Close

Preliminary Tests (swimming platform)

A project log for EMDrive/satellite

Developing a small fuelless microwave thruster

paul-kocylaPaul Kocyla 07/05/2015 at 13:358 Comments

Optical markers (LEDs) are tracked with a camera to determine the angle of the platform.

It shows that the platform reacts very sensitive to high power loads in the house. When the TV is on, it gets noisy. When people take shower in the morning hours (electric heating here), the angle lowers significantly. Where the jump in the pic below comes from - I don´t know, maybe a neighbor toggled some power load.
For the preliminary test, I took the data from the flat areas of the graph - in this case in the sections before and after the jump.

The cavity has been turned on and off alternately. For the preliminary tests, two ON and two OFF phases were recorded with cavity in CW and CCW direction. Green graphs show cavity OFF, red graphs show cavity ON. Due to the jump (see above), the second pair of values is raised.

Cavity attached to push CCW:

When the cavity is on, the angle drops below the off-level.

Second test with cavity meant to push CW:

The angle rises above the off-level.

This looks plausible so far. Anyway this is neither a proof nor a real valuable data.

Another think I noticed is the poor resolution in determining the angle.

It´s because the marker detection algorithm works only on pixel accuracy. I will improve it to calculate the marker position on subpixel accuracy.

Although I am happy to see these results, it might be just a coincidence that they look good. I´ll need to perform long duration tests to be sure.

Discussions

Marvin Macportain wrote 07/08/2015 at 22:00 point

You're on the right track, glad to see that my magnet idea was fruitful!

  Are you sure? yes | no

Chris wrote 07/07/2015 at 14:02 point

If the shack in the woods is not an option, insulating the test chamber with a Faraday cage (e.g. metal screening or aluminum foil may be good enough for your purposes) and powering the system via a battery may isolate the system from most noise.

Also - is there any chance the platform is getting hot enough to generate steam?  This could give you false positives.

Otherwise - very neat setup!

  Are you sure? yes | no

Paul Kocyla wrote 07/06/2015 at 20:19 point

Ah yes, seen this one - motion amplification.
It´s something different. In our case, we just need to increase the resolution of our coordinate system. The algorithm is pretty simple, here´s my code. It calculates the center of gravity of an image area based on pixel intensity. Simple and works well:

int cnt=0;
 double fx=0,fy=0;
 for (int x=xp-wsp;x<=xp+wsp;x++)
  for (int y=yp-wsp;y<=yp+wsp;y++)
  {
   fx+=(double)iimg[y][x]*x;
   fy+=(double)iimg[y][x]*y;
   cnt+=           iimg[y][x];
  }
 fx/=cnt;
 fy/=cnt;

  Are you sure? yes | no

lolno wrote 07/06/2015 at 02:44 point

This is a big improvement in the rigor. Being able to see the fluctuations of the apparatus overall and compare directly on and off really help.

  Are you sure? yes | no

Paul Kocyla wrote 07/06/2015 at 09:19 point

Thanx. I improved the tracking algorithm over the weekend and added a function to automatically track the on/off status.

The software is now able to automatically calculate the means of the on/off phases.

I ran a CW test over night (hoping the disturbances at night are lower) and it looks really good.

This night I will run a CCW test and post the results.

If they show positive results, I will run a bigger series to be sure it´s not only a coincidence.

  Are you sure? yes | no

Taven wrote 07/05/2015 at 16:40 point

The setup is looking great, but that's still quite a bit of noise. Your environment seems to be your number one enemy. Is there anywhere else you can run these tests? I don't know how practical it would be, but maybe taking a weekend to run the tests in another location, like a friend or family's house where you can turn everything off, or in a shack in the middle of the woods would be worth the hassle to clear up the noise. Otherwise you might keep running test after test with too much noise and that would just waste your time and efforts. A few very clear tests would be better than fifty somewhat noisy tests. In any case, keep up the great work!
Edit: Also, this might be a bit silly, but I think you should use green for on and red for off, like a traffic light. It's somewhat hard to imagine them backwards, but that might just be me!

  Are you sure? yes | no

Paul Kocyla wrote 07/05/2015 at 17:03 point

Thanx for the suggestions. Maybe there will be less EMI in the company I am working in. At night, there is nothing running, it´s worth a try. The forrest shack would be great, I wish I had one. According the colors: makes sense, I´ll change that :-D

  Are you sure? yes | no