i don't believe the exact answer to this question is known. i haven't seen any research that mentioned the effect of various interval lengths when training. therefore, all i can provide you with is my best guess. your mileage may vary.
by "intervals" i assume you're referring to the time between each trial, where the default length is 3 seconds in most DNB programs. i haven't experimented with different interval lengths, but i imagine shorter intervals will be more difficult, more taxing on your working memory. i also have a sense that the more difficult the training is, the better.
so if i had to guess, i'd say that decreasing time per trial could make your training more effective because i imagine the task would become substantially more difficult. furthermore, you could also complete more sets per unit time because each set would take less time to complete. again, i haven't tried this, but i think it could work. i don't see anything wrong with experimenting, i make changes to my training routine all the time.
hope that makes sense. sorry i couldn't give a conclusive answer, i think very little is known about this. i'm happy to help if you have more questions.