PDLVDS-RX — photodiode LVDS receiver

Clean-room receiver for the Miniscope Zero optical uplink: Excelitas C30737LH Si APD → OPA657 TIA → LMH7220 comparator (LVDS out, DAC-set threshold) → ADN2813 CDR → STM32H723 + USB3300 USB-HS. Design target: 50 Mbps line rate, 8b/10b (40 Mbps payload); the paper's mode: 8 Mbps Manchester (16 Mbaud), as transmitted. LT3757 flyback APD bias 150–420 V. USB-C powered; verify on any laptop over USB, no FPGA attached.

Pass 146

2026-11-06 01:19 PST (2026-11-06T09:19:00Z UTC) · commit a3a5b7a (a3a5b7a31d32444611a9b015fa9f9606324d6546)

ADOPTED

Copper-edge campaign, part 2: the creepage slot is now CLEAR of copper, and C26 was living in it. Diagnosis chain: DRC slot items read ‘actual 0.0000 mm’ – tracks CROSS the slot (docs/02 HV-island creepage slot, x12..46 / y23.6..25.1), they do not hug it. The heal router ignores Edge.Cuts and re-lays crossings (proven on +5V: delete the 19 mm In2 crossing, heal, crossing census unchanged). New mechanism, proven on a +5V probe first: a sacrificial SLOTFENCE net fenced across the band on all 4 layers (scratch/fence_slot.py) forces heal detours around the slot ends; fence removed before the referee runs. Production: deleted 23 crossing segments (HV_EN 11, DAC_VREF 6, HV_DAC 3, +5V 1, CMP_QN 1, GND 1; scratch/delete_slot_crossings.py), fenced, healed all 5 nets (4 connected immediately; +5V U7.4 and HV_EN U6.97 remain the standing isolated pads). DAC_VREF’s residue exposed the real defect: C26 was PLACED INSIDE THE SLOT (centre (43.5,24.0), pads over the cutout) with its fanout via also in the slot (43.125,24.125, deleted). C26 is U8’s VREF bypass (DAC_VREF has exactly two pads: U8.10 and C26.1). First move target (46.8,30.8) landed in U8’s +3V3D fanout (4 shorts, heal NO_EXIT – rejected by DRC); second move to (49.5,23.5), east of the slot end and north of U8, courtyard-clean, healed in 1 s. A VM replacement ate the first heal run mid-flight; the stage was rebuilt from exp175 after a census showed the dead run’s checkpoint had contaminated the work file. KiCad DRC referee, exp175 -> exp176: violations 249 -> 229, unconnected items 45 -> 44 (DAC_VREF whole again), copper_edge_clearance 68 -> 40 (slot band 35 -> 0 by the crossing census), shorting 0, clearance 34 -> 35 (one new item: CMP_QN detour vs THR_DAC via south of the slot – named for the clearance campaign), dangling 23 -> 30 (fresh heal stubs, next prune). Edge residue: J1 NPTH field 18 (incl. footprint-intrinsic pad items), east-edge trunks 15.

Nets fully routednot recorded
Pins failednot recorded
Clearance violations68
Unconnected nets31
Audit verdictFAIL

Routes — all copper layers and each layer

F.Cu In1.Cu In2.Cu B.Cu

Top

Pass 146 — top

Bottom

Pass 146 — bottom

← Pass 145 · all passes · Pass 147 →