Tim,
Modelhead is being modest - you will be hard pressed to find better feedback than he has provided. The only info that I would add is what I always reference when looking at the scope of paying work. What is the client expecting? What is your deliverable? Is this a one-off model or will you be working on a whole series of similar appliances. I think Modelhead's time estimate is pretty accurate if the intent is to provide an efficient native sketchup model that is of comparable detail to the cadd file. For one model - probably not worth it. But if you will be doing multiple and can share common or similar components - working from scratch in SU becomes much more practical. Two more things I would share with you;
1 - Ruby scripts can do a lot of the time consuming tedious work and they should be carefully considered. Looking closely at the ones available to you may provide you with new methods to model, and you can often learn new tricks just by looking at how some scripts approach specific tasks.
2 - While other applications may be quicker or slicker at certain modeling functions (like organics) in general - SU is a very capable and efficient modeler where the only limit to the level of detail you can achieve is your skill and patience it may take to "get there".
Good luck and have fun while you are doing it!
Bytor