Should I be worried about different amounts of latency from the different outputs, for example phase problems with different drum mics on different outputs (snare on analog 1, kick on analog 2, OHs on ADAT 1+2)?

In general I don't think you need to worry as much as understand and be aware of possible outcomes based on your routing. I don't know about your specific example, but I have heard and know others that have complained that using different combos of analog and digital as well as when using different converters A2D or D2A that there were instances when lantency and other artifacts were noticed.

In general D2A has less latency than A2D. Whatever latency that is conversion specifc should be fairly consistant within that moment of time. There is going to be a difference between the Digi, the Lucid and the Alesis. How much? Hard to tell and what ever specs you can get or find are not going to be solid and accurate enough to rely on or think that you can compensate for to perfectly re-align tracks. I would think that the routing through the PC, software, drivers, soundcard that there would be more inconsistantcy even though it would be within a very small margin, everytime play and/or record button is used. Measuring latency of any kind with any accuracy requires expensive test gear under specific and consistant conditions, and even then it is often a result of an average of many consecutive tests as the result can vary every time it is measured. By the time you factor in variables such as capacitance, temp, clocking, and many other factors, you can only make an educated guess within a best and worst case range.

Just some of the current issues that half to be dealt with in the "perfect' digital audio world. So what do you do? You go through every possible combo and note the results. You then choose the one that does the least amount of damage that you can live with and still stay within your working style based on your overall limitations of the gear you have, the time you have to do it, and what you and/or the client is willing to accept.

Of course the only real true answer is that you now have go out and buy more gear... And this time, make sure you buy enough of the exact same models and brands that will cover all of your I/O and conversion needs now, and in the future.

I use a similar set up. I try to send drums, or whatever shares the same instantaneous acoustic environment, through the same set of converters. That's when latency would become an issue. Everything else, which would be overdubs, I send to available converters. Try not to bounce. That defeats the purpose of going out of the box. As far as latency consistency, I have no idea. If you want to see the latency afterplug-inson a certain channel on your pro tools window, simply control click (command-click on mac) on the little window that shows numeric level indicator. So if you are using a limiter on the snare, you bettter delay other drum channels the same amount.

I am encouraged by what you guys are telling me. Obviously I will ultimately have to go with what my ears tell me and experiment to see which sounds best.
BTW AudioGaff the Digi 001 has one ADAT I/O, one SPDIF I/O, and the indwelling 8 ins/outs. thus the AI3 and the Lucid work well to fit the bill (it isn't a case where I haphazardly bought different types of converters). Thanks, David

record a click made in PT to one channel of every diffrent D/A bring back into PT check the Latency
and delay tracks as needed
There will be differnces.

PTLE doesnt show delay in the lil box