QAM Modulation and OFDM Symbols
To transmit data a carrier signal is jiggled around to represent binary 1’s and 0’s – a process called “modulation”. 802.11ad and 802.11ac differ in the degree to which they use the most complex types of modulation. The simplest modulation method is called Binary Phase Shift Keying (BPSK) – bits are represented by shifting carrier phase. In BPSK, one phase shift (or absence of shift) represents 1 bit – BPSK is a simple modulation technique. Bear in mind that simple modulation techniques are more immune to noise disruption than more elaborate schemes. Quadrature Amplitude Modulation (QAM) represents bits by a combination of amplitude and phase shifting – much more elaborate than BPSK. We’ll discuss the implications of 802.11ac having “256-QAM” capabilities while 802.11ad provides the lower bit density “QAM-16” maximum modulation rate. With QAM modulation a group of bits are encoded into a single RF “constellation” and transmitted as a single “pulse”. The more bits represented in a QAM constellation, the more subtle the changes in phase and amplitude of the RF signal “pulse” that’s transmitted. Subtle changes in phase and amplitude can only be recovered (i.e. received properly) if there is very little noise and interference in the environment. As the number of bits represented in a QAM constellation increases the requirements for low noise and low interference increase. If a transmitter doesn’t get ACK’s back for transmitted data then it reduces the modulation complexity. A high-quality radio circuit is required to recover the highest levels of QAM modulation (i.e. a more expensive radio device).
The term “QAM-16” means that 16 different bit patterns can be represented in a single transmitted constellation. The 16 binary values 0000 through 1111 can be represented in a single QAM-16 constellation hence 4 bits per symbol. The term “64-QAM” means 64 bit patterns are represented per constellation : 000000 through 111111 hence 6 bits per constellation. The term “256-QAM” means that an entire 8-bit byte can be represented by a single QAM constellation.
To construct a QAM constellation requires that the circuitry adjust both the phase and amplitude of each wave cycle of the transmitted carrier. Adjusting on a per-cycle basis at 2.4 or 5 GHz frequencies is significantly easier than trying to keep up with a 60 GHz frequency. 802.11ac offers up to 256-QAM modulation which can be done because it operates under 6 GHz. Because of the challenges of “keeping up” with a 60 GHz carrier, 802.11ad only offers 64-QAM modulation. Remember that modulation is applied to each carrier wave cycle and a 60 GHz carrier provides 10 times more cycles to modulate than a 6 GHz carrier. There are economic reasons to limit 60 GHz modulation – the price is kept competitive.
In addition to the raw modulation method the actual data bit stream is mathematically adjusted to maximize the probability of successful recovery. Algorithms referred to as “convolution”, “coding” and “Forward Error Correction” (FEC) are applied to the data prior to modulation. In addition to the pre-modulation bit manipulation a number of other techniques are used to improve recovery. When transmitted, multiple QAM constellations are assembled into a single Orthogonal Frequency Division Multiplexing (OFDM) “symbol”. The individual QAM constellations in an OFDM symbol are each constructed using a separate, narrowband carrier frequency called a “subcarrier”. OFDM symbols include some number of “pilot subcarriers” that are used to correlate and synchronize the receiver with the transmitter. Different applications of OFDM (i.e. different standards) specify a different number of pilot subcarriers in each symbol. Pilot subcarriers are overhead in the transmission process – they carry no useful end-user data. There’s an engineering trade-off between increased bit recovery (more pilot subcarriers) and increased bit density (more QAM carriers) in an OFDM symbol.
Let’s summarize: An OFDM symbol consists of multiple QAM subcarriers and interspersed pilot subcarriers. Higher QAM modulation (256-QAM) encodes a higher number of bits but is more sensitive to corruption by noise and interference.