I'm pretty sure .1 second intervals would be fast enough, assuming you meant that the way I understand it, as ten frames per seconds, rather than as .1 seconds slower than the signal's normal repeat rate.

What did you do to cause or measure the .1 second intervals? I didn't think the Pronto gave you that sort of flexibility.

Post the Pronto Hex for the signal. We can then see whether it is one that needs to be broken into two parts in order to be extended by a sequence.

What model Pronto are you using? The NG Pronto can't reduce the delay between steps in a sequence enough to accomplish what you tried. Some of the older Prontos can. But as mentioned above, some NG Prontos have the ability to directly specify duration, which is much simpler than constructing it as a sequence.