Alright we've discussed this issue amongst ourselves, and here's what people said:
It does appear that there is something occurring (maybe phase cancellation) going on in this case. Both the delays are set to the same time on this preset "Another Brick" in Digital Delay and you hear double to delay time on the repeats. Instead of 450ms delay for each repeat, you get 900ms delays only (after an initial one at 450ms). When you add a slight modulation or offset a delay by a ms the repeats change to their normal length.
I checked out some waveforms. It appears that the A delay in this case is getting phase flipped every other repeat. Any ideas why or is this how it's supposed to be?
I can hear what he's talking about. It happens in TF as well. I can see where it is occurring in the sigfile. The Feedback coef in DlyA is negative so it's flipping the phase with each successive pass. With A and B mixed equally with the exact same delay times, this will of course cancel out. A little modulation goes a long way in that the waveforms are never exactly opposite, so it goes away. Equally, changing the delay times a little or the mix will get rid of it. I think this was probably intentional, because it creates more interesting evolving feedback effects, especially when combined with modulation, but we'd have to ask Dan.
Either way, the easiest solution is to just use one of the delays with no modulation if wants that clean single repeat.
This is intentional, as Russ suggests. Think of it as a feature, not a bug. If you want it to go away, pan to delay A or B.