At 09:41 AM 11/8/2009, Terry Blanton wrote:
For those interested in prior discussions on the emdrive, here is the first:
http://www.mail-archive.com/vortex-l@eskimo.com/msg15755.html
which references an article from 2002:
http://www.shelleys.demon.co.uk/fdec02em.htm
with piccys of the early experiment.
From that article (that's the Eureka article).
The device uses a resonator made of copper, filled with microwaves
from a commercial magnetron running at 2.5GHz, delivering 850W at an
efficiency of around 70%. Enclosed in an EMC enclosure for safety
reasons, the total weight of the box of apparatus is 15kg. When the
box is placed upon a balance one way up and is switched on, it
exerts a downward force of 15kgf + 2gf, and when it is placed the
other way up, it exerts a force of 15kgf - 2gf. (The force motor and
microwave generator weigh only 9.4kg, the remaining weight is that
of the EMC enclosure).
So, he's using a balance. Out of 15 kg, he's seeing a variation of 2
g. Notice that the actual weights measured are not stated, it's
highly unlikely that this thing was 15.000 kg. Any idea how difficult
it is to weigh 15 kg to an accuracy of 2 grams? Okay, it's hard to
see, but he's using a beam balance, it seems. And then we see this:
A curiosity of the constructed prototype is that when switched on,
it takes some seconds to apparently build up to full thrust. Shawyer
at first suspected that the apparent thrust might be due to some
buoyancy effect arising from heat generated within the EMC
enclosure. Careful modelling and analysis, however, shows that the
effect arises purely from the time constants of the pulsed output of
the microwave source, and the way these interact with the time
constant of the balance system used to measure the forces developed.
Apparently these are pulsed microwaves, with the pulse frequency low
enough to be able to affect a balance system. (He doesn't mention
pulsing or pulse frequency in what I've seen.)
In the article update in 2007, Shawyer says he's going to do a
demonstration where he'll "fly" the emdrive on an air bearing. Great
stuff, air bearings, I've been looking at videos of them.
The demonstration doesn't fly the device on an air bearing, it
rotates it. Instead of doing a linear acceleration test, he
apparently constrained the movement so that it would only rotate. He
has his equipment with the emdrive cavity at one end, so that it will
cause the thing to rotate. He imputes the force from the rotation,
but it doesn't look like an ordinary force on an air bearing, which
is special because of the low friction. You'd want the floor to be
very level, or the surface on which this thing is floating, because
any deviation from level would cause a torque. He did not turn the
thing all the way around.
Figure 10 on page 8 of the paper,
http://www.emdrive.com/IAC-08-C4-4-7.pdf, shows a plot of velocity,
frequency, and power. The frequency plot shows lock. There are a
number of really odd things about this graph, I've pointed them out before.
If a small force is applied to a large mass, and friction has been
reduced to negligible, which would be the case with an air bearing,
the mass will start to accelerate when the force is applied and, of
course, the force can be calculated from the acceleration, if the
mass is known. Rotating a mess, though, gets messier. Why didn't he
just move the thing along the floor? That would have been "flying
it," and if it moved more than its length, the denonstration would
have been clearly not due to mass distribution effects inside the
drive. But rotating it, of course, could have been caused by the
water pump and water circulation.
For whatever reason, the "velocity" begins to rise from zero after
the point of peak power, but before lock. at lock, the power is
declining. When the frequency locks, the velocity actually decreases
a little, then starts to rise as the power drops. When they turn off
the power, the acceleration decreases, it shows a knee, though the
velocity continues to rise for some time (about 45 seconds). Then the
velocity falls off, with increasing deacceleration, the
deacceleration reaches a value that is greater than the initial acceleration.
Therefore there is some force acting on the emdrive that is greater
than the emdrive force. And unless you know what that force is, you
don't know jack about what this test shows. The force is not
friction, the deacceleration increases as the velocity decreases.
Using a rotational measurement introduced lots of complications, eh?
If the device wasn't perfectly balanced on the bearing, if it wasn't
perfectly level, it could rotate just from that....
