B-1: build a Halbach pair at the 12 mm design gap and traverse a Hall probe. The cheapest route to the first measured number in a project whose standing caveat is E4 — nothing built, fired or measured at any scale.
Status 2026-08-10: it is now an order, not a procedure. docs/B1_ORDER.md.
B-1 has had a method since 2026-07-29 and a bill of materials since 2026-07-30 and has never been ordered. ADR-021 froze the defect register partly on that observation: 17 model-against-model analyses were produced around a ~₹22,000 experiment that was never bought.
What the order adds beyond the method
The magnetisation sketch and quotation wording. The near-certain failure mode is a supplier magnetising through the largest face — the default tooling orientation — which yields eight identical blocks and no Halbach array. It is invisible on arrival because the blocks are correctly magnetised, just on the wrong axis. The order carries an ASCII sketch of the four-block wavelength and paste-ready wording requiring written confirmation of the axis per group before production.
A receiving inspection. Check every block individually with a compass and sort into two piles of four before anything closes across a 12 mm gap — after that it does not come apart by hand, and the mistake has already cost the fixture.
The hazard, quantified. The arrays attract at ~828 N across the gap on brittle sintered NdFeB. Captive spacer, shim outward from a closed stack, never approach from open.
The bands, unchanged
Declared 2026-07-30, derived 2026-08-06, not editable after a reading:
| Row |
Model |
Rig contributes |
Accept if |
| Peak gap field |
0.694 T |
4.4 % |
±15 % → 0.590–0.798 T |
| Winding mean |B| |
0.552 T |
4.5 % |
±15 % → 0.469–0.635 T |
| Stray at 10 mm |
22.7 mT |
3.7 % |
±20 % |
| Stray at 20 mm |
4.33 mT |
4.2 % |
±40 % |
| Stray at 50 mm |
0.38 mT |
3.3 % |
order of magnitude |
Biggest cost lever: the teslameter is two thirds of the total and 1 T range is binding — the model predicts 0.694 T and most cheap gauss meters are 200 mT full scale. Borrowing one takes the order from ~₹52,000 to ~₹22,000.
B-1: build a Halbach pair at the 12 mm design gap and traverse a Hall probe. The cheapest route to the first measured number in a project whose standing caveat is E4 — nothing built, fired or measured at any scale.
Status 2026-08-10: it is now an order, not a procedure.
docs/B1_ORDER.md.B-1 has had a method since 2026-07-29 and a bill of materials since 2026-07-30 and has never been ordered. ADR-021 froze the defect register partly on that observation: 17 model-against-model analyses were produced around a ~₹22,000 experiment that was never bought.
What the order adds beyond the method
The magnetisation sketch and quotation wording. The near-certain failure mode is a supplier magnetising through the largest face — the default tooling orientation — which yields eight identical blocks and no Halbach array. It is invisible on arrival because the blocks are correctly magnetised, just on the wrong axis. The order carries an ASCII sketch of the four-block wavelength and paste-ready wording requiring written confirmation of the axis per group before production.
A receiving inspection. Check every block individually with a compass and sort into two piles of four before anything closes across a 12 mm gap — after that it does not come apart by hand, and the mistake has already cost the fixture.
The hazard, quantified. The arrays attract at ~828 N across the gap on brittle sintered NdFeB. Captive spacer, shim outward from a closed stack, never approach from open.
The bands, unchanged
Declared 2026-07-30, derived 2026-08-06, not editable after a reading:
Biggest cost lever: the teslameter is two thirds of the total and 1 T range is binding — the model predicts 0.694 T and most cheap gauss meters are 200 mT full scale. Borrowing one takes the order from ~₹52,000 to ~₹22,000.