[morzh]

Well, it is just a simple parametric regulator, standard circuit, but it will need a heat at any sensible current. A 50mA at 25V dropout (which is not even what it can be) will result in 1.25W and TO-220 will not like it without at least some sort of a heatsink.
At 100mA 2.5W will make it not touchable.
The R(theta-j)A is 62.5C/W. So 2.5W will raise the crystal temperature by 156 degrees C. At room temp it already is 31 degrees above the maximum allowed.

So I hope they have that thing mounted on a heatsink.

This said, yes, a regulator does a good job to eliminate ripple.

But. If they used a classic regulator (op-amp-based), they would achieve better cleaning. Like an LDO.