Skip to content

SUPERFRAME_FLOWGRAPHS.md

Latest commit

 

History

History
120 lines (90 loc) · 3.47 KB

SUPERFRAME_FLOWGRAPHS.md

File metadata and controls

120 lines (90 loc) · 3.47 KB
Raw

Superframe Flowgraph Modifications

This document describes the flowgraph modifications needed for superframe transmission.

Overview

The superframe system uses separate flowgraphs for:

  1. Authentication Frame (Frame 0): 256 bits -> 768 bits coded (rate 1/3)
  2. Voice Frames (Frames 1-24): 384 bits -> 576 bits coded (rate 2/3)

Voice Frame Flowgraph: tx_4fsk_opus_voice.grc

Modifications from tx_4fsk_opus.grc

  1. Replace WAV source with file source:

    • Remove: blocks_wavfile_source_0
    • Remove: low_pass_filter_0
    • Remove: gr_opus_opus_encoder_0
    • Add: blocks_file_source_0 (reads 48-byte payload file, type: byte)

  2. Update LDPC matrix:

    • Change ldpc_matrix_file to: "../ldpc_matrices/ldpc_voice_576_384.alist"
    • Update ldpc_encoder_def file path accordingly
    • Note: Rate is 2/3 (384 bits info -> 576 bits coded)

  3. Payload size:

    • Input: 48 bytes (384 bits) payload file
    • After LDPC: 72 bytes (576 bits coded)
    • After 4FSK: 288 symbols (576 bits / 2 bits per symbol)

Flowgraph Structure

File Source (48 bytes) 
  -> LDPC Encoder (384->576 bits, rate 2/3)
  -> Packed to Unpacked (2 bits per chunk)
  -> Chunks to Symbols (4FSK mapping)
  -> Root Raised Cosine Filter
  -> Frequency Modulator
  -> File Sink (IQ output)

Authentication Frame Flowgraph: tx_auth_frame.grc

New Flowgraph

  1. File source:

    • Reads 32-byte payload file (256 bits signature)
    • Type: byte

  2. LDPC matrix:

    • Use: "../ldpc_matrices/ldpc_auth_768_256.alist"
    • Rate: 1/3 (256 bits info -> 768 bits coded)

  3. Payload size:

    • Input: 32 bytes (256 bits) payload file
    • After LDPC: 96 bytes (768 bits coded)
    • After 4FSK: 384 symbols (768 bits / 2 bits per symbol)

Flowgraph Structure

File Source (32 bytes)
  -> LDPC Encoder (256->768 bits, rate 1/3)
  -> Packed to Unpacked (2 bits per chunk)
  -> Chunks to Symbols (4FSK mapping)
  -> Root Raised Cosine Filter
  -> Frequency Modulator
  -> File Sink (IQ output)

Common Parameters

Both flowgraphs share:

  • RF Sample Rate: 48000 Hz
  • Symbol Rate: 4800 symbols/second
  • FSK Deviation: 2400 Hz
  • 4FSK Symbol Mapping: [-3, -1, 1, 3]
  • Root Raised Cosine: alpha=0.35, 110 taps

Usage from Python

The superframe controller writes payload files and executes flowgraphs:

# Voice frame
payload_file = "frame_1_payload.bin"  # 48 bytes
subprocess.run(["grcc", "-o", "output_dir", "tx_4fsk_opus_voice.grc"])
# Modify generated script to use payload_file
subprocess.run(["python3", "tx_4fsk_opus_voice.py"])

# Auth frame
payload_file = "frame_0_payload.bin"  # 32 bytes
subprocess.run(["grcc", "-o", "output_dir", "tx_auth_frame.grc"])
# Modify generated script to use payload_file
subprocess.run(["python3", "tx_auth_frame.py"])

Implementation Notes

  1. File Source Configuration:

    • Set repeat: False to read exactly one frame
    • Use type: byte for binary payload files

  2. LDPC Encoder:

    • Use fec_extended_encoder block
    • Configure with appropriate matrix file
    • No puncturing needed (use puncpat: "11")

  3. Timing:

    • Voice frame: 576 bits / 4800 bps = 120 ms (but we transmit in 40ms chunks)
    • Auth frame: 768 bits / 4800 bps = 160 ms
    • Note: Actual transmission time depends on symbol rate and filtering

  4. Output Concatenation:

    • The superframe controller concatenates frame outputs
    • Each frame produces a separate IQ file
    • Final superframe is assembled from 25 frame files