When you don't know what you are doing it is impossible to design a minimum cost approach.Bringing it back to that first off topic; MSimon, how would you structure the design-build-operate team to bring it in at the lowst cost, and how low do you think that cost coould possibly go?
We did get a big help when some one recently found a very low cost power supply vendor. About 25 cents a watt for HV megawatts. I was estimating $1 a watt for a very long time.
1. We need a continuous operation machine. SC or LN2 depending
2. A doubled size pulsed machine - for scaling
Once we have data from them we can think about WB-100.
For WB-100 I'd go with the approach suggested by Brooks. Six or eight master surgeons (depending on project breakdown) supported by teams expert in the surgeon's field.
I would not break it down along the lines of a drafting dept. A machining dept. etc except for general support functions like accounting. And maybe not even then. Maybe we contract out payroll. Each team is autonomous. No vying for people and tools. I don't want scheduling conflicts to get in the way of getting something done. If there is enough work to keep 1/2 person busy we hire a whole one. If they have spare time that goes into cross training.
If we have a lot of 10% tasks from different groups that might be handled by one person. We make sure the loading on that one is not above 70% or so. That gives surge capacity.
The cheapest way to get things done is to design a fair amount of "waste" into the system. Bottlenecks are what drive costs the most. Six teams waiting for the seventh to finish an activity.
PERT chart it to death before you start. Timely revisions as knowledge is gained. Again: spare resources to throw at bottlenecks.