board
Network cabinet fan controller
Public Unreviewedby matt
Two independently controlled 12 V fans, Ethernet and Wi-Fi, and an RGB speed display in a custom printed puck. Routed development preview, not approved for manufacture.
main
matt
Publish 0.1.3
34ed3ff
Review of advisory field-screen findings
MTA-156 revision fine-grid report: aiflow/fields-mta156/fields.json, 0.1 mm grid, full declared loads. Current and per-part regional-temperature screens pass. The following findings remain visible; there is no EMC, ESD, junction-temperature or hardware qualification claim.
- Electro-thermal coupling: the uncoupled model estimates about 19% more copper Joule loss at its solved temperatures. This applies to approximately 0.036 W of copper loss, not to the full 1.874 W package loss. A self-consistent sensitivity solve and closed-enclosure measurements remain release work.
- Inductor ripple: the percentage is high relative to the 0.6 A design load. The chosen inductance follows the converter topology and the separately retained periodic-ripple calculations; peak current, minimum inductance, saturation and input-capacitor ripple must be checked in bring-up. The model does not establish startup or compensation behavior.
- Ground-loop risk: each fan and the sensor have their own ground at their connector. Keep signal and return together. U2 and U3 are on-board devices, not external cable connectors; the generic detector overcounts those.
- Nine stitching-gap suggestions: the present four ground planes are connected and native DRC has no opens. Candidate edge positions from the rectangular field grid are not automatically valid near rounded corners. Additional perimeter stitching is an EMC design option, not a substitute for measured emissions.
- Two fragmented-pour findings: the report explicitly finds zero isolated pieces; all outer-plane pieces connect through GND vias and the internal planes. Native connectivity passes. No floating copper is being accepted.
- U1 decoupling finding: pin 1 is the 3.3 V feedback sense input, not its power-input supply pin. The generic heuristic incorrectly asks to move C10 (U3's bypass) to U1. Actual VIN bypass C3 and bootstrap C4 were checked on their correct pins; retain those intended circuits.
- U2 temperature warning: Approximately 78–79 °C is ambient plus calculated board-region rise. It is not a verified silicon-junction temperature. The module's 85 °C operating limit and small margin require physical enclosure testing and a real package thermal model.
- Twelve ESD and nine cable-filter findings: J5 is an unpopulated service interface used only for maintenance. Fan and remote-sensor wires still expose the board to handling and cable transients; cable length, routing, sensor placement and the intended ESD environment are not confirmed. Qualification may require added clamps or filters; no immunity rating is claimed.
- Conducted emissions: the tool flags approximately 30 dB over its class-B estimate using an assumed 8 ns switching edge and an uncalibrated cable/LISN model. Treat this as an open input-filter and bench-emissions item. Do not convert the tool's rounded 0.0 µH suggestion into a component selection, and do not claim compliance.
The design package is for review and prototype planning. The fans are confirmed as ARCTIC P14 Max; verify startup/stall behavior and the mated mechanical stack before releasing a revision for manufacture.
# Review of advisory field-screen findings
MTA-156 revision fine-grid report: `aiflow/fields-mta156/fields.json`, 0.1 mm grid, full declared loads. Current and per-part regional-temperature screens pass. The following findings remain visible; there is no EMC, ESD, junction-temperature or hardware qualification claim.
- **Electro-thermal coupling:** the uncoupled model estimates about 19% more copper Joule loss at its solved temperatures. This applies to approximately 0.036 W of copper loss, not to the full 1.874 W package loss. A self-consistent sensitivity solve and closed-enclosure measurements remain release work.
- **Inductor ripple:** the percentage is high relative to the 0.6 A design load. The chosen inductance follows the converter topology and the separately retained periodic-ripple calculations; peak current, minimum inductance, saturation and input-capacitor ripple must be checked in bring-up. The model does not establish startup or compensation behavior.
- **Ground-loop risk:** each fan and the sensor have their own ground at their connector. Keep signal and return together. U2 and U3 are on-board devices, not external cable connectors; the generic detector overcounts those.
- **Nine stitching-gap suggestions:** the present four ground planes are connected and native DRC has no opens. Candidate edge positions from the rectangular field grid are not automatically valid near rounded corners. Additional perimeter stitching is an EMC design option, not a substitute for measured emissions.
- **Two fragmented-pour findings:** the report explicitly finds zero isolated pieces; all outer-plane pieces connect through GND vias and the internal planes. Native connectivity passes. No floating copper is being accepted.
- **U1 decoupling finding:** pin 1 is the 3.3 V feedback sense input, not its power-input supply pin. The generic heuristic incorrectly asks to move C10 (U3's bypass) to U1. Actual VIN bypass C3 and bootstrap C4 were checked on their correct pins; retain those intended circuits.
- **U2 temperature warning:** Approximately 78–79 °C is ambient plus calculated board-region rise. It is not a verified silicon-junction temperature. The module's 85 °C operating limit and small margin require physical enclosure testing and a real package thermal model.
- **Twelve ESD and nine cable-filter findings:** J5 is an unpopulated service interface used only for maintenance. Fan and remote-sensor wires still expose the board to handling and cable transients; cable length, routing, sensor placement and the intended ESD environment are not confirmed. Qualification may require added clamps or filters; no immunity rating is claimed.
- **Conducted emissions:** the tool flags approximately 30 dB over its class-B estimate using an assumed 8 ns switching edge and an uncalibrated cable/LISN model. Treat this as an open input-filter and bench-emissions item. Do not convert the tool's rounded 0.0 µH suggestion into a component selection, and do not claim compliance.
The design package is for review and prototype planning. The fans are confirmed as ARCTIC P14 Max; verify startup/stall behavior and the mated mechanical stack before releasing a revision for manufacture.