The pics below illustrate the behavior:
Any ideas? I have a lot of breakdowns to do

I'm open to other approaches or tools suggestions as well.
Update 8/11/2018:
Found a forked version of the Papagayo software called Papagayo-NG at https://morevnaproject.org/papagayo-ng/. It's a little more sluggish but usable. I think this may be because it's written in Python where the Papagayo software I was using was based written in C. Anyway, it still doesn't solve the problem if I add text to the spoken text area. It just doesn't issue the breakdown automatically if I remove the text.