its just a header issue.
it was expecting a 44.1kHz signal and you were sending 96kHz. that's why the clock was moving so fast.
if the device counts 1 second for every 44.1k samples that are sent, but you're sending 96k sample per second, the counter will go a little more than twice as fast as normal.
there are lots of threads dealing with this topc, as the MT isn't the only recorder with issues like these. in fact, I think there's a sticky thread somewhere regarding this topic.
anyway, try this:
before you turn on the MTII, ahev the coax connected and the V3 sending a 24/96 signal. then, when you turn it on and set it to S/PDIF, it'll hoepfully lock onto the 96 kHz signal and set itself appropriately.
edit to add: this was an issue with the original MT. if the S/PDIF signal wasn't active when the MT was turned on, it would write a a file with the wrong header info