximum signal value, and goes to label 170 of FIG. 11B.

If the determination of decision block 168 is negative, the program enters block 177 and sets a "negative" pen position flag which indicates that the pen position is located before the grid conductor having maximum signal strength.

If the determination of decision block 161 is affirmative, i.e., if the signal strength of the grid conductor presently being scanned exceeds the "proximity" level, then the program goes to decision block 162 and determines if the signal strength of the present line exceeds a certain predetermined maximum level. If this is the case, it is necessary to adjust the gain of the amplifiers in the graphic module 69 by changing the level on conductor 90, i.e., the gain adjust input, thereby "deleting" amplifier 87 of FIG. 6B from the amplifying circuitry, as indicated in block 172. (The resulting place shift has no significant effect on the operation of the system.) The program then goes to label 170 of FIG. 11B. If the graphic module is already at this gain level, the DC offset is adjusted from its "nominal" value, as predetermined by the interrupt routine, to assume that the grid lines are maintained within the dynamic range of the system.

If the determination of decision block 162 is negative, the program enters decision block 163 and determines if the maximum signal strength for the present group of conductors being scanned has already been found. If this determination is affirmative, the program goes to decision block 173 and determines if the present signal strength is less than the signal strength of the previously scanned line of the present group. This determination should never be negative, but if it is due to a noise pulse or other system error, the program goes to decision block 174 and increases the "proximity" level with reference to which the determination of decision block 161 is made, so that, in essence, "meaningless" signals are filtered out. The program goes to label 170 of FIG. 11B.

If the determination of decision block 173 is affirmative, the program goes to label 170 of FIG. 11B.

If the determination of decision block 163 is negative, the program goes to decision block 164 and determines if the previously scanned signal level is also in the predetermined "range". If the determination of decision block 164 is negative, the program goes to block 175 and sets a "previous line in range" flag and goes to label 170 of FIG. 11B. Note that the "previous line in range" flag of block 175 is the reference for the next scan. At the present time, it actually means that the current line is in the predetermined range. If the determination of decision block 164 is negative, the program goes to decision block 165 and determines if the signal strength of the present grid line is greater than for the previous line. If this determination is affirmative, it means that the scan process is continuing to find major grid conductors with larger signal levels than the previously scanned lines, so the line having the maximum signal level has not yet been found. The program then goes to label 170 of FIG. 11B. If the determination of decision block 165 is negative, the program goes to decision block 166 and determines if the previous line is the first line of the present group, i.e., if it is the "X.sub.0 " or "Y.sub.0 " line. If this determination is negative, the program goes to decision block 167, previously described. If the determination of decision block 166 is affirmative, the program goes to block 176 and sets a flag indicating that the pen position is located to the left of or below the scanning range of the digitizer major grid conductors. The program then goes to label 170.

It should be noted that the capability of the foregoing routine to cause the microcomputer 72 to automatically adjust the gain and offset in accordance with the signal levels being scanned (as in block 172 of FIG. 11A) is important to the present invention because of the considerable variation in the pen signal level capacitively coupled to the grid conductors due to the absence of a fixed ground in the pen. The automatic adjustment of gain and offset also enables the circuit design to be accomplished in such a way as to minimize power dissipation. Those skilled in the art will recognize that if a "worst case" design were to be performed, wherein an extremely wide range of grid conductor signal levels could be adequately processed at one range setting, a much higher level of system power dissipation would occur than is the case for the present digitizer.

Referring next to FIG. 11B, the program enters decision block 181 and determines if all sixteen grid lines have been scanned during the present sequence of the scanning operation. If this determination is affirmative, the program goes to block 189 and sets flags and certain software registers to prepare for the evaluation of data obtained by scanning X conductors, so that the program knows that the next pass through the EVAL subroutine is being performed for the X group of grid conductors rather than the Y group.

The program then goes to decision block 190 and determines whether the Y line with the greatest signal strength has been found. If this determination is affirmative, the program goes to decision block 191 and determines if the X line with the greatest signal strength has been found. If this determination also is affirmative, the program exits to the CALC subroutine (label 192) to compute the precise binary coordinates of the pen tip position. If the determination of decision block 191 is negative, the program returns to the main spare time loop (block 102) of the program of FIG. 10 via RTS (return from subroutine) label 188.

If the determination of decision block 190 is negative, the program enters decision block 193 and determines if the signal level of the last line scanned is the only one within the predetermined "range". If this determination is affirmative, the program goes to block 194 and sets a flag indicating that the pen is positioned to the right of or below the active grid scanning area of the tablet surface, and returns to block 104 of FIG. 10.

If the determination of decision block 193 is negative, the program goes to block 195 and indicates that the signal levels of all Y lines scanned are out of the "range" of the grid conductors, i.e., that the signal levels are less than the "proximity" level, the minimum signal level which indicates that the pen tip is in the "proximity" of the digitizing surface. The program then goes to block 196, decreases the value in the "Y proximity register", and/or adjusts the range and offset registers to achieve higher gains, and then returns to the main program via label 188.

If the determination of decision block 181 is negative, the program goes to decision block 182 and determines if the first nine lines (for "X" data) have been scanned. If this determination is negative, the computer returns to the main program of FIG. 10 via label 188.

If the determination of block 182 is affirmative, the program goes to block 183 and sets appropriate flags and register values to cause the EVAL routine to evaluate the signal levels for the next group of conductors scanned, which will be the Y conductors.

The program then goes to decision block 184 and determines if the X conductor having the maximum signal strength has been found. If this determination is affirmative, the program goes to block 197 and saves that X line number, and returns to the main program. If the determination of decision block 184 is negative, the program goes to decision block 185 and determines if the last X conductor scanned is the only line in the present range. If this determination is positive, the program goes to block 198 and sets a flag indicating that the pen is positioned to the right or below the digitizing surface, and returns to the main program. If the determination of decision block 185 is negative, the program goes to block 186 and sets a flag indicating that all of the X lines are out of the range of the digitizing surface. The program then goes to block 187 and decreases the value in the X proximity register and/or adjusts the range and offset registers to produce higher gain in the graphic module 69. The program then returns to the main program of FIG. 10.

Referring now to FIG. 13, the CALC subroutine referred to in block 105 of FIG. 10 is entered at label 200, for the purpose of computing the precise binary X and Y pen coordinates after the three X lines and three Y lines having maximum signal levels have been found. (Note that in the EVAL subroutine, it is inherent that the next highest signal levels are on the lines adjacent to the one having the maximum signal level.) The program goes to block 201 and calls up a subroutine called PSWSTA (not described in detail, as it is very straight-forward and can be easily provided by one skilled in the art) for decoding the pen switch status on the basis of the pen frequency reading saved in accordance with block 114 of FIG. 10 during the last frequency scanning operation. The program then goes to block 202 and sets a flag that indicates that the present calculation is for a binary Y coordinate for the pen position, rather than a binary X coordinate.

Next, the program goes to block 203 and recalls the stored levels B, A, and C for the three highest signal levels, respectively, obtained during the previous scanning of the Y grid conductors.

The program then goes to block 204 and computes a "divisor" in accordance with the equation ##EQU1##

The above equation is an empirically determined equation that has been found, in conjunction with equation (2) below, to provide accurate results in computing a precise pen coordinate on the basis of signal strengths of relatively widely spaced grid conductors with the above-described minor grid conductors positioned and resistively coupled therebetween.

Next, the CALC subroutine goes to block 205 and computes a "dividend" in accordance with the equation

The "128" in equation (2) represents half of the number of points between each major grid line. The major grid lines are 1.28 inches apart, with 256 points between them, i.e., 200 points per inch.

Next, the program goes to block 206 and computes the quotient obtained by dividing the dividend of block 205 by the divisor of block 204.

The program then goes to decision block 207 and determines (by reading the flag set in blocks 169 and 177 of FIG. 11A) whether the pen is located to the right or left of the grid conductor having the maximum signal strength (assuming scanning occurs from left to right), or whether the pen is located above or below the grid conductor if a "Y" scan is being done, wherein scanning is done from the bottom to the top, rather than left to right. If the pen position is located after (to the right of) the grid conductor with maximum signal strength, the program goes to block 212 and computes the least significant bit (LSB) of the binary pen position coordinate as equal to 128 plus the value of the quotient coupled in block 206. Otherwise, the program goes to block 213 and sets the LSB of the binary pen position coordinate to 128 minus the quotient computed in block 206. In either case, the progam goes to block 208 and computes the most significant bit (MSB) of the binary pen position coordinate being computed in accordance with the equation

where MAX LINE NO. is equal to 0 through 6 for X coordinates and is equal to 0 through 4 for Y coordinates. This value is used to reference a stored "look-up" table, which precisely defines the pen location and corrects for small errors in the above formula.

The program then goes to decision block 209 to determine if the binary pen position coordinate just completed is for an X coordinate. If this determination is affirmative, the program goes to a format routine (not described, as it would be easy for one skilled in the art to format the data as it desired to be input to an external computer, such as an IBM PC) via label 210 to format the binary X and Y pen position coordinates in a predetermined manner so that they can be output to an external buffer (not shown) in a suitable format to be read by an external computer. If the determination of decision block 209 is negative, the program goes to block 214, sets the flag of block 202 to a value that indicates that the binary coordinate being calculated is an X coordinate, and returns to block 203 and repeats the above-described process.

Next, the interrupt subroutine (previously briefly described with reference to blocks 110-115 of FIG. 10) that is executed by microprocessor 72 in response to each pulse supplied to its IRQ input is described in detail with reference to FIG. 12. The interrupt subroutine is entered via label 125. The subroutine goes first to decision block 126 and determines if the present time is an appropriate time to transmit a bit of serial data from microprocessor 72 to the above-mentioned external parallel buffer or peripheral device. This determination is made by testing two conditions, first, a flag (set in block 120 of FIG. 10) that indicates whether the pen position has been computed and properly formatted, and hence is ready to be transmitted. The second condition tested is whether the time is right to change the data bit in order to transmit it at the predetermined baud rate. This is determined by an internal counter, which keeps track of the number of interrupts, since the interrupt rate occurs at 4800 hertz, and the transmit speed is a derivative of this rate, i.e., is 2400 baud (bits per second), 1200 baud, 600 baud, etc. If this decision is affirmative, the program goes to block 127 and sets the serial output port (PA7) of microprocessor 72 to a logical "one" or "zero" depending on the value of the bit to be transmitted. The interrupt routine then goes to decision block 128. If the determination of decision block 126 is negative, the program goes directly to decision block 128.

In block 128, the interrupt routine determines if the timer data register has been properly loaded from the input data applied on the DATA/FREQ input (conductor 81 of FIG. A6). At every fourth interrupt pulse, the timer data register of the microprocessor 72 will be valid. If it is not valid, the interrupt routine returns to the point in the main program at which the interrupt occurred.

If the timer data register is valid, the program goes to decision block 129 and determines if a data scan operation or a frequency scan operation is presently being performed. If a data scan is being performed, the program goes to block 131 and performs the function of reading the value of an internal input timer data register. The routine then goes to block 132 and sequentially saves the data just input to the input timer of the microprocessor and transferred into one of sixteen locations in an internal buffer. The position of the data in that buffer determines to which of the sixteen major grid conductors the presently saved signal strength corresponds. The interrupt routine then goes to decision block 133 and determines if all sixteen grid conductors have been scanned and if the corresponding signal strengths are determined and saved in the software buffer of block 132. If this determination is affirmative, the interrupt routine goes to block 151 and resets a pointer that points to the internal buffer, and then goes to decision block 152.

In decision block 152 the program determines if the external counter 71 is in proper synchronization with the internal software counters in microprocessor 72, in case a system error has interrupt the proper synchronization. This is done by reading a few of the outputs of counter 71 to see if they have proper values, which are read via ports PB0 through PB2. If this determination is negative, the interrupt routine goes to block 154 and loops until the interrupt line goes high, i.e., one-half of the interrupt cycle has elapsed. As soon as this occurs, the program then causes microprocessor 72 to send out a reset pulse to counter 71, as indicated in block 155. This synchronizes the internal microprocessor counters with the external counter 71. The routine then goes to block 156 and resets the internal software counters to begin the data scanning operation over again. The interrupt routine then goes to block 136 and resets the timer data register to an initial state and returns to the main program via label 137.

If the determination of block 152 is that the external counter and internal software registers are in sync, the program goes to block 153 and sets a flag which causes the routine to ignore the next eight lines of the data input to the data timer of microprocessor 72, and then causes it to measure the frequency, i.e., count the pulses, of the remaining eight grid lines scanned, i.e., perform the "frequency scan". The interrupt routine then goes to block 136.

If the determination of block 133 is negative, the program goes to block 134 and determines if the first nine grid conductors, i.e., the X conductors, have all been properly scanned. If this determination is negative, the program goes to block 136 and resets the timer data register of microprocessor 72. If the X lines all have been scanned, the program goes to block 135 and sets the offset and range ports to appropriate values for the Y scanning operation. These values have been previously determined by blocks 172, 174, or 196 of the "EVAL" subroutine of FIGS. 11A and 11B. The program then goes to block 136 and resets the microprocessor timer input register.

Returning to decision block 129, if this determination is that a frequency scan is being done, the interrupt routine goes to decision block 139 and determines if the present pass is the first pass of the frequency scanning operation. Block 153 of FIG. 12 will have already set the interrupt routine to ignore the first eight X lines. If these lines are the ones presently being scanned, then they are ignored by the "no" condition of block 128. The "FREQ" branch of decision block 129 is only taken during the scan of the last eight grid lines, those being the X8, Y0, Y1 . . . Y6 lines. Since the first line looked at in block 129 is the X8 line, it is the only one ignored, but this now allows the program to set up the frequency scan operation to look at the next seven Y lines for the frequency information. If this is the case, the interrupt routine goes to block 148, and sets the microprocessor input timer to ignore the data presently on the FREQ OUT input conductor 233 of FIG. 6A. The interrupt routine then goes to block 149 and sets an internal timer control register to function as a "timer" input only (i.e., to count the input pulses instead of measuring their widths).

If the determination of decision block 139 is that the present pass is not the first time through, the routine goes to decision block 140.

If the determination of decision block 140 is that the present line being scanned is not the last Y line in the group, the routine goes to decision block 141 and determines if the present line is the Y line having the maximum signal level. If this determination is negative, the routine goes to block 136. Otherwise, it goes to block 142 and cause the input timer of microprocessor 72 to read the number of pulses of the maximum Y conductor signal level, and then goes to block 143 and saves this frequency count so that it can be decoded, as previously explained, to determine the present settings of switches 57, 58, and 59 of FIG. 4. The interrupt routine then goes to block 136, and then returns to the main program via label 137.

If the determination of decision block 140 is that the present conductor is the last one to be scanned during the frequency scanning operation, the program goes to block 145 and sets the output ports PB6 and PB7 of microprocessor 72 to indicate that the next scanning operation is to be a data scan. Blocks 146 and 147 then set internal registers properly for the next X data scan on the X lines, and then returns to the main program via block 136 and label 137.

Referring now to FIGS. 7 and 8, instead of using the cordless pen 9 of FIGS. 1 and 2 in conjunction with the digitizing tablet 2, a cordless keyboard 240 having a plurality of depressable keys 241 thereon is positioned on digitizing surface 2. Mechanical alignment indicia can be utilized to ensure proper alignment of keyboard 240 with digitizing surface 2.

Portable keyboard 240 includes a ground plane 242 having a metal conductive strip 244 (FIG. 8) thereon. Metal conductor 244 conducts a signal produced by an oscillator identical or similar to the oscillator 46A shown in FIG. 4 that is contained within cordless keyboard 240.

Ground plane 242 has a plurality of openings 243 therein. Each of the depressable keys 241 has a conductive shaft 247 that extends through one of the holes 243. When a particular key 241 is depressed, its conductive shaft or plunger 247 extends down to or nearly to the upper surface of digitizing tablet 2.

As seen in FIG. 8, the bottom surface of each of the key caps 241 has an annular conductive plate 245 thereon. A variable capacitor formed by the conductive plate 245 of each key and the conductive strip 244 carrying the oscillator signal is designated by reference numeral 246. When the key is depressed, the spacing between the plates of the variable capacitor 246 decreases greatly increasing the value of the capacitor 246. The increased capacitance greatly increases the capacitive coupling of the oscillator output signal from the conductor 244 to the conductive plate 245. The conductive plates 245 are electrically connected to the plungers 247, so the effect of depressing any particular key, such as 241A in FIG. 8, is to effectively couple the oscillator signal to the lower end of the plunger 247 of that key. (The conductive plates 245 alternately could be capacitively coupled to the plungers 247.) The lowering of the plunger 247 by depressing a key increases the capacitive coupling between it and the major grid conductors such as 3 and 4 previously described with reference to FIGS. 1 and 3 and the minor grid conductors resistively coupled therebetween. This enables the previously described circuitry to quickly digitize the location of any depressed key. This information then is suitably formatted and output to an external computer. The external software, characters or items of which are selected by means of the digitized information obtained by the external computer from the digitizer tablet 2.

In accordance with this embodiment of the invention, signal transmitting indicia can be provided on the bottom of the ground plate 242 for coupling oscillator signals to the digitizer tablet 2 in order to enable the digitizer tablet and the external computer connected thereto to interpret the coordinates of such signal-transmitting indicia in order to enable the external computer to identify the type of keyboard placed on the digitizer tablet 2. This enables the external computer to determine which "menu" should be used for interpretation of depressed keys of that keyboard.

Of course, it is not essential to this embodiment of the invention that the keyboard 240 be cordless. A separate power supply and oscillator circuit could be provided, and the oscillator output signal could be connected by a cable to the conductor 244. As in the case of the previously described cordless pen 9, not all information produced by the keys needs to be locational or coordinate information. As before, the frequency of the oscillator can be shifted by depressing certain keys, and the resulting frequency can be decoded in precisely the manner previously described.

Thus, it can be seen that the invention provides a digitizing tablet with a cordless pen that can operated in either a dignitizing mode or a mouse mode. The described system dissipates far lower levels of power than typical prior art digitizing system, and nevertheless has adequate noise immunity despite the necessarily low pen signal levels, and the very low resulting signal levels capacitively coupled from the cordless pen to the grid conductors in the digitizing surface. The dual mode operation, i.e., the digitizing mode and the mouse mode, is provided at very little additional cost. The entire system achieves the objective of low overall cost by implementing the desired functions with a minimum number of system components, including printed, passive grid components. By relying upon only signal levels of three X lines and three Y lines, the susceptibility of the system to externally induced noise is greatly reduced. The disclosed removable keyboard structure provides previously unavailable versatility in a computer input device, allowing digitizing, mouse, and keyboard input operations to a computer with a single low cost system level upon a single digitizing tablet.

While the invention has been described with reference to a particular embodiment thereof, those skilled in the art will be able to make various modifications to the described embodiment of the invention. For example, it is not necessary that the control of the scanning operations be done by down counter 71, as described herein, but may be directly controlled by the microprocessor 72. The technique of using down counter 71 as described herein simply lowers the cost of the system, at some sacrifice in operating speed, which we found to be acceptable in this case. In some instances, it may be desirable to drive the pen synchronously. Also, it is not essential that the tubular shield 16 actually be tubular, although this is a preferred configuration. All that is really necessary is that the shield have a sufficiently greater capacitance than the pen cartridge that the shield effectively acts as a virtual ground, so that it undergoes potential transitions which are much smaller than the potential variations of the pen tip. Other techniques for routing the oscillator signal to the various signal transmitting elements associated with the various keys in the embodiment of FIGS. 7 and 8 can be provided. Furthermore, it is not necessary that batteries be used as the power source in the cordless pen or in the cordless keyboard described. For example, inductive coupling between a power source associated with the digitizer tablet could be used to couple AC signals into a sensing circuit located in the cordless pen, which rectifies the induced signals to produce the DC levels necessary to operate the pen circuitry. The plastic tip 12 in FIG. 2 can be metal, and thereby make electrical contact to pen tip 10. This improves the capacitive coupling to the grid conductors and improves the grid conductor signal levels. Techniques other than those disclosed can be provided for varying the oscillator frequency in response to closure of the pen switches 57, 58, and 59. For example, the resistor 57 that is switched into the oscillator circuit by pressing down on the pen tip 10 could be a pressure-responsive resistor that changes the oscillator frequency as a function of pen pressure on tip 10, and the various frequencies could be decoded to represent different line widths. By way of definition, it is to be understood that the pen need not be a writing instrument. The term "pen" as used in the claims is intended to encompass any hand-held instrument that has a conductive tip or element that capacitively couples a signal to the grid conductors in the manner described herein. It is not essential that the grid conductors be in orthogonal sets as described. Linear resistive parallel conductors could be used, wherein the conductor locations determine one coordinate of the pen and resistive voltage division proportional to the pen position in such conductors determine the other coordinate.