This is an infrared gate with two sensors planned to use in the wall in the way behind a door. It can be applied in a toilet to keep track of that someone is inside exceeding a certain amount of time. After that time elapsed, the circuit triggers the digital output wich can turn on a ventillator. The time period the output is turned on can be separately controlled by a second timer.
If you plan to build this circuit, beware that you may have lots of difficulties though the schematic may seem simple. The construction of the circit requires some amount of equipment like an oscilloscope and a DVM, too. Without them, the device will do weird things you wouldn't expect, and even if it is correctly put together, you must adjust it with care both mechanically in its final place and electronically with the help of an oscilloscope. Only if you want to span about less than 20-30 inches with the infra diodes can forget about this calibration. Alternatively you can take ideas from this construction.
The device consists of several parts, the most critical one is the panel with the infra LEDs. I tried to use several receiver transistors, but best result was given by infra receiver diodes used in TV remote control receivers. The receiver diodes must be properly shielded from the transmitter LED(s) otherwise the infra light will surely drive the receiver with a large enough signal. These photodiodes should only see infrared light coming from the mirror. The two very sensitive receiver parts should also be isolated from the transmitter electrically or the TX signal will get across the wires to the RX lines, which results the same effect as weak optical shielding. Use metal shielding around the receiver amplifiers where possible. The infrared transmitter LEDs should be close in wavelength to the max. sensitivity band of the receivers. You can experiment with using more LEDs and more current testing several resistor values, but don't exceed the 500 mA current limit flowing on the diodes or they will burn out. Do not shield the transmitters, allow the maximum amount of infralight to reach the mirror to extend the possible range.
Update: The TTL IC called '74123' is actually a 74HC123, and is used out of the specifications - although it is rated up to 7VDC supply voltage as absolute maximum, in this project it is powered from VCC=9-12VDC without problems. I am ashamed of neglecting such an important aspect.
To start testing the infra LED panel,
you wil need the infragate amplifier panel and the small
transmiter driver. The TX driver will generate the digital signal
for the LED driver on the LED panel. The digital signal is 1:10
on/off to achive good performance with lower power dissipation on
the LEDs. Connect GND, VCC planes and LEFT, RIGHT wires of the
LED panel with the amplifier panel, and drive the TX line from
the TX driver. Now you are able to start testing and calibrating
the analogue part of the circuit. If everything is ok, holding a
mirror in front of the LED panel will reflect enough signal to
overdrive the amplifier and you can check the output on the OPA
1, 7 pins with an oscilloscope. Taking the mirror farther on will
result a weakening signal on the amplifier output. Set the
orientation of the diodes to be able to get the maximum signal
amplitude on the oscilloscope screen. This is the heaviest part
of the work, don't deal too much with it until the complete
circuit is not built. Just adjust a static state of the
construction to give the maximum signal amplitude on the output
when nothing is between the diodes and the mirror and give a
small noise only when the line of sight is covered. If you are
ready with it, you can adjust the schmitt triggers built of the
other two OPA parts to generate TTL pulses when the analog signal
is at its maximum and stay on the same DC level when the received
signal is missing.
It is also important to protect the receiver diodes from direct light as natural light will weaken the sensitivity of the diodes, and lamps will transform the 50/60 Hz modulation present in the line power. Small noise is not problem, but the received signal from the TX generator should be stronger to be able to detect it. After the ST adjustments, connect LEDs to the 74123's TTL outputs through proper value resistors. The 74123 here is used as a demodulator. If there is a periodic signal change on the input, the output will be high, while if there is no activity on the input for a given period of time, the output falls low. When you cover the line of sight of one receiver diode, the corresponding LED turns off. There should not be any flickering in the turning on/off, the output should immediately respond to the change without blinking.
If still everything is correctly working at this point, the remaining digital circuit is the easy part of the work. The outputs of the previous circuit (LEFT, RIGHT) directly connect to the remaining part. The RS memory built from two NAND gates remembers the way of the last movement direction, so if someone is in or not. If you experience problems, connect another LED to pin 10 of the RS and check if this part does what it should. If there was any activity in the past minutes, the first timer is running, but it can only trigger the second timer part, if someone is still inside. The diode from the second timer output prevents resetting itself before the timing period is over in case of another movement. For a 1 minute timing (first timer) R=470k C=100u can be used, the second part would use R=1.5M C=470u for about a 15 minute timing (t=1.1RC). The output of the second timer (pin 9) can drive a relay activating the ventillator.