The VSU-2 sounds very similar to the VSU-100. Although the presumed speech decoding algorithm used by the TSI S14001A is replicated, there are numerous hardware differences which result in a slightly different sound.
Most notably is the use of a ~250khz PWM signal for audio generation. This requires less agressive filtering than on the original board, and results in superior speech which sounds less "underwater".

Listen to the examples on the main page to decide for yourself.

The sample rate of the VSU-100 is controlled by a 555 timer IC and RC circuit. Variation in component tolerances could create differences in pitch between boards, as well as differences due to components aging over the last ~40 years. The VSU-2 is factory programmed to simulate the pitch of an original VSU-100.

This may not be satisfactory to everyone, so the ability to program custom pitch settings over a serial connection is included. See the code and instructions at our github: https://github.com/glodstone/pyVSU

A game which did not originally include a speech board (Gamatron in particular) may require additional modifications prior to using the VSU-2.
    • A lamp matrix pass-through cable
    • A custom J3 connector (GI power, Lamp strobe #2, Lamp interrupt, voice out)
    • SB-300 / Pinstar board modification
    • Mounting standoffs to secure the board to the cabinet
A preliminary guide can be found in the following pinside post: https://pinside.com/pinball/forum/topic/gamatron-vsu-100#post-5941811

The VSU-100 was known to intermittently produce a loud "BONK" on some games in certain MPU configurations during power up.
While this is not 100% eliminated with the VSU-2, it is drastically reduced.