Microprocessor Controlled Garage Door Operator

1. Field of the Invention

This invention relates in general to garage door operators and in particular to a microprocessor controlled garage door operator.

2. Description of the Prior Art

Garage door operators of the prior art utilize upper limit and down limit switches which are actuated by movement of the door so as to set the up and down limits for the door. Prior art garage door operators also have force adjusting means so that the up and down force for the door can be adjusted by setting such force limit adjustment means. The up and down limit switches are set manually by the operator and the up and down force limits are set manually as, for example, by increasing the force limit by rotating a knob the first direction and decreasing the force limit by rotating a knob in a second direction.

Since the size of the door openings and the size of the door vary from installation to installation and since the weight of the door and setting of the springs require different force limits for different doors it has previously been necessary that the up and down limits and the force limits be set by the installer or the home owner when the garage door operator is installed. This leads to faulty installation in that the installers and home owners do not properly set the up and down limits and also they set the force limits too high or too low which results in a dangerous condition, particularly if a child or other object is caught under the door as it is moving down.

The present invention provides a microprocessor controlled garage door operator which includes a track and head unit mounted to the ceiling of the garage and which has a trolley connected to the door so as to move it up and down. A radio receiver is mounted in the head unit and is connected to the microprocessor unit and a control unit is mounted on the wall of the garage and is connected to the microprocessor. A security switch may be mounted to the outside of the garage and be connected to the microprocessor. The control unit and microprocessor may be put into a program mode in which condition the door may be moved downwardly with a down switch mounted on the control unit to the desired down position of the door and this position will be automatically set into the microprocessor memory. Then the door may be moved to the full up position with ar up switch which will set the up limit of the garage door in the microprocessor memory.

After the up and down limits have been set, the unit is placed out of the program mode and into the operate mode and the garage door is operated through a complete cycle which will automatically set the up and down force limits for the door. Subsequently, the door may be operated up and down with the up and down set limits and with the set force.

If for any reason the door does not properly operate, a door condition light will indicate that there may be something wrong with the operation of the control or the door. When this condition exists, the operator will be given a warning and can take proper correction action to remedy any such defects.

It is seen that the present invention comprises a microprocessor control garage door operator wherein the up and down limits and the forces set for the door are set in a manner such that they are stored in the memory of the microprocessor and the invention makes it impossible for the home owner to set the force limits too high which results in a dangerous condition. This is because the forces are automatically set for the particular installation in the microprocessor by measuring the actual forces needed to raise and lower the door and by setting the force limits slightly as, for example, 10% above these required forces. Thus, it is not possible for the home owner to set the forces to limits which are much greater than are necessary which can result in a dangerous situation.

Also, since the up and down limits are set by the home owner by noting and controlling the actual up and down positions of the door, these will be accurately set. This also prevents the down and upper limits from being set at points further than the door can actually move since the actual position of the door at the full down and full up position sets the up and down limits in the memory of the microprocessor.

A security switch may be utilized to allow the garage door operator to be actuated externally of the garage upon the use of the proper code. The attempt to open the garage door with the improper code will prevent the garage door operator from being actuated.

Other objects, features and advantages of the invention will be readily apparent from the following description of certain preferred embodiments thereof taken in conjunction with the accompanying drawings although variations and modifications may be effected without departing from the spirit and scope of the novel concepts of the disclosure and in which:

FIG. 1 illustrates the garage door operator of the invention installed to operate a garage door;

FIG. 1A illustrates the security switch;

FIG. 1B illustrates the control unit;

FIG. 2 is an electrical schematic of the control unit of the invention;

FIG. 3 is an electrical schematic of the transmitter of the invention;

FIGS. 4A, 4B and 4C illustrate the electrical schematic of the garage door operation; and

FIG. 5 illustrates the motor and speed control of the invention.

FIG. 1 illustrates the garage door operator 10 of the invention installed to move a garage door 14 which is mounted on tracks in a conventional manner. The head unit 11 of the garage door operator 10 includes a motor which is mounted in the head unit 11 and has an output gear 21 that drives an endless chain 19 which also passes over a gear 22 mounted on the other end of the track 12. A trolley 13 is engageable with the chain 19 and moves on the rail 12 and has an arm 16 which is connected by bracket 17 to the door 14 so as to move it up and down. A light cover 18 is mounted to the head unit 11 and a control unit 19 is mounted on the inside wall of the garage and is connected by an electrical cable 22 to the microprocessor mounted in the head unit 18. A security switch 21 has a plurality of buttons which can be actuated in various sequences to provide a code for actuating the garage door operator 10 externally of the garage. It is connected to the garage door operator 10 by a cable 23. A transmitter 24 can be used to actuate the garage door operator 10 remotely and transmits a radio signal which is received by a receiver mounted in the head unit 11 to actuate the garage door operator.

The wall unit 19 has a control push button 25 for operating the door up and down and a control indicator light 26 which indicates when the garage door operator is being actuated. A vacation/down switch 27 is mounted on the control unit and indicator 28 indicates certain conditions when this switch is actuated. A work light/up switch 29 is mounted in the control unit 19 and an indicator 31 for this switch is also mounted in the control unit 19.

The security switch 21 is provided with a numbeer of switch actuators 32 which can be actuated in different sequences so as to provide a code to the microprocessor in the head unit 11 so as to actuate the garage door operator.

An energize switch 33 is provided for the transmitter 24 so as to remotely operate the garage door operator.

One surface 36 of the head unit 11 has a door condition indicator light 37 and a program/operate mode button 38 to allow the garage door operator to be set in the operate or/program mode. Antenna 39 of the radio receiver mounted in the head unit 11 extends from the head unit 10 as illustrated.

FIG. 2 illustrates the control unit 19 with the switches 25, 27 and 29 and the indicators 26, 28 and 31. A contact point 42 is connected to the collector of a transistor Q201 which has its emitter connected to ground and to a common terminal 43. The command switch 25 is connected in parallel with the collector-emitter of the transistor Q201. A terminal 44 is connected through the indicator 31 which may be a LED and through a resistor R201 which may be 470 ohms to a resistor R202 which may be 820 ohms and through a diode rectifier D205 to terminal 42. A Zener diode D201 and the indicator 28 which may be an LED is connected between terminal 44 and ground. A Zener diode D204 is connected between the junction point between resistors R201 and R202 and ground. A capacitor C201 which might be 10 microfrareds is connected between the junction point between the resistor R202 and diode 205 and ground. The indicator 26 which may be a LED is connected between terminal 42 and a resistor R203 which might be 1.2K ohms and ground.

An oscillator comprises a resistor R205 which may be 1.5M ohms which is connected to terminals 5 and 6 of an integrated circuit 61 which may be a type CMOS 4011 which has its output terminal connected to a resistor R206 which might be 180 k ohms and to switch 29. The other side of switch 29 is connected to a resistor R209 which might be 62.K ohms and has its other side connected to resistor R205. The switch 27 is connected from resistor R209 to a resistor R207 which may be 430K ohms which has its other side connected to the output of the circuit 61. A diode D207 is connected between resistor R209 and integrated circuit 61.

Integrated circuit 63 which may be a type CMOS 4011 receives at its number 2 input the output of the integrated circuit 61 and has its number 1 input connected to a resistor R112 which may be 91K ohms and which has its other side connected to ground. A resistor R210 which may be 100K ohms is connected between the resistor R112 and the resistor R202 as illustrated.

A capacitor C202 which may be 0.1 microfarad is connected between resistor R205 and the output of integrated circuit 63. An integrated circuit 64 which may be a type CMOS 4011 has its inputs connected to the output of integrated circuit 63 and its output connected to a resistor R211 which may be 3.9K ohms and which has its other side connected to the base of transistor Q201. A resistor R204 which may be 10M ohms is connected between the resistor R205 and the input to the integrated circuit 64 as illustrated.

FIG. 3 comprises the schematic of the transmitter which comprises an antenna 71 connected to an oscillator circuit 72 which is controlled by a transistor 73 and an integrated circuit 74. A code setting switch 76 is mounted in the transmitter and allows the selected code for the transmitter to be established by setting the switches 76 in different combinations. These switches can also be changed so as to vary the code in the transmitter. When the transmit switch 33 is depressed a coded signal is radiated by the transmitter 24 and the coded signal depends upon the setting of the switch 76.

FIGS. 4A, B and C comprise an electrical schematic of the invention. A microprocessor 101 which might be, for example, an Intel type 8049 is connected to the terminals 42, 43 and 44 from the control unit 19 by cable 22. Terminal 43 is the common grounded terminal and pin V.sub.ss of the microprocessor 101 is connected to ground. Terminal 42 which is the command terminal is connected through a resistor R18 which might be a 100K and transistor Q4 to input terminal P11 of the microprocessor 101. Terminal 44 is connected through transistor Q5 and Darlington driver 102 which might be type ULQ2023A to terminal D4 of the microprocessor 101. Terminal 44 is also connected through Darlington driver 103 to terminal D5 of microprocessor 101. The EEROM 104 which might be a type XICOR 24431 has four output terminals which are connected respectively to terminals P20, P21, P22 and P23 of microprocessor 101.

The security switch 21 is connected to five input terminals 106 through 110 which are connected by cable 23 through resistors R22 through R25 to input terminals P14, P15, P16 and P17 of microprocessor 101. Terminal 110 is connected to ground. Capacitors C20, C21, C22 and C23 which might be 0.01 microfarad are connected between ground and terminals 106 through 109, respectively.

The radio frequency receiver 111 receives an input on the antenna 39 which is connected through the inductor L2 which might be 1 microhenry to ground, couples a signal through transformer T1 and parallel capacitor C1 which might be 3.9 picofarads to a capacitor C2 which might be 39 picofarads to the base of transistor Q1. The collector of transistor Q1 is connected through capacitor C3 which might be 1 picofarad to the emitter. A capacitor C4 which might be 3 picofarads is connected between the base and emitter transistor Q1. Inductor L1 which might be one microhenries is connected from the emitter of transistor Q1 to a resistor R4 which might be 8.2K ohms and which has its other side connected to the inductor L2. A resistor R3 which might be 7.5K ohms is connected between the base of transistor Q1 and the inductor L2. The output of the receiver 111 is supplied through capacitor C6 which might be 4.7 microfarads and transistor Q2 through capacitor C12 which might be 1 microfarad and resistor R12 which might be 100 K ohms to a quad voltage comparator 112 which might be a type LM2901 and which has its output terminal connected to INT terminal of the microprocessor 101.

One inch obstruction switch 201 connects input P27 to ground when closed. It automatically sets the one inch reverse on hard obstruction. LED D6 is connected from a voltage source plus V5 through resistor R42 which might be 3.3K ohms to a Darlington driver 116 which might be a type ULQ2023A and which has its other side connected to terminal D7 of the microprocessor 101. The program switch 38 is connected between ground and terminal P13 of the microprocessor 101. A jumper JU1 is connected between ground and terminal P24 of microprocessor 101. 120 volt AC power is applied to terminals 120 and 121 and this is applied across the primary of transformer T2 which has its secondary grounded. Power is applied from one side of the secondary of the transformer T2 to the resistor R33 which may be 1 megohms, for example, to integrated circuit 122 which may be a type CD4011 and which has its output connected to terminal P12 of the microprocessor 101. A pair of diodes D3 and D4 are connected to opposite sides of the secondary of transformer T2 and have their other sides connected to resistor R34 which may be 1 megohm and which has its other side connected to circuit 123 which may be a quad voltage comparator type LM2901 which has its other side connected to terminal TP of the microprocessor 101. A capacitor C15 which might be 1 microfarad and a resistor R35 which might be 150K ohms are connected in parallel between the resistor R34 and ground. A circuit 124 which may be a quad voltage comparator LM2901 is connected to circuit 123 and has its output connected to resistor R40 which may be 1K ohm to the reset terminal of the microprocessor 101. Terminal D6 of the microprocessor 101 is connected to a resistor R41 which may be 1K ohms and through a diode D5 to an input to the circuit 124 and to a resistor R44 which may be 1 megohm which has its other side connected to circuit 123. A terminal D2 of the microprocessor 101 is connected through a Darlington driver 131 which might be a type ULQ2023A 131 to a down relay 132 which controls switch contacts 133 connected between power terminal 121 and the motor down terminal 134 such that when the relay 132 is energized the motor will run in the down direction. When the relay 132 is energized by output on terminal D2 of the microprocessor 101 contacts of switch 133 will be closed and power will be applied to the down direction rotation of the motor 136 to drive the garage door operator in the down direction. Terminal D1 of the microprocessor 101 is connected through Darlington driver 136 which might be a type ULQ20231A to relay 137 which controls switch 138 which applies power to the motor up terminal 139. The other side of the switch is connected to power input terminal 141. Thus, when the microprocessor 101 supplies output on terminal D1, relay 137 is energized which closes switch 138 to cause the motor 135 to run in the up direction to raise the door toward the up position.

Contact DO of the microprocessor 101 is connected through circuit 142 to a relay 143 which controls switch 144 which is connected to the work light on contact 146 to turn a work light 147 on.

The motor 135 has an output shaft 147 which carries a cup-shaped shutter member 148 of generally cylindrical-shape and which has cut-out portions which form segments 149. A light source 151 is supported on a suitable bracket 152 so that it illuminates the inside of the shutter member 148 and a pair of light responsive diodes or detectors 153 and 154 are mounted so as to intercept the light as the shutter members 149 move out from between the light source 151 and the sensors 153 and 154. The light sensors 153 and 154 produce output pulse waves 156 and 157 which have a phase depending upon the direction of rotation of the shaft 147. In other words, if the shaft is turning in the clockwise direction relative to FIG. 5, the pulse wave 156 would lead pulse wave 157 but if the motor shaft 148 is turning in the opposite direction the pulse wave 157 would lead the pulse wave 156 and, thus, by sensing the outputs from the sensors 153 and 154 the direction of rotation of the output shaft 148 of the motor 135 can be detected. The sensor 153 is connected to input terminal 159 which supplies an input to terminal P25 of the microprocessor 101 through circuit 161 which might be a type CD4011. The sensor 154 supplies an input to terminal 162 which is supplied to input terminal T1 of microprocessor 101 through resistor 163 which may be 11K ohms and through circuit 164 which may be a quad voltage comparator LM2901 which has its output connected to terminal T1.

The up and down limits and the forces required are obtained by processing the RPM signals from the RPM signal on input 159 from sensor 153. Both of the RPM signals from sensors 153 and 154 are used to process the limit information. The force that the motor exerts is monitored from one second after the motor turn on until the motor is turned off. The force is determined by measuring the RPM low pulse width and if either of the limits are being programmed or if the force is being programmed an obstruction will only occur if the maximum force setting is exceeded. This will be indicated by the door condition indicator light 37 being illuminated. An obstruction will be indicated if the pulse width exceeds the pulse width that has been set during programming of the force. The door condition indicator is turned off by entering the program mode.

The maximum force is determined by the jumper on pin P24 with the jumper installed the maximum low pulse width allowed is established and with the jumper removed the maximum low pulse width allowed is 6.25 milliseconds, for example. It is to be noted that the photo interrupter shutter 148 or cup produces a square wave with a duty cycle of approximately 57% due to the voltage comparator inversion and this gives the longer time as being low at the microprocessor 101.

Only one low time that exceeds the preset force is needed to create an obstruction indication while in the operating mode. The RPM input is examined approximately every eighty milliseconds for the worse case. Since both the force and the trinary data decode routine utilize the hardware timer, the force is only examined during the blank time of the transmitter. The data decoding routine assures that if noise is detected on the data input, the blank routine is immediately executed and thus the force is measured.

The up force will be programmed whenever an internal up force flag is active and the down force will be programmed whenever an internal down force flag is active. These flags are set whenever the unit is put into the program mode by moving the program switch 38 whenever the limits are changed with the work light switch or the vacation switch 27 or 29. These flags will only be reset when the respective limit is hit. This ultimately means that if the user places the unit into the program mode and then back into the operating mode on the next cycle of the door the forces will be programmed. During programming the forces, the transmitter will be disabled and this will allow every RPM pulse width signal to be examined.

The limits are kept by counting the number of falling edges on the input from the sensor 153. Since the limits could be effected by noise, the program only counts falling edges while the motor is on and for ten seconds after the motor turns off. Noise at any other time will not effect the limit information.

The two RPM signals from sensors 153 and 154 are needed so as to increment or decrement the limits properly by examination of the second RPM signal upon the falling edge of signal T1. The door direction can be positively determined and there are two situations where the door direction cannot be assumed. The first is for door backup at the down limit due to excessive door pressure. The second is during a "soft" door reversal. During a soft reversal (no hard obstruction) while the door is still coasting down, the up relay is pulled in. It may take 0 to 4 RPM counts until the motor can overcome the doors momentum and actually start to move up. These falling RPM edges must be counted in the proper direction.

The door direction is obtained for one second after motor startup and for 10 seconds after motor turnoff. After the motor turns off, the door direction is obtained by a hardware interrupt on the T1 input for 11/2 seconds by hardware interrupt and for 81/2 seconds by examining the T1 and P25 inputs. During the 11/2 second immediately following the motor turn off, the transmitter is disabled due to door direction routine exclusively using the microprocessors hardware timer. Immediately following the motor turn off, the RPM comes in very quickly, and polling may not read the terminal P25 input quickly enough to properly obtain the door direction information.

The receiver is input to the microprocessor on the external interrupt input pin and the first falling edge in a code frame will be the sync pulse. The sync pulse is timed out and stored. Subsequent data bits will be 1, 2 or 3 times the sync pulse width. Only sync pulse widths of 1/2 milliseconds to 2 milliseconds are accepted. Anything else will be considered to be noise and decoding will be aborted by going into a blank time-out.

During the blank time the interrupt input is polled. After bit ten of a code is determined, a blank flag is set. Every loop of the main program will poll the interrupt input. If the input is ever low this flag will be reset. After the blank timer times out this flag is examined. If high, then the blank time is considered good. If low, then the blank time is considered bad. The next code to come in will be considered invalid.

In the control unit 19, the three switches and three lights are arranged so that the three switches and the command LED 26 are multiplexed over the command wire 42. The command LED 26 will be on as long as power on this contact is high. The command switch 25 shorts out the power supply to the control unit. The microprocessor will detect a constant low and interpret this as a command switch closure. The vacation switch 27 and the work light switch 29 when closed generate a low frequency low duty cycle wave form. The microprocessor will determine the frequency of oscillation to determine the switch that is closed. Particular frequencies were chosen to allow the AC line frequency to time out the switch closure.

The two indicators 31 and 28 are controlled by one wire. A high will cause the vacation light 28 to be on. A low will cause the work light 31 to be on and a floating circuit will cause both lights 28 and 31 to be off. If both functions are to be on by closing both switches 27 and 29, the microprocessor will oscillate between the two lights at a rate of 8.33 milliseconds and this will give the appearance that both lights 28 and 31 are continuously on.

Power up and down is accomplished by two hardware signals. Advanced power fail is applied to terminal TO of the microprocessor and the reset terminal. Power up will have the advance power fail input going high followed by the reset going high. Reset going high allows the microprocessor to start executing instructions. The circuit is configured to allow approximately 5 cycles of AC power to dropout without affecting either the off or reset.

Power down begins by the advanced power fail signal going low. The circuit is constructed to allow small AC line drops without effecting either the APF or the reset signals. The amount of time is very dependent on the AC line amplitude just before dropout. The microprocessor will detect this and store the 16 interval RAM buffers into the EEROM 104. The microprocessor will then cause the hardware reset by pulling I/O port D6 low. The circuit is configured so that either the microprocessor will reset itself or the APF being low long enough will reset the microprocessor.

1. Initialization

2. Power Down

3. Timers

4. Motor Operation

5. Motor and Light Relay Subroutines

6. Door Limits of Travel and Direction of Travel

7. Force Routine

8. Control Case

A. Command, work and vacation switches

B. Work and Vacation Leds

9. Program Mode Switch

10. Auxiliary Obstruction

11. Receiver Decoding

12. EEROM Information Storage and Retrieval

13. Keyboard

14. RAM Organization and Flag Definitions

15. Timer Definitions

NOTES:

A. All labels that are in the software document will be references in this paper by putting a colon after the word, i.e., the label start will appear as start:

B. All numbers are decimal unless followed by an H for hexidecimal.

C. A dash preceding a flag or a signal name indicates that it is active low.

The initialization routine begins at INITL: The I/O ports are set for their proper outputs. The flags 1 register is saved and the registers from 32H to 73H are zeroed except for the upper byte of the keyboard timer (register 4EH) which is loaded with an FFH. The RAM locations from 24H to 31H are checked against the CHECKSUM register at location 20H. If the numbers do not match, then the EEROM is read and its CHECKSUM is compared to its own registers. If the EEROM CHECKSUM is no good, the limits and CLC register pairs are set to FFFFH and the limits unknown flag is set. This will prevent the motor from operating until the unit is programmed. If the EEROM'S CHECKSUM matches the calculated CHECKSUM, this information is used to operate the GDO.

If the numbers match then an erroneous hardware reset occurred (A glitch?) and the RAM will be assumed to be good. The flags1 register is restored. An erroneous hardware reset may not have been preceded by the power down routine and thus the information that is in the EEROM may not match the information that is in the RAM.

The power down routine is located at STR:. At this point, the advanced power fail signal (APF) is polled. If the APF is active then it is debounced for 60 USEC. If the debounce fails then the program continues as if nothing ever happened. If the APF remains active then all relays are turned off along with the door condition indicator to minimize the current drain on the power supply capacitor. The flags1 register is stored into Sflags1 and the registers from 20H to 31H are stored into the EEROM chip. After storage, a delay of 10 MSEC occurs to allow the EEROM to complete the storage operation. The software finishes by resetting I/O pin D6 low to cause a hardware reset of the chip.

Almost all timing over 10 MECS is timed out from the A/C line frequency. The section of code that runs these timers starts at label start. The signal into pin P12 is a 16.6 MSEC square wave generated from secondary side of the transformer. The timer routine will begin by comparing the P12 input with the RTC flag (real time clock). If they match then the timer routine is skipped. If not, the RTC flag is inverted in order to match the present P12 input and the timer routine proceeds. Execution of the timer routine begins by reinitializing the input ports of port 2.

The timers are located sequentially in RAM. This is to allow RO to always be able to point to the next timer by a mere increment. Subroutines byteo (one byte) or bytet (two bytes) are used to see if a timer is active or not and, if active, increments the timer. Notice that a timer starts at a negative number and times out upon incrementing to zero. These subroutines will also signal the calling routine if the timer has timed out.

The timer routine will check each timer one at a time. There are 13 timers, 3 of them are 2 byte timers. Also, timers TIM30, T1M1, and Tkeys are dual function timers. That is, depending on the state of things. The timers are used for different purposes, i.e., upon the timeout of the TKY timer, The T/-F flag is checked to see if it was a five or ten second timer.

The timer functions are listed in the section fifteen of this paper. The timers will be discussed in context of the appropriate function, i.e., the 4.5 minute timer will be discussed with the light turn on routines.

The routine that controls the motor is located at main and main is executed if the SYSC flag is active. The motor routine begins by disabling the force routine and data interrupts so as to allow all status to be checked without changing any flags halfway through the status checkout. Along with disabling interrupts, the RPM fail timer (TIMRPM) is turned off.

Exiting the main routine can occur in two ways via the MAC:or MAC2: Labels. Execution of MAC: will result in the execution of MAC2: MAC:will activate a hardware interrupt in order to get the proper door direction. See section six. MAC2 deactivates the CHCKPT register and the SYSC flag.

The CHCKPT register is a double check of the SYSC

See the SYSC subroutine. Both the SYSC bit and the CHCKPT register must be set properly in order to activate the motor. This was done so that an erroneous setting of the SYSC bit (A glitch?) would not start the motor. The SYSC flag and the CHCKPT register are reset at the end of the motor routine. It is desirable to have both of them as close as possible to prevent a hardware reset glitch from deactivating one and not the other.

Upon going into the main routine, the motor is checked to see if it is going up or down. If it is going up, under all conditions, the motor is turned off. If the motor is going down and either the down limit is active (set by the limits routine) or the GDO is in program mode, the motor is turned off. Otherwise there has either been an obstruction or a command and the door must be reversed.

A reversal in the proceeds as follows. The reverse timer is enabled (T1MREV) for one-half second and the motor is turned off. Upon timeout (at TREV.), If the up limit is active nothing more happens. If the up limit is not active then the auto reverse flag is activated and the SYSC: routine is executed. This will eventually allow the motor routine to start the motor up.

Upon going into the main routine, if the motor is off, the light is activated to start the four and one-half minute timer. If the unit is in operate mode, the limits unknown flag is checked. If active then the motor is turned off (redundancy) and the routine is exited. The status is checked for the auto REV flag, down limit flag, and auxiliary obstruction flag being active. If any one of these are active then the motor is activated in the up direction as long as the up limit flag is not active.

Operation of the motor routine in program mode is as follows. If the door is going up or down the motor is turned off. If the motor is already off then program control is directed to INPRG1. At INPRG1 the limit flags are set according to the PUP/-PDWN Flag. The CCSW register is checked for either the vacation or worklight switch depression if the worklight switch is active the motor down relay is turned on. If the vacation switch is active the up motor relay is turned on. If neither switch is active the motor off subroutine is executed.

The motor subroutines (MOTOF:, MOTDWN:, and MOTUP:) are only called from the main routine. The light subroutines (LGHTON: and LGHTOF:) may be called from several spots in the program.

There are several pieces of code that are common to all five routines. Each routine starts out by disabling all interrupts and selecting RAM register bank one. The motor up and motor down routines will preceed this by checking the CHCKPT register and the RAM CHECKSUM to see if both are good. If not the motor will not be activated and the routine simply returns.

All five routines will finish by using the same piece of code. This last piece of code starts at CHT1:. All routines will have found the zero crossing of the A/C line voltage, loaded register 3 with the I/O information that is to be output to the DBUS port, and loaded the hardware timer with a number to allow the proper delay from the zero crossing of the line voltage to allow the particular relay to turn on at the proper time. After the timer times out, the bus information is output to the DBUS I/O port.

The subroutine that finds the zero crossing of the A/C line voltage is called tick. The tick: routine will remain in a loop until the real time clock input changes state. During this time the edge routine is continuously executed so as not to miss any edges coming in on the RPM input.

The up and down motor routines are also responsible for activating the one second and 42 second timers. The motor off routine reactivates the 42 second and one second timers as ten second and one and one-half second timers. Notice that the only way that the timer routines know if the timer is from the motor on routines or from the motor off routine is by reading the I/O port. If the motor is off then the timers must be the ten second and one and one-half second.

The light on routine activates the four and one-half minute timer. If the work light flag is inactive then the light is turned off otherwise the light is left on.

Detection of a change in the RPM input (T1) is detected by the edge: subroutine. The edge: subroutine debounces the RPM input and sets or resets the LRI flag (last RPM input) accordingly. A falling edge on the input will cause the CTRPM register to be incremented. It is the limits routine that will pick up the CTRPM register to increment or decrement either the CLC or limits register pair.

There are several circumstances where the direction of door travel cannot be assumed. That is, the door itself must be monitored for the proper direction. There are two clear cases of this problem. First is backup of the door at the down limit due to excessive door pressure. The second case is during a "SOFT" door reversal (no hard obstruction), even though the up relay is actuated, the door momentum may still take the door down until the motor has had enough time to counter act the door force.

The GETDR: subroutine will determine the door direction by detecting the phase relationship of the phototransistors on the RPM board. GETDR is executed in two ways, by hardware interrupt or by being polled from the edge subroutine. Upon motor turn on or turn off the STDR: subroutine is executed and enables a hardware interrupt on the RPM input. Upon motor turn on the direction is obtained by interrupt for one second. Upon motor turn off the direction is gotten by interrupt for one and one-half seconds and by polling from the edge: Subroutine for another eight and one-half seconds. The interrupt method is used when the motor speed is fast. In order to guarantee that the second phototransistor is in the proper state after the falling edge of T1 (note that this is the edge used to increment the CTRPM register for the limits routine), the other RPM input (P25) must be read very quickly in order to determine correctly the direction of travel. The problem with using the interrupt method is that data from the receiver data cannot be decoded at this time since the hardware timer must be used for both the data decode and the RPM interrupt. Thus, data decoding is disabled for one and one-half seconds after motor turn off. For the remainder of the eight and one-half seconds the GETDR subroutine will be polled since the motor is going very slow. This will allow the data decoding to take place.

When the GETDR: is executed via interrupts, the routine exits by loading the blank timer with a minus one. It is the blank timer routine that will execute the STDR: subroutine to reenable interrupts again to execute the GETDR subroutine. This will insure that on every next falling edge of the RPM signal the door direction flag, U/-D, will be set correctly.

Counting the RPM falling edges and managing the current limit counter (CLC) and the limits registers is done by the routine starting at MBYE. The CLC maintains the current position of the door. The limits register pairs remember what the total length of door travel is.

The limit routine starts by testing if the forty-two/ten second timer is active. If it is not active the CTRPM register is cleared and the limits routine is skipped. If it is active, the CTRPM register is examined. If the CTRPM is zero the limits routine is again skipped. Otherwise the count in CTRPM is used to adjust either the CLC or the limits register pair depending if the unit is in operate or program mode.

Assuming that the unit is in operate mode, the CTRPM count must be added or subtracted from the CLC. This determination is made from the U/-D flag that is set by the GETDR subroutine. If the door is going down, the CTRPM count is subtracted from the CLC. The down limit is hit by the CLC decrementing below zero. When the down limit is hit, the down limit flag is set, the open door indicator is turned off, the program force down flag (PFD) is reset and the SYSC subroutine is executed to allow the main routine to turn the motor off. The situation is similar with the up limit. When the U/-D flag is set, the CTRPM count is added to the CLC register. The up limit is hit when the CLC increments over the limits register count. When the up limit is hit the up limit flag is activated, the program force up (PFU) flag is reset, and the SYSC subroutine is executed to allow the main routine to turn the motor off.

The limit program is such to allow the user to program the limits of door travel easily. The prcgram can determine if the limits being programmed are "good" limits or not. A "good" limit is when the number in the limits register pair is between -80 (FFBOH) and 2303 (08FFH). 2303 counts represents a door travel of over twelve feet. If this count is exceeded (notice that an underflow will also look like a count that is too high) then the CLC and limits registers are automatically set to FFFFH and the limits unknown flag is activated. The -80 count is used to allow coast to accumulate when the registers are set to FFFFH.

The limit that is being programmed is indicated by the PUP/-PDWN flag in the flags1 register. This flag is set or reset while in program mode, in the motor routine, whenever the command switch is actuated. The control case LED routine located at tiled: (time LED) also uses this flag in program mode to flash the proper LED to indicate which limit being programmed. The work light LED indicates the up limit and the vacation LED indicates the down limit. The respective switches will also run the door in the indicated direction. The work light switch moves the door up and the vacation switch moves the door down. These switches will only move the door while they are depressed. Releasing them will cause the motor to stop. Pressing the command switch will change the PUP/-PDWN flag and thus changes the limit to be programmed.

Briefly programming the limits via the work light switch or the vacation switch is as follows.

The above chart is relatively self explanatory. The coast refers to when either the vacation or the work light switch is let go and the motor turns off and coasts to a halt. If the door is coasting the same direction as the indicated limit (Up limit and coasting up) then the limits register will be frozen and the coast added to the CLC. It is assumed that the coast in the programming of the limits will be the same as the coast when the door is runs in operate mode. If the door is going the opposite direction of the indicated limit (programming up limit and going down) the coast will be subtracted from the limits registers in order to further reduce the length of door travel.

During programming of the limits, whenever the limits number is changed a new CHECKSUM number is calculated and stored.

The 1" obstruction set switch will decrease the down limit every time it is depressed. This allows for automatic adjustment for the 1" obstruction test required by U.L.

Every time this switch is pressed a count of 12 is added to the CLC and the limits register pairs. This will effectively lower the down limit by 3/8". Also 12 will be added to the tweak register located at 30H in the RAM. If the door is installed properly, the down force will be programmed from the up limit to when the door just touches the floor. After setting the 1" reversal via the 1" obstruction switch, the door will push harder into the floor. If the forces ever need to be relearned, then the down force should be learned from the up limit until the door just touches the floor. During reprogramming of the down force, the forces are examined until the door is within this registers count of the down limit. At this time the force from here on out will not be considered for permanent storage.

The force routine has three modes. Operate mode, program force mode and programming the limits mode.

In the operate mode the force routine is run from one second after the motor is turned on until the motor is turned off again. The force routine must share the hardware timer with the data decode routine. This means that the force is only run during the blank time of the receiver. The force is measured by interrupting on a falling edge of the RPM1 input. Loading the hardware timer with number that was preset during the program force routine, interrupting from that timeout and looking again at the RPM input. If the input is high then the RPM frequency is faster then the timer and the force is ok. If the RPM input is low the RPM frequency is longer than the timer. The force has exceeded the programmed in value. Upon determination of an obstruction, the SYSC subroutine is executed and the motor routine will determine the proper motor action to be taken. This method of two hardware interrupts provides for a very accurate count of the RPM input signals low time.

The blank timer routine (TBLNK) enables the force routine by setting the edge/-timer (E/-T) flag. Activating the DO force flag (DOFCE), and enabling the hardware timer to interrupt on the next falling edge of the RPM signal. Upon the interrupt, the RPM routine will determine that the E/-T flag is set for edge. The flag will be reset and the RPM input will be debounced to determine if a noise spike caused the interrupt. If the signal is high, the routine exits. Otherwise, if the GDO is in operate mode, the proper force is loaded into the hardware timer and the timer is enabled as a timer. Upon timer overflow another interrupt will occur. The E/-T flag will be low so that the RPM input will be checked. If the RPM input is low, then the RPM signal's frequency is longer than the timer which indicates there is excessive force. The routine will execute the SYSC subroutine to signal an obstruction. If the RPM input is high then the RPM signal's frequency is faster than the timer and there is no obstruction detected.

If the GDO is programming the limits, an interrupt on the RPM edge will cause the hardware timer to be loaded with the maximum force. During programming of the limits, the force is not known so the unit is allowed to go to maximum force. There are two values for maximum force and the value used is determined by a hardware jumper on P24. When the timer overflows and causes an interrupt, the RPM input is checked. If the signal is low, then the force is too high and the SYSC subroutine is executed and the DCI will be turned on.

If the GDO is programming the force, an interrupt from the RPM edge will cause the timer to be loaded with the minimum force setting (approximately 4 MSEC). Upon the timer overflow interrupt occuring the low time is timed out. The actual time that is obtained is increased slightly to account for door changes, weather, etc. This force number is compared to the force number that is already stored. The heaviest count stays in the register as the force to be used during operate mode.

The up and down forces are programmed independent of each other. The force for a particular direction is to be programmed whenever its program force flag is active (PFU for up and PDF for down). The flags are activated whenever the unit is put into program mode or whenever either limit is adjusted via the work light or vacation switches. The program force flags are reset whenever the particular limit is hit. That is--after programming the limits, a command is given and the door runs from the up limit to the down limit. Upon hitting the down limit the PFD flag is reset.

There is an RPM fail timer (TIMRPM) that is used to detect if the RPM signal has stopped coming in altogether. Notice that the force routine just described will not be executed if the RPM input maintains itself at a DC level. TIMRPM is initialized for one-half second when the one second motor start timer (TIM1) times out. Every time the force routine times out the RPM low time (E/-T flag reset at the timer interrupt) then TIMRPM is reinitialized for another one-half second. If this timer ever times out, then the force routine has not been run in a half-second and thus a hardware failure has occurred. This will start the TIMDCI timer to blink the DCI LED to indicate to the user that a failure in the GDO head has occurred.

A. COMMAND, WORK, AND VACATION SWITCHES

The control case has several features that are served by a total of three wires. Ground, command, and the indicator wire is all that is needed to control three switches and three LEDS. This routine starts at EMAIN:.

The three switches are multiplexed over the command wire. Power to the control case also comes from this wire. If there is no key depression the command line will be high to the control case and low into Pll of the microprocessor. The software has written refering to the voltage level with respect to the control case and not the microprocessor input pin. All references will be made accordingly. If the command switch is pressed, the CC input is constantly low. The work light and vacation switches will generate a low frequency, low duty cycle waveform. The timing is approximately 10 MSECS low and 50 MSECS high for the vacation switch and 10 MSECS low and 25 MSECS high for the work light. The frequency was chosen in order to use the A/C line frequency to time out the switch depression. The vacation switch should take between five and eight counts of the A/C line. The work light will be between two and four. The low duty cycle was chosen because power to the control case comes from the command line and the oscillator circuit will only last a short time if power is removed. The low time was chosen according to the nyquist criteria for sampling a waveform. This input is polled.

Each key depression is debounced. The command switch is debounced after sixteen to twenty four milliseconds of being constantly low. The vacation and work light switches are debounced by getting four timeouts in a row of the same period. Thus the vacation switch will take 4*(10 MSEC+50 MSECS) or 240 MSECS to activate. All switches are debounced off by the command timer timing out. This will take approximately one-hundred milliseconds.

A key depression will start the control case timer (TIMCC) to start at negative nine. As long as the control case input is low, TIMCC continues to increment. Assuming that the command switch was pressed, eventually TIMCC increments three times. The switch routine detects this and turns off the TIMCC (sets it to zero) and executes the SYSC. Subroutine so that the motor routine will turn the motor on or off. Further holding down of the command switch will only keep the command flag in the CCSW register active.

If the key that was depressed was either the vacation or the work light the switch routines must detect the frequency of oscillation, debounce the switch, and activate the proper flags for other routines to examiner.

The frequency of oscillation is determined by examining the TIMCC timer at each falling edge of the control case signal as long as there are no active switches, a falling edge will activate the TIMCC to negative nine. TIMCC is incremented by the A/C line timer routine (at TCCT). Upon the next falling edge of the control case signal, TIMCC is examined for its content. A count of negative seven to negative five indicates the work light switch is pressed. A count of negative four to negative one indicates that the vacation switch has been pressed. Any other numbers will cause TIMCC to be reinitialized and the routine exited. After four consecutive time outs of the same period, the switch is considered debounced.

If the unit is in operate mode, then pressing the work light switch will flip the work light bit in the CCSW register and flip the incandesent lights on or off. If the vacation switch is pressed then the vacation bit in the CCSW register is flipped.

If the unit is in program mode and the motor is off, then pressing either the work light switch or the vacation switch will result in the flag in the CCSW register being set and the SYSC: subroutine being executed to start the motor in the proper direction.

Debouncing off the work light or vacation switch while in the program mode will result in the SYSC. Subroutine being executed in order to turn the motor off. Also the command, work light, and vacation flags in the CCSW are reset.

B. WORK AND VACATION LEDS

The routine that controls the control case LEDS is located at tiled. When in operate mode the CCSW register is examined to see if either the work light or vacation switches are active. If only one switch is active then the appropriate LED is turned on. Both switches being active means that both LEDS should be on. With only one wire to the control case this is not possible. What is possible is to alternately turn on each LED at a rate of 8.33 MSECS. this way both LEDS will look on.

If the unit is in program mode then either the vacation or work light LED is to BLINK at a rate of three times every two secons. If the PUP/-PDWN flag in the flagsl register is set then the work light LED is activated otherwise the vacation LED blinks.

Both the program mode and the 1" obstruction switches are momentary switches. The debounce routine is started at DBYE. This routine will check if the motor is on. If it is then the routine is skipped. If the motor is off then the switch inputs are polled to see if either one has changed state. If one has the new depressions are stored into the SW register and the debounce timer (TIMDCI) is started for 24-32 MSEC timeout.

The code at TSWDEB. times out the debounce timer and determines which switch changed state. If the 1" obstruction set switch changed the program determines if the switch is being pressed. If it is then the tweak register, the CLC registers, and the limits registers are incremented by a count of 12. If the switch is not being pressed then the routine exits.

If the program switch is active then the vacation and work flags in the CCSW register are reset. If the unit is going into operate mode the routine exits. If the unit is going into program mode the routine jumps to GOPROG. At GOPROG. the up or down limit flags are set according to the PUP/-PDWN flag, the PFD and PFU flags are set, the up and down forces are set to minimum and the D.C.I is turned off.

This routine is located at KBYE: and sets the AUX OBS flag in the flagsl register to signal to the motor routine that an auxiliary obstruction is occuring. A low signal of eight to sixteen milliseconds will set the AUX OBS flag. The same time is used to debounce the auxiliary obstruction off.

Activation of this flag while the motor is going down will result in the SYSC subroutine being called to activate a motor reversal. See TAUXOB:routine in the A/C line timer routine. If the motor is off and the AUX OBS flag is active then the motor routine will not allow the door to go down.Commands will only allow the motor to go up until the up limit is reached.

Data decoding is split into two routines. The interrupt routines will decode one code and signal to the housekeeping routine if the code was valid or invalid via the flags3 register. One code consists of one blank time followed by ten data bits. each bit can be one, two, or three times the length of the SYNC pulse. The sync pulse is the first bit to be transmitted after the blank time.

Data is not decoded one second after motor start up, one and one half seconds after the motor is turned off, during programming of the limits, and during programming of the force. In the first case, motor turn on, data is not decoded because of the RF dead spot effect. In both the motor turn on and turn off cases, the direction flag for the limits routine must be set properly by execution of the GETDR:subroutine via interrupts. Data decoding and GETDR: cannot be done simultaneously.

During programming of the force and programming of the limits the limits the transmitter is disabled only while the motor is on. During programming of the force, the data is not decoded in order to allow the force routine to find the longest RPM pulse width by looking at as many pulse widths as possible. There is no data decoding while programming the limits just happened to workout that way and there is not a good reason to add code to circumvent the situation.

The data code housekeeping routine is located at CHBITS. Four valid codes are needed to actuate a command. The valid code counter keeps incrementing on each valid code flag. If three invalid codes come in a row, the valid code counter is reset. As soon a four valid codes come in, if the unit is in operate mode, the SYSC subroutine is executed. If the unit is in program mode, the data stored in the TRIDT locations (56H to 58H) are stored into the TRIDAT locations (29H to 2BH) and the work light is flipped on or off. The transmitter is considered active as long as there is one valid code separated by no more than six invalid codes.

The program has been set up to easily add decoding for auxiliary codes. If an auxiliary output is desired, after four valid codes are acknowledged, the tenth bit or the channel bit will be in register 59H. The value can be zero, one or two.

The trinary data bits are stored in the TRIDAT locations (29H to 2BH). The TRIDT locations are only used as buffers during programming the data bits in (locations 56H-58H). The bits are stored in a compacted manner. The first location of TRIDAT, 29H, contains bits one, two, and three. Each data bit needs two bits to remember its value since there are one, two, and three millisecond pulse widths. The DATAIN register will assemble incoming bits. When three bits are in, the interrupt routine will compare it with the appropriate switch register, i.e. a typical bit pattern will look like this--SW1, SW2, SW3, 00. For a code of SW1=1MSEC, SW2=2MSEC and SW3=3MSEC.fwdarw.00, 01, 10, 00 or just 00011000. Notice that the last two bits contain no information.

Data decoding is done through hardware interrupts on the -INT PIN and interrupts by the hardware timer. The description that follows is the decoding for the SYNC pulse and the first data bit. The other bits are decoded by the same method.

To start decoding the blank timer (TBLNK routine) enables the interrupt pin. The timer routine also loads the blank timer for a timeout in sixteen to twenty-five milliseconds. Upon timeout, the blank timer will assume that a sync pulse and ten data bits have come in. Also the blank timer routine will activate the -SYNC find bit in flags4.

The first falling edge that occurs on the interrupt pin will be decoded as the SYNC pulse. The blank timer is restarted for a 8.33 to 16.66 millisecond timeout. The bit pointer, data pointer, and datain registers are initialized and the SYNC pulse is timed out in a software delay loop. The delay loop uses the hardware timer to accumulate the time. The timer increments every ninety-six microseconds. A SYNC pulse is only valid if it is between one-half and two milliseconds or a timer count between five and twenty-one. Any value other than that will cause the bit valid flag to be reset.

After a valid SYNC pulse is detected, the timer count is stored in the SYNC register for use in the decoding of the subsequent data bits. The INT pin is enabled and the interrupt routine is exited.

The next interrupt should be the first data bit. The blank timer is restarted again for 8.33 to 16.66 MSEC timeout. The SYNC register contents is gotten, two is added to it, and this value is loaded into the hardware timer. The timer is set to interrupt upon overflow. One hundred and ninety two microseconds was added to the sync pulse time to decode a data bit in order to compensate for pulse stretching and to give more of a window to decode data bits.

The next interrupt will occur from the hardware timer. The interrupt input pin is read. If it is still low then the datain register is incremented and tested to see if this bit has exceeded a three times sync pulse width. If it has, the bit valid flag is reset and decoding of this code frame halts. If it is still less than three times the sync pulse, the hardware timer is restarted again for the sync pulse time plus zero.

If the interrupt input pin is high when the hardware timer interrupt occurs, the datain register is shifted left twice. This will effectively saves this bit time. The bit pointer is incremented to see if three bits are in. If not, as in the case of the first data bit, interrupts are enable on the interrupt pin and the routine exits. If three bits are in then the bit pointer is reinitialized to negative three and the datain register is compared to wherever the data pointer is pointing to. If the data does not match then the bit valid flag is reset. Either way, data pointer is incremented to point to the next register and the datain register is cleared in order to assemble and compare the next three data bits that are to come in.

The only deviation that will occur from the above sequence is if the data is to be programmed in or if the code is at bit ten.

When programming in a new transmitter code, the first code to come in (determined by the valid counter equal to zero) is stored in the TRIDT locations (56H to 58H). Subsequent codes are compared to the first one. After four valid codes are found the housekeeping program will store the TRIDT locations into the TRIDAT locations. This method assures the microprocessor of storing only a valid transmission when in program mode.

Bit ten in a code sequence is very similar to program mode. Only the first valid code input will store bit ten. Subsequent codes will compare their bit ten to the stored bit ten for validity. Bit ten is stored in register 59H. This will allow future reprogramming of the microprocessor to enable auxiliary codes using the channel bit. using the channel bit.

The code assembling routine ends when bit ten is detected. That is, upon getting bit ten the bit valid flag is checked. If it is set, then the last code was good and the valid code bit flag is set. If the bit valid flag is reset then this bit is set again in order to decode the next code sequence and the invalid code flag is set. The valid code and invalid code flags are used by the housekeeping routine to increment the valid and invalid code counters. The blank timer is restarted to timeout in less than 8.33 MSEC so that the force routine can be enabled if necessary. Finally the blank time detect flag is reset. A blank time is valid only if this flag is not set before the blank timer times out.

The EEROM routines read and write registers 20H to 31H to and from RAM. Three RAM registers are used to aid in keeping the subroutines general. These are the EAINST, WORDPT, and eACT registers. The EAINST register has the instruction that is to be output to the EEROM. The WORDPT register holds the address of the microprocessor RAM location that is to be written into or read out of the EACT contains the number of register pairs plus two to be written or read.

Both the EEROM read routine (at EADAIN) and the EEROM write routine (at IWRITE) begin by outputting the appropriate instruction to the EEROM chip. The data is then clocked in or out. Upon completion, the data write routine simply returns to the calling routine. The data read routine will check the information from the EEROM chip to see if it is good or not. This is done by a CHECKSUM on locations 24H to 31H. If the information is good then SFLAGS1 is put into the Flags1 location and the CCSW register is stored into SCCSW location. If the CHECKSUM that is calculated does not match the CHECKSUM that was read then the CLC and the limits register pairs are set to FFFFH and the limits unknown flag is activated.

The keyboard routine allows the user to have a four digit code to actuate door motion. This code is set when the unit is in program mode by depressing the desired four key combination. The keyboard also acts as a panic switch when the door is in motion. If the door is moving, then the keyboard routine needs to see only one switch closure to execute the SYSC subroutine. This effectively will act like a door obstruction.

The keyboard routine has three timers associated with it. One timer is used to debounce the key depressions for twenty-five to thirty-two milliseconds. The other two timers use the same RAM locations (TKY). Obviously only one can be running at any given point in time. The active timer is identified by the ten/-five flag (T/-F). The ten second timer is started upon the first key depression. If three more keys are not entered within ten seconds, the flags and registers that are associated with the keyboard are reintialized. The five second timer is used as a keyboard lockout timer to frustrate anyone that would try to activate the door through trial and error of keying in code combinations. As soon as a bad key entry is made, the keyboard is ignored for five seconds. Anybody trying to guess the code would have no indication that the board was not being read and thus there is a very good chance that when keying in the proper code the previous bad code entry has got the keyboard disabled. The panic switch feature is still operational even if the keyboard is locked out.

RAM register bank one is always used for interrupt routines and for the motor and light turn on and turn off subroutines.

RAM locations 20H to 31H are used as the buffer to the EEROM. This information has a CHECKSUM on it located at register 20H. The CHECKSUM merely sums locations 24H to 31H, adds one to the sum and stores it. The one was added just in case the RAM ever comes up all zeroes. The sum of all zeroes would equal the CHECKSUM in that case.

The A/C line timers are located in sequential order to allow the timer routine to always access the next timer by incrementing the register pointer.

The register definitions are as follows.

The four and one-half minute timer is started by the Lighton: subroutine. Upon time out the light is turned off and the flag FO (program mode flag) is reset to indicate operate mode. This will prevent the user from forgetting to take the unit out of program mode.

The forty-two second timer is a dual timer. The motor up or motor down routines start the timer for forty-two seconds. If a limit is not hit then this will signal a command to the motor routine. Upon motor turn off, this timer is turned on again for another ten seconds. The limit routine will only count RPM input while this timer is active. The purpose of this is to prevent noise glitches from changing the limits while the unit is off.

The one second timer is also a dual purpose timer. Upon motor turn on, this timer prevents the error code counter or the valid code counter from incrementing. Effectively, this turns the data decoding off to prevent spurious operation due to RF dead spots. Also, the one second timeout will reset the limit flags and turn on the open door indicator. The second function is activated at motor turn off. The timer is activated for one and one-half seconds. This signals to the blank timer routine to enable the GETDR: subroutine via interrupts in order to count RPM input in the proper direction.

The reverse timer is activated by the motor routine in order to reverse the door. If the door is going down and a command is given then the motor is turned off for one-half second and the up motor relay is activated.

The switch debounce timer is a twenty-four to thirty-two millisecond timer used to debounce the program switch and the 1" obstruction switch. Time out results in the respective switch code being executed.

The control case timer is a sixteen to twenty-four millisecond timer used to debounce the command switch when the switch is activated. The timer is also used to determine the frequency of oscillation when either the vacation or work light switch is pressed. This timer is also used to debounce any of the three switches off. The debounce off time is one-hundred milliseconds.

The Auxiliary obstruction timer is a 8.33 to 16.66 millisecond timer used to debounce the auxiliary obstruction input.

The blank timer is used to time out between the transmitter blank time so that the force routine may be executed. This timer is nominally set for twenty-five to thirty-two milliseconds. There are other factors that can set or reset this timer. Execution of the GETDR: subroutine via interrupts always finishes by setting the blank timer to negative one (less than eight milliseconds). The data decoding routine will also set this timer to time out in eight to sixteen milliseconds on every falling edge that interrupts the processor.

The keyboard debounce timer is used to debounce the keyboard entries for twenty-four to thirty-two milliseconds.

The keyboard timer is a dual function timer, either five or ten seconds. See section 13 of this report.

The control case LED timer is used to turn on or off the LEDS. During program mode either the vacation or the work light LED is flashing at a rate of twice per two seconds. While in operate mode, if both the work light and vacation functions are active, then both LEDS are to be on. To make them appear on to the user each LED is flashed on and off alternately every 8.33 milliseconds.

The DC1 time, when activated, will blink the DCl LED at a rate of 3 times a second. The only way to disable this timer is to give the microprocessor a hardware reset.

After the garage door opener has been installed, it is placed in the program mode by moving the switch 38 to the program position. By checking the control unit 19 and determining which of the lights 28 or 31 are on, it can be determined whether the system is set for setting or up or down limit. If the light 28 is flashing, the switch 27 can be closed to cause the door to move down until the desired down limit of travel has been reached. When the door has been moved by the motor 135 to the desired down position the switch 27 is opened. Switch 25 is pressed to allow the up limit to be adjusted. Then the up limit can be set by closing the work light switch 29 which will cause the motor 135 to drive the door in the up position and it will continue to drive the door until the full up position is reached at which time the switch 29 should be released and the up limit will have been set. In the event the door moves past the desired up limit, it can be moved backward by closing switch 27 to the proper position.

When the up and down limits have been set, press momentarily switch 38 to place the machine in the operate mode. Then the press switch 25 which will cause the door to close if it has been in the full open position. After it is closed, then push the switch 25 againto cause the door to move to the full open position. During this sequence, of one complete cycle after the up and down limits have been set the force limits will be automatically set by the microprocessor. The forces are automatically set at a force of approximately 10% above that required to move the door in the up and down positions and these cannot be changed by the operator to cause the limits to be set at a dangerous and unsafe force.

Thus, the open and closed forces are set automatically and the garage door opener is ready for a normal operation if the force is to be reset at a later time you would press and release the program/operate mode switch 38 to enter the program mode. Press and release again to return to the operate mode. Press the wall control unit switch button to run the door through a full cycle and the force will automatically be reset.

In the event the door will not operate when programming limits or force check the door condition indicator light on the back panel of the garage door operator which is the light 37. If this light is continuously on, the garage door may not be operating properly as for example the spring force may be too high and it may be necessary to disconnect trolley and move the garage door up and down manually. If this light is flashing it indicates that the garage door opener has a fault.

The one inch obstruction switch 201 when depressed lowers the down limit a fixed amount as for example, 3/8 inch.

Although the invention has been described with respect to preferred embodiments, it is not to be so limited as changes and modifications can be made which are within the full intended scope of the invention as defined by the appended claims.

The following is the program. ##SPC1##