Skip to content
RFrftools.io
Signal

BER Calculator — Bit Error Rate from SNR

Free BER calculator for BPSK, QPSK, 8PSK, 16-QAM. Enter Eb/N0 to instantly compute bit error rate. Compare modulation schemes and optimize link performance.

Loading calculator...

Formula

BER=12erfc(Eb/N0)BER = \frac{1}{2} \text{erfc}\left(\sqrt{E_b/N_0}\right)
BER— Bit error rate
Eb/N0— Energy per bit to noise density (dB)
erfc— Complementary error function

How It Works

The BER-SNR Calculator computes bit error rate from Eb/N0 for digital modulation schemes — essential for communication link budget analysis, modem design, and wireless system planning. RF engineers, telecom designers, and satellite communication specialists use this to predict link reliability and select appropriate modulation. Per Proakis "Digital Communications" (5th ed., Ch. 5), BER depends on modulation type and Eb/N0 (energy per bit to noise spectral density). BPSK/QPSK achieve BER = 0.5*erfc(sqrt(Eb/N0)) — at 10 dB Eb/N0, BER = 3.9e-6 (approximately 1 error per 256,000 bits). 16-QAM requires 4 dB higher Eb/N0 for same BER; 64-QAM needs 8 dB more. Per 3GPP TS 36.101, LTE targets BER < 1e-3 before FEC, achieving < 1e-6 after decoding. Modern 5G NR uses 256-QAM requiring 24 dB Eb/N0 for uncoded BER = 1e-5.

Modulation comparison matrix (uncoded, AWGN channel)

The table below shows the required Eb/N0 for common BER targets across the modulation schemes this calculator supports. Values follow Proakis Ch. 5 formulas with high-precision erfc evaluation.

ModulationBits/symbolEb/N0 for BER=1e-3Eb/N0 for BER=1e-6Eb/N0 for BER=1e-9
BPSK16.8 dB10.5 dB12.6 dB
QPSK26.8 dB10.5 dB12.6 dB
8-PSK310.0 dB14.0 dB16.2 dB
16-QAM410.5 dB14.5 dB16.6 dB
64-QAM614.8 dB18.5 dB20.6 dB
256-QAM819.5 dB23.0 dB25.2 dB
Two patterns: doubling the constellation size (4-QAM → 16-QAM → 64-QAM → 256-QAM) costs roughly 4-5 dB per step for the same BER. Each 3 dB of Eb/N0 pushes the BER down by about two orders of magnitude in the steep "waterfall" region.

Why this matters for link budget work

The modem's required Eb/N0 sets the receiver sensitivity, which sets the maximum path loss a link can tolerate. If a 100 Mbps 64-QAM link requires 18.5 dB Eb/N0 for 1e-6 BER, a 20 MHz thermal-noise floor at 5 dB NF gives sensitivity ≈ -96 dBm + 18.5 dB = -77.5 dBm. Drop to QPSK and the number becomes -85.5 dBm — 8 dB more path-loss headroom, or roughly 2.5× range, at the cost of 3× lower throughput. Adaptive modulation and coding (DVB-S2X, 5G NR) navigate this trade dynamically.

Worked Example

Worked example 1 — QPSK uplink sizing for a LEO satellite

Size uplink power for LEO satellite with QPSK modem requiring BER < 1e-6.

  1. From QPSK BER formula: BER = 0.5 × erfc(sqrt(Eb/N0)). Solve 1e-6 = 0.5 × erfc(sqrt(x)) → x = 10.5 dB.
  2. Add 2 dB implementation loss per Proakis Table 5.3.
  3. Required Eb/N0 = 12.5 dB.
  4. For 1 Mbps data rate: required C/N0 = 12.5 + 10*log10(1e6) = 72.5 dB-Hz.
  5. With -174 dBm/Hz thermal floor + 5 dB NF + 15 dB sky temperature ≈ -154 dBm/Hz: required signal = -154 + 72.5 = -81.5 dBm.

Per ITU-R S.1062, this matches typical LEO uplink sensitivity specifications.

Worked example 2 — DVB-S2 Ku-band downlink, QPSK 3/4 LDPC

Problem: a DVB-S2 broadcast carrier at 27.5 Msym/s occupies 30 MHz of Ku-band spectrum. Target quasi-error-free reception (QEF, BER < 2e-10 after FEC).

  1. DVB-S2 uses QPSK + LDPC rate 3/4. The post-FEC QEF threshold is 4.0 dB Es/N0 per DVB-S2 spec (ETSI EN 302 307).
  2. Convert Es/N0 → Eb/N0: Eb/N0 = Es/N0 - 10*log10(bits/symbol × code rate) = 4.0 - 10*log10(2 × 0.75) = 4.0 - 1.76 = 2.24 dB.
  3. Uncoded QPSK at 2.24 dB Eb/N0 has BER ≈ 3e-2 — a 3% raw bit-error rate. LDPC pulls this to < 2e-10 (8 orders of magnitude of coding gain).
  4. Thermal noise floor in 30 MHz at 290K + 1 dB LNB NF: N = kTB = -174 + 75 + 1 = -98 dBm.
  5. Required received C: -98 + Es/N0 = -98 + 4.0 = -94 dBm.

Key lesson: strong FEC codes like LDPC let you operate 8+ dB below what uncoded QPSK would need. This is why modern satellite broadcast survives at receive signal levels only a few dB above the noise floor.

Worked example 3 — LoRa SF12 uplink, 125 kHz bandwidth

Problem: an outdoor LoRa sensor at SF12 / 125 kHz needs to cover 15 km rural range with 99% reliability.

  1. LoRa is chirp spread-spectrum — not a classical PSK/QAM modulation, but the calculator's BPSK BER curve is a reasonable approximation for the coherent-detection inner receiver below threshold.
  2. Semtech SX1276 datasheet: SF12 / 125 kHz sensitivity = -137 dBm, corresponding to Es/N0 ≈ -20 dB (negative — signal is below noise). Processing gain from 4096-chirp SF12 = 10*log10(4096) ≈ 36 dB.
  3. For 1e-3 raw BER before FEC: effective Eb/N0 after despreading = -20 + 36 = 16 dB — which looks like a QPSK curve at BER ≈ 4e-8.
  4. LoRa's coding rate 4/5 + interleaving further drops this to packet error rate ≈ 1% at the sensitivity limit.
  5. Link budget: 20 dBm Tx + 2 dBi Tx/Rx antennas - 2 dB cable - FSPL_915MHz(15km) = 20 + 4 - 2 - 115.2 = -93.2 dBm at receiver. Margin over sensitivity = -93.2 - (-137) = 43.8 dB.

The calculator can't natively handle the spreading-gain math, but it gives the right despread Eb/N0 → BER mapping. For the spreading part, use the LoRa-specific relation: effective Eb/N0 = C/N0 - 10*log10(chip_rate/bit_rate).

Key lesson: when a modem operates below the noise floor (like LoRa or GPS), the calculator is useful for the inner BER curve after despreading, not the outer receive signal.

Practical Tips

  • ✓Per 3GPP standards, budget 2-3 dB implementation margin above theoretical Eb/N0 for real hardware
  • ✓Use Gray coding for QAM constellations to minimize adjacent symbol errors — reduces BER by factor of log2(M) per Proakis
  • ✓Forward error correction (FEC) provides 5-10 dB coding gain: rate-1/2 turbo code achieves BER=1e-6 at 2 dB Eb/N0
  • ✓For fading channels, use diversity techniques — 2x diversity provides 10 dB gain at BER=1e-3 per Rappaport
  • ✓When feeding this BER into a link budget, subtract the modem's implementation margin (typically 1-3 dB) from Eb/N0 before looking up the BER — real hardware never hits theoretical performance
  • ✓For ADC-limited systems, check the quantization-noise floor as well — an 8-bit ADC has SQNR ≈ 50 dB, which caps effective Eb/N0 even if the RF SNR is higher

Common Mistakes

  • ✗Confusing Eb/N0 (dB) with linear ratio — must convert: 10 dB = 10 linear, not 10 for erfc calculation
  • ✗Using BPSK formula for higher-order modulations — 16-QAM BER is approximately 4x higher at same Eb/N0 per Proakis
  • ✗Neglecting erfc function precision — polynomial approximations introduce 1-5% error; use IEEE 754 compliant implementations
  • ✗Comparing uncoded BER against post-FEC QEF thresholds — a modem datasheet citing "BER = 1e-10" almost always means after FEC; the uncoded BER on the channel may be 1e-2 or worse
  • ✗Mixing Es/N0 and Eb/N0 — Es/N0 measures energy per modulation symbol; Eb/N0 normalizes to information bits. For QPSK with no coding: Eb/N0 = Es/N0 - 3 dB; for rate-3/4 LDPC-coded QPSK: Eb/N0 = Es/N0 - 1.76 dB

Frequently Asked Questions

BER depends on modulation type and Eb/N0 (energy per bit to noise density). For BPSK and QPSK: BER = 0.5 × erfc(√(Eb/N0)). At 10 dB Eb/N0, BER = 3.9×10⁻⁶. To convert SNR to Eb/N0: Eb/N0 = SNR + 10×log10(bandwidth/bitrate). This calculator handles the erfc computation with high precision.
BPSK BER = ½ × erfc(√(Eb/N0)), where erfc is the complementary error function. At Eb/N0 = 0 dB: BER = 7.9×10⁻². At 5 dB: BER = 6×10⁻³. At 10 dB: BER = 3.9×10⁻⁶. QPSK has identical BER performance but doubles spectral efficiency by transmitting 2 bits per symbol.
For BER = 10⁻⁶ (one error per million bits): BPSK/QPSK requires 10.5 dB Eb/N0. 8PSK needs ~14 dB. 16-QAM needs ~14.5 dB. 64-QAM needs ~18.5 dB. Add 2-3 dB for real-world implementation margin. With rate-½ turbo coding, you can achieve BER = 10⁻⁶ at only 2 dB Eb/N0.
Eb/N0 (energy per bit to noise spectral density) is the fundamental SNR metric for digital communications per Proakis. Eb/N0 = C/N0 - 10×log10(Rb) where Rb = bit rate. It normalizes SNR to bit rate, enabling fair comparison across systems with different data rates. 10 dB Eb/N0 means each bit has 10× energy compared to noise in 1 Hz bandwidth.
BER decreases exponentially as SNR increases. Each 1 dB improvement in Eb/N0 roughly halves the BER in the waterfall region. At low SNR (< 5 dB), BER is high (> 1%). At moderate SNR (8-12 dB), BER drops rapidly. At high SNR (> 15 dB), BER approaches zero. The exact relationship depends on modulation — BPSK/QPSK is most power-efficient, while 64-QAM needs 8 dB more for the same BER.
16-QAM packs 16 symbols into the constellation versus QPSK's 4, making symbols closer together. The minimum distance between adjacent symbols is smaller, so noise is more likely to cause detection errors. 16-QAM needs ~4 dB higher Eb/N0 for the same BER. The tradeoff: 16-QAM transmits 4 bits per symbol (vs 2 for QPSK), doubling spectral efficiency — useful when bandwidth is limited but power is available.
Receiver sensitivity P_sens = thermal noise floor + NF + required Eb/N0 + 10*log10(bit_rate) + implementation margin. Example: 100 Mbps 64-QAM, 5 dB NF, 1e-6 BER target, 2 dB impl. margin. Thermal floor in 1 Hz = -174 dBm. Sensitivity = -174 + 5 + 18.5 + 80 + 2 = -68.5 dBm. Feed that into the RF Link Budget calculator to compute link margin at your operating range.
No — this calculator computes uncoded (raw, pre-FEC) BER as a function of Eb/N0 and modulation. FEC coding gain (3-10 dB depending on code family) is applied separately. To model a coded link: (1) subtract the coding gain from the required Eb/N0 (rate-1/2 convolutional ≈ 5 dB, turbo ≈ 7 dB, LDPC ≈ 10 dB), or (2) adjust Eb/N0 to Es/N0 scale and use the modem datasheet's post-FEC threshold directly. See Worked Example 2 (DVB-S2) for how to reconcile the two in a real design.
Yes — every feature on this page runs in your browser, requires no signup, and supports URL-shareable scenarios for handoff to colleagues. Pro tier adds cloud-saved scenarios, CSV export, and API access; core math is always free.
BPSK, QPSK, 8-PSK, 16-QAM, and 64-QAM with exact Proakis/Sklar BER formulas. For 256-QAM, 1024-QAM, APSK (used in DVB-S2), or GFSK/MSK, use the comparison table in the theory section above for approximate Eb/N0 vs BER values. We're adding 256-QAM + APSK to the dropdown; if you need them sooner, open a request via /request.

Related Articles

Signal Processing

AM Modulation Index: Calculation and Importance

Learn how to calculate AM modulation index, sideband frequencies, bandwidth, and power efficiency with real worked examples for RF engineers.

Signal

BER vs SNR: Understanding Digital Communication Performance

Understand the relationship between Bit Error Rate (BER) and Signal-to-Noise Ratio (SNR). Compare BPSK, QPSK, and QAM modulation performance with Eb/N0 curves and worked examples.

Shop Components

As an Amazon Associate we earn from qualifying purchases.

Op-Amp IC Assortment

General-purpose and precision operational amplifiers

Amazon →

Function Generator

DDS function generator for signal and filter testing

Amazon →

Related Calculators

Signal

SNR

Calculate signal-to-noise ratio, noise floor, receiver sensitivity, and dynamic range for RF systems. Analyze your signal chain performance. Free, instant results.

Signal

Quantization Noise

Calculate ADC quantization noise, SQNR, ENOB, and noise spectral density. Analyze dynamic range for analog-to-digital converter design. Free, instant results.

Signal

ADC SNR & ENOB

Calculate ADC signal-to-noise ratio, ENOB, and SFDR with aperture jitter effects. Analyze converter performance for data acquisition design. Free, instant results.

RF

Link Budget

Free RF link budget calculator: enter Tx power, antenna gains, frequency, and distance to get received signal level, link margin, and max range. Covers satellite, terrestrial, and IoT links.