Thinking about it I'd suggest riding the mixer channel faders to cope with different people's voices and mic technique (or lack thereof), far simpler and less likely to go horribly wrong than a compressor. Unless you have plenty of time to soundcheck that is. If hardware compression on vocal channels just for speech is what is required then I'd seriously consider a mixer like the Yamaha ones that have a "one knob" compressor on each channel, or a compressor that is simpler to use than having to set attack/delay times, ratio, control output gain etc. as separate functions.
To be honest, if you're recording to a computer, and the recording is all you need to compress, personally I'd just record the incoming audio "as is" into a DAW (even a free application like Audacity) then do any processing in the DAW. Far easier to add compression and any other processing that way because you don't have to commit to the results until you're happy. Also much cheaper.
The compressor plugins that come with Sonar, even the Sonnitus, can do at least as good a job as a 3630. No, make that a better job. I have three 3630s kicking about and they're OK for some jobs but "smooth, subtle and uncoloured" they are not.
Just record 24bit and keep the incoming levels well below the red.