= Solution
The <Hall-plateau voltmeter calibration> is obtained as follows. Let $D_p$ be a plateau height in displayed units after any voltage offset is removed, with identified integer plateau index $p$. The true <Hall voltage> is $I_xR_K/p$, so the <calibration> factor would be
$$
\boxed{\lambda=\frac{I_xR_K}{pD_p}
=\frac{3.8719\times10^{-4}}{pD_p}\,\mathrm{V/(displayed\ unit)}.}
$$
The nominal current is $15\,\mathrm{nA}$. If several plateau heights are available, their inverse-integer ratios help identify indices and a fit of $D_p$ against $1/p$ can separate a constant offset from the gain. Reversing <magnetic field> and antisymmetrizing the voltage also removes an even offset. A unique absolute index should not be asserted solely from an unspecified sequence of steps.
For an established exact integer $p$ and independent current and height errors,
$$
\boxed{\left(\frac{\delta\lambda}{\lambda}\right)^2
=\left(\frac{0.3}{15}\right)^2+\left(\frac{\delta D_p}{D_p}\right)^2.}
$$
The current alone contributes $2\%$ relative uncertainty; it is a common <calibration> uncertainty and does not average away by fitting more plateaux. Voltage-offset uncertainty and deviations from quantization must also be included where relevant. The $25\pm2\,\mu\mathrm m$ width cancels because the Hall \i[voltage] divided by current is a sheet Hall resistance; it would enter the transverse <electric field>, not this <calibration>.
\b[TODO — missing source data:] the held PDF's last-page stream refers to an absent `/Im6` graphic, and its recovery log reports an unknown XObject. The supplied TeX records the same source gap. No plateau heights, labels or axis values survive, so neither a numerical $D_p$ nor a unique $p$ can be read. The numerical voltmeter <calibration> and its complete graph-reading error cannot be supplied without that graph; the formula and known $2\%$ current contribution above are determined.
Back to article page
