[thecrow1213]

> The obvious explanation is that they have an error in the measurement.

Then why have 6 different teams verified it? Not saying EmDrive is the real deal but either the error is hard enough to catch 6 teams missed it, or theres something else going on. Saying all these teams are wrong, to me, is far from "obvious".

[gus_massa]

Do you have a link to the 6 experiments? I'd like to compare the results. This link has a list of 3 or 4 experiments before December 2014, that is more than 1 year ago. http://forum.nasaspaceflight.com/index.php?topic=36313.msg13... I'd like to compare that with the new results.

With the "current" physic laws, the theoretical maximum of ForcePerPowerInput is 1/c = 0.0033 mN/kW. In that table, the ForcePerPowerInput varies from 300000x to 3x. That's a lot of variation, not and exact value that coincides with a theoretical prediction. It's a pity that the list is not ordered by date.

If you order the experiments by date and all of them have roughly the same result, they probably are measuring a well known effect were all the variables are well understood. Like measuring g with a pendulum.

If you order the experiments by date and the value increase a lot, it's perhaps a new phenomena that is still not well understood and they are tweaking the materials to get more efficiency. For example, let's consider measuring the critical temperature of a high temperature superconductor. If you pick a fixed simple superconductor, you expect to get approximately the same result in any laboratory, but small changes in the fabrication process can increase or decrease the temperature. But any time someone discover a new superconductor material or method of production, you will get a new record, so the world record will increase and the other laboratories will try to reproduce and increase it.

If you order the experiments by date and the value decrease a lot, it's possible a sign that they are fixing some experiential details and reducing the experimental errors, and they get a smaller result because the correct value is 0.

[mcnamaratw]

Yeah imagine if someone had made similarly exciting claims about an unexpected and so-far impossible-sounding wrinkle in how gcc works. A bunch of people tested gcc and got all different results. Nobody cites any of the results or discusses any detail.

Everyone would be asking to see the code and the output for themselves before they got excited.

Wikipedia says that the experimental tests disagree on even the sign of the measured force:

"An article published by Shawyer in Acta Astronautica summarises the existing tests on the EmDrive. Of seven tests, four produced a measured force in the intended direction, and three produced thrust in the opposite direction. Furthermore, in one of the tests, thrust could be produced in either direction by varying the spring constants in the measuring apparatus. Shawyer argues that the thrust measured in the opposite direction is the reaction force from the drive, and therefore it is consistent with Newtonian mechanics.[1]"

https://en.wikipedia.org/wiki/RF_resonant_cavity_thruster#St...

[yxhuvud]

One thing to note though I'd that the theory of the OP could actually explain that sign shift.

[mcnamaratw]

Another way to explain the sign shift, though, is that the thrust might be zero and experimental errors come out both positive and negative.

[simonh]

Obvious explanation, not only explanation.

Would it really be do surprising if similar experiments all working from the same basic design using similar components under similar conditions suffered from the same design flaw?

[jakub_h]

Wasn't there a problem that while there was an anomaly of some sorts, different teams still reported significantly different results? That would still mean that n-1 teams were wrong, which could mean that the nth team could have been wrong as well. Or was that some other similar device?