TL;DR: The Sun is well-resolved in small telescopes so a focal plane occulter works well. Other stars are not resolved (except in a few exciting cases using interferometry) so an internal blocking disk would be useless.
Local blocking disks for telescope coronagraphs are placed inside the instrument at a focal plane, usually the first focal plane. That way they have sharp edges and can block the Sun (or star) but allow light nearby to reach a detector.
A blocking disk "on an arm" some short distance in front of a telescope would be way way out of focus and so act like a "fuzzy blob" at best. SOHO does not have one of these.
It does have a small disk on a small arm inside the telescope.
The disk of the Sun is huge compared to the resolution of the telescope, roughly 2000 arcseconds compared to order 1 arcsecond resolution (depending on wavelength). So a disk at the focal plane is sufficient to block the Sun and allow the larger corona to still reach the focal plane.
However a star's angular size is way, way, way smaller than the resolution of a modern space telescope. There's no hope of blocking the star and letting light from nearby orbiting exoplanets to reach the focal plane.
So the only alternative is to move the disk so far from the telescope that it appears as small as the star itself, and blocks the light.