Boards and fields

The copper, solved for its near field.

A schematic and its board are two files in code. The near-field and cross-section solvers read the board as drawn and show where its switching current couples into its sensor lines.

Simulated: the magnetic near field of the robot hand board's 6 V servo regulator, drawn as a glowing surface over the copper, then a cut through the board where the regulator's switching trace crosses over a touch-sensor line, with its electric field lines.
Simulated: the near field of the Hand board's 6 V servo regulator, then a cut through the board. Rendered from the solver's numbers.

Case study · UpperBody-001's Hand board

84 A/m over the coil, 0.7 kA/m on the touch lines.

The hand's 6 V servo regulator switches about 1.07 A of ripple through its output loop. The near-field solver sums that current's field over the board's copper. A probe plane 3 mm over the top face reads its peak above the coil, L1. The touch-sensor lines run under the loop on an inner layer, and the field there is more than eight times stronger.

at most on the probe plane, 3 mm over L1
84 A/m
at most along ELE10, on the inner layer under the loop
724 A/m
peak to peak of ripple, from the board's L1 and rails at an assumed 500 kHz
1.07 A

Simulated

The strongest magnetic field along each sensor line of the Hand board
The numbers
The strongest magnetic field along each sensor line of the Hand board
NetStrongest field along the line
ELE10724 A/m
ELE11718 A/m
ELE6700 A/m
ELE9699 A/m
FSRN_INDEX248 A/m
ELE5243 A/m
FSR_THUMB220 A/m
ELE8168 A/m
ProvenanceFile hardware/modules/UpperBody-001/ee/Hand board.pcb.tsxOutput case-studies/boards-and-fields/media/hand-board.emi.jsonTool scripts/sites/emi-study.mts (board-emi.ts: near-field.ts, cross-section.ts)Ran on This computer (dell-gb10-1)When

The cut through the board

More than half of ELE9's capacitance is to the switch node.

Where the switch node crosses the touch lines, the cross-section solver cuts the board and solves the electric field between 14 conductors. The touch electrode ELE9 has 97 pF/m in all, and 55 pF/m of it is to the switch node, so each switching edge reaches its reading.

The board's file names no layer stack, so the study assumed a 4-layer 1.6 mm one (JLC04161H-7628). Every assumption is in the study's file.

Simulated

of ELE9 is to the switch node
55 of 97 pF/m
conductors in the cut, solved in 1,357 iterations
14
schematics and 9 boards in code across our products (case-studies/boards-and-fields/stats.json)
20
ProvenanceFile hardware/modules/UpperBody-001/ee/Hand board.pcb.tsxOutput case-studies/boards-and-fields/media/hand-board.emi.jsonTool scripts/sites/emi-study.mts (cross-section.ts)Ran on This computer (dell-gb10-1)When

Find the coupling before the board is made.

Write the board in code, run the near field and the cut on it, move a trace, and run them again.