Talk-Polywell.org

MSimon wrote:might be acceptable for initial experimental purposes. (360 - 10 second runs).

That should do nicely.

Now how much ?

Roger wrote:

MSimon wrote:might be acceptable for initial experimental purposes. (360 - 10 second runs).

That should do nicely.

Now how much ?

$20 to $50 million depending on power supplies.

Roger wrote:http://www.dailykos.com/story/2008/7/1/ ... /43/545079

Good grief. I had to chase down three levels of indirection to find the original comment. It's actually on the MSNBC story comments thread.

ScareDuck, whaddiya think?

20-50 million for Dr Nebel to build a 100MW size machine .....

MSimon wrote:Hard to do because of the coil scaling laws. On time goes down as the square of the linear size or something like that due to coil heating.

Excuse me Simon but I dont get it, are you sure about that?

I see it like this: Lets say the linear scaling factor is X and we want to scale B at the same rate. If we keep the same geometry, material, and current density for the coils then their mass is going up at X^3, amperoturns at X^2, Joule's heat at X^3, and B at X^1 as required.

Since mass and heat scale up at the same rate (X^3) then the time any machine should be able to work before coil overheating is also the same regardless of its size (as long as we dont change anything else, and neglecting heat received from outside).

Am I making a dumb mistake I cant see?

I estimated that the WB-6 coils weighted between 6 and 10 kg and dissipated from 11 to 7 kW each (depending on wire gauge, AWG 11-9) to generate a field of 0.13 T. If that figures are right they went from 20 to 120 C in no less than 25 secs (if feeded with a constant voltage, at constant intensity only 18 secs).

That should also be the minimum pulse lenght a bigger machine with copper coils could endure (uncooled).

I figure a scaled up machine running uncooled wouldn't have any shorter test run, just a much longer time between test runs to cool off. I'm expecting power to the coils and coil mass to scale at the same rate.

MSimon wrote:The bigger machine needs longer to reach equilibrium than a small machine.

I'm guessing Simon is right, if so the mS runs wont do. It may be that 10 seconds will do, or even one second. Regardless, I think run times will have to be longer. WB-7 has about 1 ft interior space, a 160cm WB-8, over 4ft.

Roger wrote:ScareDuck, whaddiya think?

20-50 million for Dr Nebel to build a 100MW size machine .....

Sure, I'll just reach into my back pocket and ... uh ... missing that spare $20M I had there just a minute ago...

For CU coils the scaling works like this:

Resistance per turn goes up linearly with size. Number of turns goes up as the square of the size. (It must go up at least linearly just to keep the field constant as size increases.)


So the power required could go up as the cube of size. It is at least the square of the size.

No, 200million is for a whole program. 3-5 machines.

I don't think so; both Bussard and Nebel have said after WB-7/8 (Bussard called these the remining small-scale experiments) we might as well go ahead and build the 1.5M. I think that's $50M to build it, $100M to solve all the engineering/operation/etc. problems (another $50M for p-B11). The 100MW would presumably be a dodec.

The control issues alone are going to be hellacious. Someone is going to have to write some very clever code with very fast response times to keep this thing in the sweet spot.

The problem with building intermediate machines is they don't answer the scaling problems at 100MW, which are likely to be very different than those at 1MW or 1W, so they probably don't teach us much that's useful.

Bussard is actually pretty emphatic on the point:

The only next useful step is to conclude small scale work (as
described previously) and then undertake a full-scale netpower
demonstration IEF system, to show total plant
feasibility.

It is important to emphasize that there is nothing
significantly new to be gained by further tests at sub-scale
sizes (i.e. less than that needed for net power).
...
Because of this B4R3 scaling of fusion output, which makes
fusion power scale as the 7th power of size, and the corollary
5th power scaling of system gain, it is obvious that little can
be gained short of building the next system at full-scale.

So the power required could go up as the cube of size. It is at least the square of the size.

And remember, Polywells don't ignite, so there will be a hefty power bill to run the thing regardless. There is no self-sustaining reaction like in a tokamak, just a ratio of input power to output.

The control issues alone are going to be hellacious.

I think interesting rather than hellacious. The first wall problem and cooling are hellacious.

True.

But, as a programmer I'm required to say "Bah, that's just engineering."

scareduck wrote:
Sure, I'll just reach into my back pocket and ... uh ... missing that spare $20M I had there just a minute ago...

:-_)

Gee, I 'm not asking you to cough up the dough. I'm asking you is it a good idea..... if someone else, not you, like the Federal Gubermint. pays for it ?