Loading learning content...
In the previous page, we explored how amplitude modulation carries information by varying a carrier wave's amplitude according to a continuous message signal. But in the digital age, we rarely transmit continuous signals. Instead, we transmit discrete symbols—ones and zeros, or more complex multilevel patterns.
Amplitude Shift Keying (ASK) applies the principles of AM to digital data. Rather than smoothly varying the amplitude according to an analog signal, ASK switches (or 'shifts') the carrier amplitude between discrete levels to represent different digital values. This is the bridge between classic AM and modern digital communications.
By the end of this page, you will understand how ASK encodes digital data onto carrier waves, analyze binary and multilevel ASK variants, explore bandwidth and bit rate relationships, examine signal constellation representations, and evaluate ASK's role in modern communication systems.
Before diving into ASK specifically, let's establish the terminology and concepts common to all digital modulation schemes.
Key Definitions:
The Relationship:
$$\text{Bit Rate} = \text{Symbol Rate} \times \log_2(M)$$
Where M is the number of distinct symbols (amplitude levels in ASK).
Example:
By increasing the number of amplitude levels, we can transmit more bits per symbol—at the cost of increased susceptibility to noise.
| Term | Definition | Unit | Example |
|---|---|---|---|
| Symbol | Basic unit of transmission | One amplitude level | |
| Symbol Rate | Symbols per second | Baud | 1000 baud = 1000 symbols/sec |
| Bit Rate | Bits per second | bps | 2000 bps with 4-level ASK at 1000 baud |
| Bandwidth | Frequency range occupied | Hz | Proportional to symbol rate |
| Signal-to-Noise Ratio | Signal power vs. noise power | dB | Higher = fewer errors |
A common source of confusion: baud rate and bit rate are only equal when each symbol represents exactly one bit (binary signaling). With multilevel modulation, bit rate exceeds baud rate. The baud rate determines bandwidth requirements; the bit rate determines throughput.
Binary ASK (BASK) uses two amplitude levels to represent binary data—typically one non-zero amplitude for '1' and another amplitude (often zero) for '0'.
Mathematical Representation:
$$s_{BASK}(t) = A(t) \cdot \cos(2\pi f_c t)$$
Where: $$A(t) = \begin{cases} A_1 & \text{for binary '1'} \ A_0 & \text{for binary '0'} \end{cases}$$
Common Binary ASK Variants:
Unipolar ASK (also called OOK):
Bipolar/Symmetric ASK:
The unipolar variant (OOK) is most common because it's simpler to implement and demodulate.
Bandwidth Considerations:
The bandwidth of a BASK signal depends on the symbol rate and pulse shaping:
$$BW_{minimum} = \frac{R_s}{2} \quad \text{(theoretical minimum, Nyquist)}$$
$$BW_{practical} = R_s \quad \text{(first null bandwidth)}$$
$$BW_{Carson} = 2 \times B \times (1 + \beta)$$
Where:
For rectangular pulses (worst case), the first null bandwidth equals the symbol rate.
Example Calculation:
To transmit 10,000 bits per second using BASK:
On-Off Keying (OOK) is the most widely implemented form of ASK, where the carrier is simply turned on for '1' and turned off for '0'. Its simplicity makes it the foundation for many practical systems, from infrared remote controls to fiber optic communications.
Mathematical Definition:
$$s_{OOK}(t) = \begin{cases} A \cos(2\pi f_c t) & \text{for binary '1'} \ 0 & \text{for binary '0'} \end{cases}$$
This can be written compactly as:
$$s_{OOK}(t) = d(t) \cdot A \cos(2\pi f_c t)$$
Where d(t) is the digital signal (0 or 1).
| Application | Carrier Medium | Data Rate | Range |
|---|---|---|---|
| Infrared Remote Controls | IR light (940nm) | 1-115 kbps | ~10 meters |
| RFID (Passive Tags) | RF (125 kHz - 13.56 MHz) | 10-100 kbps | ~1 meter |
| Fiber Optic (Short Range) | Light (850-1550nm) | 100 Mbps - 10 Gbps | km scale |
| Wireless Sensors | RF (433 MHz, 2.4 GHz) | 1-100 kbps | 10-100 meters |
| Visible Light Communication | LED light | 1-10 Mbps | Meters |
In fiber optic systems, OOK is exceptionally practical because turning a laser or LED on/off is straightforward. The absence of complex phase or frequency control simplifies transmitters. While more advanced modulation schemes offer higher spectral efficiency, OOK remains the workhorse of optical networks due to its implementation simplicity.
Multilevel ASK (M-ASK) extends the concept to M distinct amplitude levels, allowing each symbol to represent log₂(M) bits.
Mathematical Representation:
$$s_{M-ASK}(t) = A_i \cdot \cos(2\pi f_c t), \quad i \in {0, 1, 2, ..., M-1}$$
Common M-ASK Variants:
4-ASK (4-level):
8-ASK (8-level):
16-ASK (16-level):
The Noise Challenge:
The fundamental problem with M-ASK is that as we increase M, the amplitude levels get closer together. With a fixed maximum amplitude (limited by transmitter power), doubling M roughly halves the distance between adjacent levels.
Minimum distance between levels:
$$d_{min} = \frac{A_{max} - A_{min}}{M - 1}$$
Since the probability of error depends on the ratio of level spacing to noise power, more levels means dramatically higher error rates for the same SNR.
SNR Penalty for M-ASK:
Compared to binary ASK, M-ASK requires additional SNR to maintain the same bit error rate:
$$\text{SNR Penalty (dB)} \approx 10\log_{10}\left(\frac{(M-1)^2}{3\log_2(M)}\right)$$
For 4-ASK vs binary: ~5 dB penalty For 8-ASK vs binary: ~10 dB penalty For 16-ASK vs binary: ~14 dB penalty
This severe penalty limits pure M-ASK to low-order modulation in practice.
Pure high-order M-ASK is seldom used because it wastes the phase dimension. Modern systems use QAM (Quadrature Amplitude Modulation), which varies both amplitude AND phase, placing symbols on a 2D constellation. A 4x4 grid of 16-QAM symbols is far more noise-resistant than 16 amplitude levels on a line.
A constellation diagram is a powerful visualization tool that plots modulated signal points in a 2D space defined by the in-phase (I) and quadrature (Q) components. For ASK, all points lie on the real (I) axis since only amplitude varies.
Understanding the I-Q Plane:
Any sinusoidal signal can be expressed as:
$$s(t) = I \cdot \cos(2\pi f_c t) - Q \cdot \sin(2\pi f_c t)$$
Where:
The resultant amplitude: $A = \sqrt{I^2 + Q^2}$ The phase: $\phi = \arctan(Q/I)$
For pure ASK:
| Modulation | Number of Points | Constellation Shape | Bits per Symbol |
|---|---|---|---|
| OOK (2-ASK) | 2 points | 0 and A on I-axis | 1 |
| 4-ASK | 4 points | -3A, -A, +A, +3A on I-axis | 2 |
| 8-ASK | 8 points | 8 evenly spaced on I-axis | 3 |
| 16-ASK | 16 points | 16 evenly spaced on I-axis | 4 |
Interpreting the Constellation:
Point Position: Each point represents a valid transmitted symbol. Its I-coordinate indicates the amplitude.
Distance Between Points: Larger distances mean better noise immunity. Points closer together are more easily confused by noise.
Decision Regions: The receiver divides the I-Q plane into decision regions. A received signal falling in a region is decoded as that symbol.
Noise Effect: Noise adds a random displacement to each received point. If the displacement exceeds half the distance to an adjacent point, an error occurs.
The ASK Limitation is Visible:
Looking at the constellation, ASK's fundamental limitation becomes clear: all points are on a line. We're using only one dimension (amplitude) while leaving the orthogonal dimension (phase) completely unused. This is why QAM—which uses both dimensions—is far more common in modern systems.
Constellation diagrams are the universal language of digital modulation. Master reading them: horizontal axis = in-phase amplitude, vertical axis = quadrature amplitude. For ASK, all points are horizontal. For PSK, all points are on a circle. For QAM, points form a grid. The spacing tells you noise resistance; more spread = better.
Understanding how ASK signals are generated and recovered is essential for practical implementation.
ASK Modulator (Transmitter):
The simplest ASK modulator multiplies the digital signal by the carrier:
$$s_{ASK}(t) = m(t) \cdot A_c \cos(2\pi f_c t)$$
Where m(t) is the digital baseband signal (pulse-shaped sequence of symbols).
Implementation Options:
Coherent detection multiplies the received signal by a synchronized local carrier. For a signal A·cos(2πfct), multiplying by cos(2πfct) gives A·cos²(2πfct) = A/2 · [1 + cos(4πfct)]. The low-pass filter removes the double-frequency term, leaving A/2—directly proportional to the transmitted amplitude.
The Bit Error Rate (BER) quantifies the reliability of a digital communication system—the probability that a transmitted bit is received incorrectly.
For Binary ASK with Coherent Detection:
$$P_b = Q\left(\sqrt{\frac{E_b}{N_0}}\right)$$
Where:
For Binary ASK with Non-Coherent (Envelope) Detection:
$$P_b = \frac{1}{2}e^{-\frac{E_b}{2N_0}}$$
At the same E_b/N_0, non-coherent detection performs approximately 1-3 dB worse than coherent.
| Target BER | Required E_b/N_0 (dB) | Practical Meaning |
|---|---|---|
| 10⁻¹ (1 in 10) | ~2 dB | Barely intelligible voice |
| 10⁻³ (1 in 1,000) | ~6.8 dB | Acceptable voice quality |
| 10⁻⁵ (1 in 100,000) | ~9.6 dB | Good data transmission |
| 10⁻⁶ (1 in million) | ~10.5 dB | Typical data requirement |
| 10⁻⁹ (1 in billion) | ~12.5 dB | Critical/financial data |
BER Comparison Across ASK Variants:
| Modulation | E_b/N_0 for 10⁻⁶ BER | Spectral Efficiency | Notes |
|---|---|---|---|
| OOK (Non-Coherent) | ~13.5 dB | 1 bit/Hz | Simplest implementation |
| OOK (Coherent) | ~10.5 dB | 1 bit/Hz | Requires carrier sync |
| 4-ASK (Coherent) | ~15 dB | 2 bits/Hz | 4.5 dB penalty vs binary |
| 8-ASK (Coherent) | ~20 dB | 3 bits/Hz | Severe noise penalty |
Key Insight:
As we increase the number of levels in M-ASK, spectral efficiency improves (more bits per Hz) but power efficiency degrades (more E_b/N_0 needed). This fundamental trade-off defines modulation scheme selection in system design.
In real systems, imperfections like timing jitter, carrier frequency offset, and intersymbol interference create a 'BER floor'—a minimum error rate that cannot be overcome simply by increasing power. System design must account for these implementation losses, typically 1-3 dB beyond theoretical requirements.
We've thoroughly explored ASK as the digital extension of amplitude modulation. Let's consolidate the key concepts:
What's Next:
We've examined OOK as a specific, important case of ASK. The next page expands on On-Off Keying in greater depth—its mathematical properties, practical implementations, detector design, and widespread applications in optical and wireless systems.
You now understand Amplitude Shift Keying—how digital data maps to amplitude levels, the trade-offs of binary vs. multilevel variants, constellation representations, and error rate performance. This foundation prepares you for understanding both OOK specifics and the evolution toward more advanced schemes like QAM.