Battery Converter Design
Table
| # | Functional cluster | Directly attributable | Notes |
|---|---|---|---|
| 1 | Potentiostat analog drive/current sense/reference/AGND | 3 | U_DAC/AGND and AGND–DGND link contacts |
| 2 | Fault/current-limit/load-detect/shutdown | 12 | All are around U_FAULT_LATCH |
| 3 | USB-C and power regulation | 9 | J_USB shell plus U_LDO_D/U_LDO_A |
| 4 | Arduino/OLED/encoder/digital control | 8 | ADC alert, FTDI, Arduino/OLED grounds, encoder mount |
| 5 | SMB connectors/INT-EXT selector/output connectors | 1 | Directly identified J_SMB_RE shell-pad contact |
| 6 | Same-net trace-pad/via artifacts or unclassified UID-only findings | 243 | Route/via/pad findings lacking enough net metadata |
| Total | 276 | Exact current DRC total |
route UID → net mapping.U_DAC:~LDAC~02b33325…__b7589d9c… ↔ Trace 05f949ac…R_AGND_DGND:P17a2ef6c8…__dff57d78… ↔ Trace 6d968da8…R_AGND_DGND:P17a2ef6c8…__dff57d78… ↔ Trace f3f05564…~LDAC~ contact is known to be an AGND same-net pad-entry artifact. The two star-link resistor endpoint contacts also look like intentional route-to-pad entries under the global keepout rule. No current different-net analog short is proven from these three pairs alone.238d (six pairs), 271b (seven), 017c (ten), and 7f6d (four) should not be dismissed merely because many other findings are artifacts; their nets are absent from the DRC payload.U_FAULT_LATCH:~Q2 ↔ Trace 971c1df2… — 1~Q1 ↔ Traces 390c3a33…, a62859dc… — 2RESET1 ↔ Traces 390c3a33…, 3f29e212…, 4f4ccedc…, a62859dc… — 4SET1 ↔ Traces 3f29e212…, 4f4ccedc…, c7b8f58b…, e2abafba…, ecb87307… — 5~Q1/~Q2 contacts are the highest-risk representatives. Those latch outputs have previously appeared as logically unused/no-connect outputs; copper touching them should be treated as a credible real risk until net identity is verified.RESET1/SET1 pairs form repeated pad/branch networks and are much more characteristic of same-net trace–pad/trace–trace keepout artifacts.U_LDO_D:NC31161b42…__30f3e345… ↔ Trace 3e0f0717…U_LDO_D:EN31161b42…__9a6ce0c6… ↔ Trace 091187a2…U_LDO_A:OUT9c206145…__5ad4e767… ↔ Trace 122a5ed1…U_LDO_A:GND9c206145…__8c60cffa… ↔ Trace 1e9a82ee…J_USB:S1 shell/mount pad8bbcaffd…mounts.21c2dc92… ↔:
00c9b93a…534bbe72…6cbf7b65…cb954fac…df47c909…U_LDO_D:NC ↔ Trace 3e0f. A routed copper object touching an NC pad is not an ordinary intended junction.J_USB:S1 contacts are credible shell/mount-pad risks, because S1 is a footprint shell pad rather than one of the explicitly named GND_A/GND_B signal pads. They may be intentional shell grounding, but the current logical data does not prove that.U_ADC:ALERT/RDY ↔ Trace 1fc1dbad… — 1J_FTDI:Pin_6 ↔ Trace a7e275a8… — 1J_FTDI:Pin_1 ↔ Trace 343116ad… — 1OLED1:GND ↔ Trace 4ea4338f… — 1U_ARD:GND pads ↔ Trace 90e09d93…, plus one ↔ Trace a99c449f… — 3ENC1:P$6 mechanical mount ↔ Trace ee2aa7d7… — 1U_ADC:ALERT/RDY ↔ Trace 1fc1 is a confirmed same-net ADC_ALERT pad-entry artifact.ee2a is the highest-risk digital representative because the mount is not an ordinary signal pad.J_SMB_RE:2 shell pad0d88c73c…__78f5c63d… ↔ Trace a3b89a5d…a3b89a5d is on SMB_SHIELD_STAR; otherwise it is a real center/signal-to-shell-area clearance problem.P1/P2 reports almost certainly belong to SMB, selector, or output connectors, but their owning component is not encoded in the DRC check name and they therefore remain in cluster 6.017c — 10 reported neighbors271b — 7238d — 61554 — 533b1 — 583fc — 57f6d — 44f4c — 46da8 — 4U_LDO_D:NC ↔ Trace 3e0fU_FAULT_LATCH:~Q2 ↔ Trace 971cU_FAULT_LATCH:~Q1 ↔ Traces 390c, a628J_USB:S1 ↔ five tracesJ_SMB_RE:2 ↔ Trace a3b8ENC1:P$6 mount ↔ Trace ee2a017c, 271b, 238d, and 83fcU_DAC:~LDAC~ ↔ Trace 05f9 — same-net AGNDU_ADC:ALERT/RDY ↔ Trace 1fc1 — same-net ADC_ALERTRESET1 and SET1 route contactsWE_DRV, WE_INT, I_SENSE_POS/NEG, I_SENSE, and references.U_LDO_D:NC and unused U_FAULT_LATCH:~Q1/~Q2.017c, 271b, 238d, and 83fc.GND net and existing full-board GND fill, I cannot guarantee the Rev 2 requirement of separate analog/digital ground pours joined at exactly one point near the analog LDO/reference, nor can I guarantee no extra SMB_SHIELD_STAR/GND stitching if bulk autoroute is allowed to proceed.U_REF manufacturer part number updated to LM4040A20IDBZR, Texas Instruments LM4040A 2.048 V A-grade 0.1% SOT-23/DBZ reference.R_SENSE is already specified as 10.0 Ω, 0.1%, 0603, 0.1 W.LM4040A20IDBZR because the library exposed only D-grade LM4040 records.WE_DRV shunt-side load path and Kelvin branchWE_INT shunt-side Kelvin branch and feedback/ADC monitor branchesI_SENSE_POSI_SENSE_NEGI_SENSE feedback/output segmentVREF_2V048 LM4040-to-DAC/reference-bypass pathDAC_VSETOPA_B_FOLLOWERVREF_MID_1V024AGND, DGND, and SMB_SHIELD_STAR domains, connect AGND/DGND at one net-tie or 0 Ω link near the analog LDO/reference, then route/fill each domain explicitly. This makes the single-point join enforceable by DRC.GND as-is, manually delete/replace the current full-board GND fill with shaped copper regions and a documented bridge. This is faster but easier to accidentally violate because DRC still sees one logical GND net.U_LDO_A/U_REF. This is the only robust way to satisfy the Rev 2 grounding requirement during routing and manufacturing review.Why routing stopped
Work completed before stopping
What is not complete
Required next decision
Current recommendation