Course status and readiness matrix¶

This page is the compact top-level engineering status board for the course. It shows what is already strong, what is executable, and what still needs hardware validation.

Detailed bring-up logs should live on dedicated evidence pages rather than inside this matrix.

Readiness legend¶

Mark	Meaning
`Ready`	Stable learner-facing material.
`Executable`	Scripts, tests, plots or reproducible checks exist.
`Measured`	Real hardware or IQ capture evidence exists.
`Hardware pending`	Board-level validation or real capture data is still needed.
`Complete`	Documentation, reproducible artifacts and measured evidence with stated limits are present.

Quick decodings¶

Term	Meaning here
`stock-shell`	The board's normal vendor Linux and PL baseline after boot.
`runtime overlay`	A PL payload loaded after Linux boot for course experiments.
`BPSK`	Binary phase-shift keying.
`CSS`	Chirp spread spectrum; the current labs implement an educational waveform and detector, not yet a complete LoRa packet PHY.
`evidence`	A manifest, JSON report, plot or log proving the result.

Block readiness matrix¶

Block	Topic	Current state	Evidence level	Main gap / next improvement
01	Intro to SDR	Ready	Real RTL-SDR captures and controlled Zynq tone witness	Add a short learner report comparing passive capture and controlled tone.
02	Signals and sampling	Executable	Python labs and generated figures	Add MATLAB/C++ translations and metadata-mistake replay examples.
03	DSP basics	Executable	Python / MATLAB / C++ path	Add direct-vs-FFT convolution threshold demo and more reference outputs.
04	Simulink and fixed-point	Executable	Python/MATLAB references and BPSK `.slx` models	Constrain the BPSK Simulink path further for HDL Coder handoff.
05	FPGA / HDL flow	Measured signoff candidate	The canonical HDL smoke runs 47 testbenches (`python tools/run_block5_hdl_smoke.py`). The payload-telemetry image is fully routed at WNS +0.009 ns / WHS +0.028 ns, clean-boots on board B and produces a live preamble/payload split; the shipped differential-QPSK image (the current board-B modem) closes at WNS +0.036 ns.	Demonstrate repeat-build/seed timing robustness of the routed image.
06	RF frontend and AD9363	Measured	RX-only, OTA tone, protected conducted-tone baseline, first relative gain table, and a measured 30 dB attenuator flat from 50 MHz to 1 GHz	Measure cable loss separately, repeat the table across frequency and establish a connector-to-connector large-signal limit.
07	TX/RX chains	Executable	DUC/DDC demos and loopback models	Add measurement package.
08	Modulation and synchronization	Executable + RTL + fabric loopback on a board	CFO/timing, QPSK, 16-QAM, GFSK, DSSS, OFDM and packet CSS models. An OFDM modem in RTL (mapper, IFFT/FFT, cyclic prefix, pilot phase tracking, per-subcarrier zero-forcing equalizer, CFO corrector, 16-QAM option), bit-exact with its Q1.15 models, packaged as an AXI4-Stream/AXI4-Lite core and run on a board behind an AXI DMA in a fabric loopback: 0 errors in 3 072 000 bits (Lab 8.14, issue #48). Frame search, coarse CFO and level from a preamble, in simulation and in a built OFDM radio image. Real-hardware BPSK metrics (Lab 8.15); a bit-exact SF7 dechirp/sequential-detector RTL regression over all symbols	The OFDM radio image has not run on a board yet: attenuated cabled loopback first, then two boards (issue #48). The compact CSS detector is educational; SF5-SF12 LoRa PHY, interoperability, oversampled correlation, packet/ToA logic and board integration remain in zynq-lora-phy.
09	Recording and analysis tools	Measured	CI16/CU8/CF32/WAV readers, fail-closed manifest CI, a public 64 KiB synthetic QPSK Git LFS fixture with zero-BER replay, and measured QPSK multi-burst BER/EVM/CFO analysis	Complete publication review or external archival for the separate measured raw QPSK WAV.
10	KiCad and basic electronics	Draft	Calculators and templates	Add measured breadboard photos and KiCad exports.
11	Integrated SDR project	Complete (two-board link measured)	The in-fabric QPSK modem is closed on a two-board 915 MHz RF link. Bit 189 was traced to the RX DC blocker tracking the modulation (fixed with a running-average estimator, Lab 11.41); the residual ~1% whole-burst rotation floor to a frame-sync false lock (LOCK_ERR_TOL 3→1, Lab 11.42) and then to the four-branch resolution itself (removed by differential QPSK + a 24-symbol preamble, Lab 11.43–11.45). Result: gross whole-burst rotation 0, payload BER ~4×10⁻⁴, rotation-invariant; fabric loopback BER < 5.34×10⁻⁷ over 5.6 M bits. Lab 11.4 is the final measurement report. On top of the modem: a PS/PL message mailbox (Lab 11.49, 498 of 500 messages delivered one way), a two-board chat (Lab 11.50), acknowledgements (Lab 11.51: 100 and 99 of 100 lines with them, 60 and 61 of 100 without) and long messages in fragments (Lab 11.52: 10 of 10 messages of 384 bytes in 16 fragments).	Chase the residual differential single-bit rate only if an application needs it: it is the inherent ~3 dB differential penalty, not a rotation failure.
12	Final projects	Ready framework + reference implementation	Final-project framework: briefs 12.1–12.4 (QPSK modem, RF capture, FPGA DSP block, full SDR report), per-tool READMEs, report templates and grading rubric. Project 12.1 (QPSK modem) is anchored by the completed Block 11 two-board modem as a worked reference exemplar (Lab 11.4).	Learners pick a track and produce their own proposal, verification set and report.

Language coverage¶

Every lab page in all twelve blocks has a full Russian version, including the detailed hardware-experiment chronologies of Labs 11.8 and 11.26. The Block 11 lab map tells you which labs to do and which to read as case studies.

Hardware validation priorities¶

Priority	Task	Done when
P0	Safe cabled loopback	Attenuation, gain settings, capture metadata and short conclusion are recorded.
P0	Runtime PL BPSK/QPSK robustness	Success count / total attempts is repeatable from clean boot and reported with limitations.
P1	QPSK demo dataset	✔ Public Git LFS payload, manifest, checksum, deterministic rebuild, constellation, BER/SER, EVM/SNR and replay command exist.
P1	Digital-link metric gate	Digital labs report SNR/EVM plus BER or FER with compared bit/frame count.
P1	AD9363 gain table	Gain settings, clipping/SNR behavior and safe starting values are documented.
P2	OFDM/CSS hardware evidence	Digital loopback, safe RF path, configuration metadata, EVM plus BER/SER/PER and limitations are recorded.
P2	Final hardware report	One report connects model, HDL, capture, metrics and engineering conclusion.

Evidence and backlog pages¶

Current release focus¶

The v0.1.0 release candidate is content-complete: it has a clean learner route, green CI, compact status pages and one flagship model-to-measurement hardware story. The remaining release action is to verify the final commit, create the annotated tag and publish the GitHub Release. The synthetic QPSK replay package is publication-ready; publication review of the separate measured OTA WAV remains open.