Formula sheet

Digital Modulation and Coding Formula Sheet

Digital communications formulas for symbol rate, coding rate, spectral efficiency, Eb/N0, EVM, BER, packet error, OFDM overhead, cyclic prefix, and MCS throughput.

Branch: Telecommunications Engineering
Content: Formula sheet
Updated: Jul 03, 2026
Revision: v1.1.0 · reviewed

This formula sheet collects first-pass calculations for digital modulation and coding. Use it to estimate symbol rate, occupied bandwidth, coding overhead, spectral efficiency, signal-to-noise requirements, error metrics, OFDM timing, cyclic-prefix margin, and modulation-and-coding mode throughput.

The equations are screening tools. Real systems require standard-specific parameters, measured channel conditions, receiver implementation loss, synchronization performance, nonlinear distortion, phase noise, interleaving behavior, retransmission rules, regulatory limits, and service acceptance tests.

How to Use This Formula Sheet

Use this sheet to connect modulation order, coding rate, bandwidth, energy metrics, receiver impairments, packet performance, OFDM timing, adaptive mode selection, and service validation. Start by defining the standard or waveform, payload boundary, coded boundary, symbol rate, occupied bandwidth, modulation order, code rate, receiver bandwidth, target BER or PER, channel model, impairment source, and measurement method. Then decide whether the calculation supports mode selection, link acceptance, receiver troubleshooting, throughput planning, or field validation.

Work through the formulas in this order:

Separate payload, coded, line, symbol, PHY, MAC, and service rates before calculating throughput.
Check occupied bandwidth, spectral efficiency, Shannon gap, $E_b/N_0$ , $E_s/N_0$ , SNR, receiver noise, EVM, BER, and PER with the same rate and bandwidth basis.
Check coding, OFDM overhead, cyclic-prefix margin, adaptive modulation, interleaving, latency, and retransmission effects before claiming user service performance.
Compare simulated thresholds with measured SNR, EVM, packet errors, spectrum mask, channel delay spread, phase noise, frequency error, IQ imbalance, interference, and field traffic load.
State the release action: accept mode, reduce MCS, add margin, change coding, adjust guard interval, retune receiver, reduce offered load, or hold release.

Do not treat raw PHY throughput as user throughput. Pilots, guards, coding, retransmissions, scheduling, headers, link adaptation, queueing, and fallback modes can consume large parts of the nominal rate.

Basis and Validity Limits

The formulas below are first-pass screens. They assume that modulation, coding, bandwidth, receiver filter, noise reference, overhead, channel model, measurement setup, and service acceptance criterion are known.

$E_b/N_0$ , SNR, and EVM comparisons are valid only when rate, bandwidth, reference plane, coding rate, target error rate, and impairment model match. A value taken from a standard, simulation, or lab receiver may not apply to a different implementation or channel.

BER-to-PER estimates are weak when errors are bursty, correlated, interleaved, coded, retransmitted, or caused by packet collisions, buffer overflow, clock slips, fading notches, or interference. Service impact must be measured at the packet or application boundary when that is the requirement.

OFDM cyclic-prefix and MCS formulas are valid only when delay spread, Doppler, carrier frequency offset, phase noise, channel estimation, equalization, pilot density, implementation loss, and traffic load are included in the acceptance evidence.

Symbols and Conventions

Symbol	Meaning	Typical units
$R_b$	bit rate	bit/s
$R_p$	payload bit rate	bit/s
$R_s$	symbol rate	symbol/s
$M$	modulation order	dimensionless
$k$	bits per symbol, $\log_2(M)$	bit/symbol
$R_c$	coding rate	dimensionless
$B$	bandwidth	Hz
$\eta$	spectral efficiency	bit/s/Hz
$SNR$	signal-to-noise ratio	linear or dB
$E_b/N_0$	energy per bit to noise density	linear or dB
$E_s/N_0$	energy per symbol to noise density	linear or dB
$BER$	bit error rate	dimensionless
$PER$	packet error rate	dimensionless
$T_u$	useful OFDM symbol time	s
$T_{cp}$	cyclic-prefix time	s
$\Delta f$	OFDM subcarrier spacing	Hz

State whether a rate is raw, coded, payload, line, physical-layer, or application throughput. Many design errors come from comparing rates that use different boundaries.

Modulation Order and Symbol Rate

Bits per ideal modulation symbol:

k=\log_2(M)

Uncoded bit rate:

R_b=kR_s

Symbol rate:

\displaystyle R_s=\frac{R_b}{k}

Payload rate with coding and overhead:

\displaystyle R_p=\frac{kR_sR_c}{1+\alpha_{oh}}

where $\alpha_{oh}$ represents pilots, framing, guards, control fields, training, and protocol overhead expressed as a fraction of payload.

Gross coded bit rate required for a target payload rate:

\displaystyle R_{coded}=\frac{R_p(1+\alpha_{oh})}{R_c}

Required symbol rate:

\displaystyle R_s=\frac{R_{coded}}{k}

These relations assume a single-carrier or aggregate symbol stream. Packet systems, OFDM systems, spread-spectrum systems, and framed waveforms require the same boundary discipline but often use more detailed standard-specific accounting.

Occupied Bandwidth and Spectral Efficiency

Raised-cosine occupied bandwidth approximation:

B\approx R_s(1+\alpha)

where $\alpha$ is roll-off factor.

Payload spectral efficiency:

\displaystyle \eta_p=\frac{R_p}{B}

Ideal coded spectral efficiency before overhead:

\eta_{coded}\approx kR_c

Efficiency after overhead:

\displaystyle \eta_{net}\approx \frac{kR_c}{1+\alpha_{oh}}

The result is not a guarantee of service throughput. Packet headers, retransmissions, scheduling gaps, acknowledgements, adaptive fallback, and queueing can reduce user throughput below the physical-layer estimate.

Shannon Capacity and Gap

Ideal channel capacity:

C=B\log_2(1+SNR)

Required ideal SNR for a target spectral efficiency:

SNR_{ideal}=2^\eta-1

Implementation gap:

G_{dB}=SNR_{required,dB}-10\log_{10}(2^\eta-1)

The gap includes nonideal coding, modulation, receiver architecture, synchronization, phase noise, equalization limits, nonlinear distortion, interference, fading margin, and measurement uncertainty.

Eb/N0, Es/N0, and SNR

Linear conversion:

\displaystyle \frac{E_b}{N_0}=SNR\frac{B}{R_b}

In decibels:

\displaystyle \left(E_b/N_0\right)_{dB}=SNR_{dB}+10\log_{10}\left(\frac{B}{R_b}\right)

Symbol energy relation:

\displaystyle \frac{E_s}{N_0}=k\frac{E_b}{N_0}

In decibels:

\left(E_s/N_0\right)_{dB}=\left(E_b/N_0\right)_{dB}+10\log_{10}(k)

For coded systems, be explicit about whether $R_b$ is coded bit rate or payload bit rate. Required $E_b/N_0$ values from standards or simulations often depend on coding rate, target error rate, channel model, and receiver implementation.

Receiver Noise and Required Signal Level

Thermal noise floor:

N_{dBm}\approx -174+10\log_{10}(B_{Hz})+NF_{dB}

Required received power:

P_{req,dBm}=N_{dBm}+SNR_{req,dB}+M_{impl,dB}

Operating margin:

M_{op}=P_{rx,dBm}-P_{req,dBm}

Bandwidth must match the receiver noise bandwidth. Using channel bandwidth, occupied bandwidth, sample bandwidth, and noise-equivalent bandwidth interchangeably can create several dB of error.

EVM and SNR Screening

RMS error vector magnitude:

\displaystyle EVM_{rms}=\sqrt{\frac{P_{error}}{P_{reference}}}

Approximate SNR from EVM:

\displaystyle SNR\approx \frac{1}{EVM_{rms}^2}

In decibels:

SNR_{dB}\approx -20\log_{10}(EVM_{rms})

This is a screening relation. EVM can include noise, phase noise, frequency error, IQ imbalance, nonlinear distortion, quantization, channel estimation error, and residual equalization error. It should be interpreted with the measurement setup.

Bit Error and Packet Error

Measured bit error rate:

\displaystyle BER=\frac{N_{bit\ errors}}{N_{bits\ tested}}

Approximate packet error rate for independent bit errors:

PER=1-(1-BER)^{N_{bits,packet}}

For small $BER$ :

PER\approx BER\cdot N_{bits,packet}

Packet success probability:

P_{success}=1-PER

The independent-error model can be weak in fading, burst noise, buffer overflow, clock slips, packet collisions, and interference. Interleaving, coding, retransmission, and packet length all change the observed service impact.

Coding Rate and Redundancy

Coding rate:

\displaystyle R_c=\frac{k_{information}}{n_{coded}}

Redundancy fraction relative to coded bits:

f_{red}=1-R_c

Payload throughput after coding and overhead:

R_p=R_{raw}R_c\eta_{oh}

where:

\displaystyle \eta_{oh}=\frac{1}{1+\alpha_{oh}}

Coding improves error performance only if the channel impairment matches the code assumptions and the decoder has enough soft information, interleaving, memory, and processing time.

OFDM Timing and Overhead

Useful OFDM symbol time:

\displaystyle T_u=\frac{1}{\Delta f}

Total OFDM symbol time:

T_s=T_u+T_{cp}

Cyclic-prefix efficiency:

\displaystyle \eta_{cp}=\frac{T_u}{T_u+T_{cp}}

Nominal OFDM occupied bandwidth:

B\approx N_{occupied}\Delta f

Sampling frequency for an FFT size $N_{FFT}$ :

f_s=N_{FFT}\Delta f

OFDM payload rate estimate:

\displaystyle R_p\approx \frac{N_{data}kR_c}{T_s}\eta_{pilot}\eta_{frame}

where $N_{data}$ is the number of data subcarriers and $\eta_{pilot}$ and $\eta_{frame}$ represent pilot and framing efficiency.

Cyclic-Prefix Margin

Cyclic-prefix time margin:

M_{cp}=T_{cp}-\tau_{sig}

where $\tau_{sig}$ is the maximum significant excess delay.

Uncertainty-aware margin:

M_{cp,conservative}=T_{cp}-(\tau_{sig}+u_\tau)

If $M_{cp}$ is negative, delayed channel energy can enter the useful symbol interval and create inter-symbol or inter-carrier interference. Increasing transmit power alone does not fix a guard-interval mismatch.

Adaptive Modulation and Coding

Usable SNR for mode selection:

SNR_{usable}=SNR_{measured}-M_{impl}-M_{fade}-M_{uncertainty}

Mode acceptance condition:

SNR_{usable}\ge SNR_{required,mode}

Approximate mode throughput:

R_{mode}=B\eta_{net,mode}

Fallback capacity ratio:

\displaystyle r_{fallback}=\frac{R_{fallback}}{R_{normal}}

Adaptive modulation should be checked with hysteresis, averaging time, traffic shaping, queueing behavior, and service priorities. A link can remain connected while the user service fails because capacity falls below offered load.

Interleaving and Latency

Interleaver delay from depth and coded bit rate:

\displaystyle t_{int}\approx \frac{D_{bits}}{R_{coded}}

Frame transmission time:

\displaystyle t_{frame}=\frac{N_{bits,frame}}{R_{line}}

Total digital PHY contribution:

t_{phy}\approx t_{frame}+t_{int}+t_{proc}+t_{retransmission}

Strong coding and interleaving can improve robustness while increasing latency and jitter. For control, voice, protection, and measurement links, the delay budget is often as important as the error-rate budget.

Worked Example 1: Payload Rate to Symbol Rate

A digital radio service must carry:

R_p=80\ \text{Mbit/s}

It uses 16-QAM, so:

M=16,\quad k=\log_2(16)=4

The code rate is:

R_c=3/4

Pilot, frame, and protocol overhead are:

\alpha_{oh}=0.15

Gross coded rate:

\displaystyle R_{coded}=\frac{80(1+0.15)}{0.75}=122.7\ \text{Mbit/s}

Required symbol rate:

\displaystyle R_s=\frac{122.7}{4}=30.7\ \text{Msymbol/s}

With raised-cosine roll-off:

\alpha=0.25

Occupied bandwidth estimate:

B\approx30.7(1+0.25)=38.4\ \text{MHz}

Engineering Comment

The payload rate is much lower than the coded line burden. The design must reserve about $38\ \text{MHz}$ before adding guard bands, regulatory mask margin, implementation filtering, and coexistence constraints.

Worked Example 2: Eb/N0 from Measured SNR

A receiver measures:

SNR=18\ \text{dB}

over:

B=20\ \text{MHz}

The coded bit rate is:

R_b=60\ \text{Mbit/s}

Convert to $E_b/N_0$ :

\displaystyle \left(E_b/N_0\right)_{dB}=18+10\log_{10}\left(\frac{20}{60}\right)

10\log_{10}(1/3)=-4.77\ \text{dB}

Therefore:

E_b/N_0=13.23\ \text{dB}

If the required value is $10\ \text{dB}$ and the implementation margin is $2\ \text{dB}$ :

M=13.23-10-2=1.23\ \text{dB}

Engineering Comment

The link passes, but not with much margin. If fading, interference, calibration error, or coding loss consumes more than about $1.2\ \text{dB}$ , this mode should not be released.

Worked Example 3: EVM to SNR

A 64-QAM signal has measured RMS EVM:

EVM_{rms}=5.0\%=0.050

Approximate SNR:

\displaystyle SNR\approx\frac{1}{0.050^2}=400

In decibels:

SNR_{dB}=10\log_{10}(400)=26.0\ \text{dB}

If the 64-QAM mode requires $24\ \text{dB}$ and implementation margin is $2\ \text{dB}$ :

M=26.0-24-2=0\ \text{dB}

Engineering Comment

The mode is only just acceptable by this screening rule. A zero-margin result should trigger a field check of phase noise, adjacent-channel interference, amplifier linearity, equalizer performance, and measurement uncertainty.

Worked Example 4: OFDM Cyclic-Prefix Margin

An OFDM waveform has subcarrier spacing:

\Delta f=15\ \text{kHz}

Useful symbol time:

\displaystyle T_u=\frac{1}{15{,}000}=66.7\ \mu\text{s}

The cyclic prefix is:

T_{cp}=4.7\ \mu\text{s}

Measured significant excess delay is:

\tau_{sig}=3.2\ \mu\text{s}

with uncertainty:

u_\tau=0.4\ \mu\text{s}

Nominal margin:

M_{cp}=4.7-3.2=1.5\ \mu\text{s}

Conservative margin:

M_{cp,conservative}=4.7-(3.2+0.4)=1.1\ \mu\text{s}

Cyclic-prefix efficiency:

\displaystyle \eta_{cp}=\frac{66.7}{66.7+4.7}=0.934

Engineering Comment

The delay spread fits inside the guard interval with about $1.1\ \mu\text{s}$ conservative margin. The link still needs EVM, packet loss, synchronization, and channel-estimation validation because delay spread is only one impairment.

Worked Example 5: Adaptive Modulation Throughput

A fixed wireless link has:

B=40\ \text{MHz}

Measured SNR is:

SNR_{measured}=21\ \text{dB}

The design reserves:

M_{impl}+M_{fade}+M_{uncertainty}=3\ \text{dB}

Usable SNR:

SNR_{usable}=21-3=18\ \text{dB}

Available modes are:

Mode	Net spectral efficiency	Required usable SNR
QPSK robust	$1.0\ \text{bit/s/Hz}$	$5\ \text{dB}$
16-QAM medium	$3.0\ \text{bit/s/Hz}$	$15\ \text{dB}$
64-QAM high	$5.0\ \text{bit/s/Hz}$	$24\ \text{dB}$

The highest allowed mode is 16-QAM medium. Mode throughput before protocol overhead is:

R_{mode}=40\times3.0=120\ \text{Mbit/s}

If additional MAC and service overhead is $12\%$ :

\displaystyle R_{service}=\frac{120}{1+0.12}=107\ \text{Mbit/s}

Engineering Comment

The link should not advertise the high mode even though measured SNR is above $20\ \text{dB}$ . After margin, it has only $18\ \text{dB}$ usable SNR, so 64-QAM would be fragile. Traffic policing should be based on the service rate, not the raw physical rate.

Worked Example 6: Packet Error from BER

A link has measured bit error rate:

BER=1.0\times10^{-6}

Each packet contains:

N_{bits,packet}=12{,}000\ \text{bits}

Approximate packet error rate:

PER=1-(1-10^{-6})^{12000}

Using the small-error approximation:

PER\approx10^{-6}(12{,}000)=0.012

Therefore:

PER\approx1.2\%

Engineering Comment

A BER that looks small can create a noticeable packet error rate for long packets. Forward error correction, interleaving, retransmission, packet size, and burst errors determine whether this becomes a service problem or stays hidden below the application layer.

Common Formula Mistakes

The most common mistake is mixing payload rate, coded rate, line rate, symbol rate, and service throughput. Every rate formula must state the boundary and the overhead included.

Another frequent error is comparing SNR, $E_b/N_0$ , and EVM values from different bandwidths or receiver reference planes. A few dB of hidden bandwidth or implementation mismatch can change the mode decision.

BER can also be misleading when packet size, burst errors, retransmission, interleaving, and coding are ignored. A low BER may still produce unacceptable packet loss, latency, or jitter.

OFDM calculations often fail when cyclic-prefix margin is checked without channel estimation, synchronization, Doppler, phase noise, pilot overhead, or inter-carrier interference.

Adaptive modulation can be unstable when mode thresholds do not include hysteresis, measurement uncertainty, fading margin, traffic load, and fallback service requirements.

Validation Evidence Package

For a modulation and coding calculation, verify:

payload, coded, line, symbol, and service rates are not mixed;
modulation order and coding rate match the selected mode;
overhead includes pilots, guards, framing, training, and protocol fields;
SNR, $E_b/N_0$ , and $E_s/N_0$ use the same bandwidth and rate basis;
EVM is interpreted with measurement conditions and impairment sources;
BER is converted to packet or service impact when packet size matters;
OFDM cyclic-prefix margin is checked against significant delay spread;
MCS selection uses usable SNR after implementation, fading, and uncertainty margin;
latency effects from interleaving, frame time, decoding, and retransmission are included;
validation evidence includes field measurements, not only simulated thresholds.
captured evidence includes receiver settings, bandwidth, reference plane, test duration, traffic load, channel condition, firmware or modem version, and calibration state.
the pass/fail decision states whether it applies to lab, field, nominal, fading, interference, mobility, overload, or service-acceptance conditions.

REF

Disciplines

Digital Modulation and Coding Formula Sheet

How to Use This Formula Sheet

Basis and Validity Limits

Symbols and Conventions

Modulation Order and Symbol Rate

Occupied Bandwidth and Spectral Efficiency

Shannon Capacity and Gap

Eb/N0, Es/N0, and SNR

Receiver Noise and Required Signal Level

EVM and SNR Screening

Bit Error and Packet Error

Coding Rate and Redundancy

OFDM Timing and Overhead

Cyclic-Prefix Margin

Adaptive Modulation and Coding

Interleaving and Latency

Worked Example 1: Payload Rate to Symbol Rate

Engineering Comment

Worked Example 2: Eb/N0 from Measured SNR

Engineering Comment

Worked Example 3: EVM to SNR

Engineering Comment

Worked Example 4: OFDM Cyclic-Prefix Margin

Engineering Comment

Worked Example 5: Adaptive Modulation Throughput

Engineering Comment

Worked Example 6: Packet Error from BER

Engineering Comment

Common Formula Mistakes

Validation Evidence Package

See also