A deeper look inside common Consumer Infrared (CIR)

A common CIR receiver chip includes more than a simple infrared photo-transistor (usually at 950nm). It includes demodulation circuitry around a specific carrier frequency (typically 38kHz) and a latch circuit for its data output pin that is triggered by that frequency demodulation. The extra circuitry is included to help block out ambient light and noise to make the incoming signal more reliable. When a button is pressed, the IR LED in a remote control turns on and off in specific timing patterns to transmit the button press data to the receiver. The "on times" aren't really fully on but rapidly turned on and off at the carrier frequency (38kHz modulation in this example). If the IR LED "on times" were fully on (like looking at a flash light), the receiver chip would think this was ambient light and ignore it. Since the "on times" are rapidly fluctuating (modulated) at 38kHz, the receiver chip's demodulation circuit triggers the latch circuit and it triggers the data pin. When the 38kHz modulated IR light turns off (the button on the IR remote control is released), the receiver chip's demodulation circuit releases the latch circuit and it releases the data pin.

To make things a little more complicated for hardware building and software programming, the CIR receiver chip's latch circuit doesn't respond immediately to any changes. This is done intentionally to help avoid noise problems and increase signal reliability. Most receiver chips' latch circuits won't activate the data pin until about a dozen pulses at the carrier frequency are received. When the IR carrier frequency is stopped, the receiver chip's latch circuit will also hold the data pin active for several to a dozen more pulses in length. This helps make up for the initial delay and will not give spurious data signals if a few pulses of the carrier frequency are missed by the demodulation circuit. The latch circuit working this way makes for clean data pulses on output.

The latch circuit delays will cause problems for critical timing applications. If the IR carrier burst is too short (under 12 cycles at 38kHz in this example), the latch circuit will likely not trigger its data pin. If the application expects short signals off the data pin, the latch circuit may hold the data pin for too long and cause a timing error again. If the latch circuit does not release in time for a short space and another IR carrier burst comes in, the output data pin will be held high for the entire time and make those two IR carrier bursts look like one. Different receiver chips (even of the same model) also have differently timed lengths for their latch circuits and may cause slight variations in the data pin time.

The specific demodulation circuitry in the CIR receiver chip doesn't mean that it will only see one carrier frequency. It will likely respond to 10-20% higher or lower than the primary carrier frequency (38kHz in this example), but the physical distance of that response will decrease as the frequency shifts away. Frequencies that are more off will mean that the data pin latch is less likely to be triggered by the demodulator. This is why it is important to keep application frequencies properly matched between hardware.

IR LED transmitters usually don't have frequency problems since they are such simple and rapidly responding devices. The transmitters also do not have to decode data from a potentially unknown and unreliable source like the receivers do. So long as the transmitting hardware is properly clocked and has all its data ready to go, it's hard to screw up a transmitter. For all the reasons given so far, this is why some IR hardware can transmit a short burst protocol but not receive it (the receiver has to be rated for the short burst protocol).