rawsocket.cloud - bench notes on serial buseshttps://rawsocket.cloud/2026-10-10T09:00:00ZIvo Kallasivo@rawsocket.cloudUART, SPI, I2C, CAN and RS-485, measured on one bench in Tartu.Two nodes at 1 Mbit/s that sampled the bit in different placeshttps://rawsocket.cloud/can-sample-point.html2026-10-10T09:00:00Z2026-10-10T09:00:00ZTwo controllers on one 1 Mbit/s segment sampling 125 ns apart, the time-quantum arithmetic behind it, and why 87.5 per cent is a default worth matching.A 95-metre CAN trunk that went quiet when I pulled one terminatorhttps://rawsocket.cloud/can-no-terminator.html2026-09-28T09:00:00Z2026-09-28T09:00:00ZSeventeen nodes, 500 kbit/s and a missing 120 ohm resistor that failed long frames while short ones kept working.A sensor that held the clock low for 1.4 mshttps://rawsocket.cloud/i2c-clock-stretch.html2026-08-05T09:00:00Z2026-08-05T09:00:00ZA device stretching SCL after the address byte, and a master whose 25 microsecond timeout aborted the read halfway.Two boards agreed at 9600 and argued at 115200https://rawsocket.cloud/uart-baud-error.html2026-06-11T09:00:00Z2026-06-11T09:00:00ZA 12 MHz crystal against an 11.0592 MHz one, a seven per cent divider error, and why the failures depended on the byte value.310 pF of bus capacitance and the rise time I ignoredhttps://rawsocket.cloud/i2c-pullups.html2026-03-02T09:00:00Z2026-03-02T09:00:00ZEleven devices and 1.8 kilohm pull-ups that passed at 100 kHz and corrupted the last acknowledgement at 400 kHz.Four hundred metres of RS-485 and two grounds that disagreedhttps://rawsocket.cloud/rs485-isolation.html2025-11-19T09:00:00Z2025-11-19T09:00:00ZA link that logged errors only in the morning, and the ground-potential measurement that led to an isolated transceiver.A SPI daisy chain where the last word came back shorthttps://rawsocket.cloud/spi-daisy-chain.html2025-09-16T09:00:00Z2025-09-16T09:00:00ZThree ADCs in series and a chip-select edge released 90 ns too early for the farthest device's top bits.Every reading came back at exactly half: a CPHA mismatchhttps://rawsocket.cloud/spi-mode-mismatch.html2025-07-08T09:00:00Z2025-07-08T09:00:00ZA mode mismatch at 8 MHz that halved every value, and why slowing the clock made the wrong answer look plausible.Three samples per bit are not evidencehttps://rawsocket.cloud/analyzer-aliasing.html2024-12-14T09:00:00Z2024-12-14T09:00:00ZThe first entry, and a confession: 24 MS/s against an 8 MHz clock, and four hours spent chasing a sampling artefact.