Estimates of the location and strength of electrical sources in the brain can be made using the magnetoencephalogram (MEG) which is the measurement of the brain’s magnetic field. Since the size and configuration of the field‐sensing coils used in MEG detectors can affect these measurements, the effects must be known and taken into consideration in making these estimates. This paper uses computer modeling methods to determine the effects of coil size and configuration on measurements of a dipolar source in a spherical model of the head. The measurements are as follows: a magnetometer, which uses a single coil; a first‐order, in‐line gradiometer, which uses two coils at different distances from the head; and a first‐order, side‐by‐side gradiometer, which uses two side‐by‐side coils at the same distance from the head. For the magnetometer it is found that the effects of coil size are large enough that the source depth can be overestimated by as much as 0.65 cm and its strength underestimated by as much as 31%. For the in‐line gradiometer, the effects of the back coil on measurements of sources near the surface of the sphere are small as compared to the effects of the finite size of the front coil; for deeper sources, the effects are found to be large but nearly independent of the size of the back coil. For the side‐by‐side gradiometer, the effects of coil size are large enough that the dipole depth can be overestimated by as much as 0.45 cm and its strength underestimated by as much as 37%. In addition, the effects of tilting the two coils used in the side‐by‐side gradiometer are investigated. It is found that this tilting does not cause the measurements to be more localized over the source. Finally, the effects of coil size on estimates of source depth made from side‐by‐side gradiometer measurements of the falloff of field amplitude with distance from the sphere are determined. It is found that the effects are large enough to cause the depth to be overestimated by as much as 0.5 cm.