Probe Mounting and IR Camera

I had four things to do on my list:

  1. Use copper mesh to seal the small gaps around the top and bottom plates – After phoning metal supply places, arts and craft places, it turns out fine copper mesh is not easily sourced.  The mesh pieces I had were too small, but I ordered more from a supplier on eBay and it should arrive by the middle of September.

    I am also going to purchase the last few CF flanges I need for the vacuum chamber (not much of savings at 14 each considering they cost $15 each)

  2. The probe casing need to be properly grounded – There needs to be a clean high current path between the cavity and the probe casing which means those rolled copper shields around the conductor needs to be properly soldered to the cavity.Here is the measurement probe casing

    Here is the power probe:

    However, the problem is the probe mount, which ideally should have a copper seat also soldered to the rolled copper, sort of shown below:

    However, actually soldering the copper seat without melting the plastic and accidentally unsoldering the rolled copper shield is difficult.  For now, I am going to make sure the rolled copper casing (which from the picture, I clearly have to redo given the humongous space, oops) presses up against the probe mount   In the meantime, I have started to talk to local CNC shops to see if they can make me a probe mount from an easily-sourced standard 1.5″W by 1.5″L by 0.75″ H copper block.  It will be sloped from top to bottom and also curved from left to right. 

    With those copper probe mounts, I will build a second cavity that will have a much cleaner probe hole and mount.  The extra plastic top and bottom plastic rings will be used to hold on the end caps and it will also expose more of the cavity in order to see any hot spots building up.

    I redid both probes with flame (at which my technique is becoming much more nuanced).  Here is the power probe which I have yet to clean up.
  3. Thicker and better tuning plate – I created another tuning plate from the thicker copper sheet and it moves much more smoothly now, however, I was a bit overzealous in sanding and it is not exactly round
  4. Buy a Thermal Imaging Camera – I won the camera auction and used it around the house to learn a few interesting things:The temperature difference between a CFL and regular light bulb is considerable:

    Which would explain why CFL’s are more efficient and last longer, because they don’t heat up nearly as much.

    A few other interesting things:


    Inside my computer at idle, note the graphics card on the bottom left, how hot is that?

    The GTX 295 graphics card at idle (which, after I put a fan on it,  now idles around 60 degrees C.)

    The images above show how the process of finding hot spots has changed from a hunt and peck with my year old Extech thermometer to an instantaneous “oh, how hot is that bright spot?”  With a digital readout on the thermometer, it is really easy to miss what exactly is causing the smoke, but with the camera, you can watch the entire apparatus through 6400 pixels (80×80 resolution).  The price difference is $70 versus $1700 though! 

    The thermal images will form an important test because a large number of mundane propulsion mechanisms can be tested for by showing where the cavity is and is not heating.  Hot jets perhaps?  Weird buoyancy effect from hot air?  Twisting action of a hot cable? The thermal images will also be useful during testing to see if the cavity is in the right mode (hopefully by the pattern of hot spots) and to find early warning signs of problems.

First Results

The research stepped across a threshold this past week as we got preliminary results from the cavity.

Earlier, I fabricated the probes and soldered them into the cavity (power carrying probe to the right, measurement probe to the left):

They are roughly 31mm in diameter, give or take a 1mm.  I then bolted everything together (here I am at the lab bench at the university)

Then handed it over to Dr. Karumudi (on the left) and Kevin, his PhD candidate on the right, to connect it to the network analyzer and begin testing:

The first tests are at low power (mW), meaning it won’t move, but we will see how the fabrication has worked out.  The low power tests are critical because they verify all the assumptions we have made with regards to the simulations, fabrication, etc.  There is still a lot of work to do to make the cavity ready for the high power (1 KW).

After calibrating the network analyzer and hooking the cavity up, we used the tuning mechanism on top and generated a bunch of graphs like this:

The graph shows that we have an arbitrary mode resonating at 2.45Ghz with an insertion loss of 1.8dB.  The great news is that even before some required modifications (mentioned below), the cavity is accepting energy and resonating!  We then moved the tuning plate back and forth to see if we could find a stronger resonance, hopefully the TE0,1,n mode we are looking for.  The best Q calculated was 350 which is significantly lower then the expected 10,000+ Q but we have not tuned the cavity to the TE0,1,n mode yet.  After the modifications, I will spend a couple hours tuning the cavity to see if I can find the TE0,1,n mode.

Here is the team calculating Q (Dr. Karumudi on the left, Kevin in the middle and Adrian on the right):

Four modifications are necessary for the cavity to handle high power:

  1. Use mesh to seal the small gaps around the top and bottom plates – I have the mesh now
  2. The probe casing need to be properly grounded – The probe casings are grounded now but need a stronger high current carrying capacity, i.e. soldered in place, not just pressed up against cavity.  The modification is quite involved and requires taking the bottom plastic shell off, soldering the wound copper shielding to the cavity, sanding everything smooth again, then cutting the plastic to fit back over.  I will also solder the probes in place inside the shell making sure they are straight.
  3. The thin tuning plate is catching on the sides, warping it and needs to be thicker.  I have sanded down the plate and it moves fairly smoothly now, but I will build a second thicker one at some point.
  4. Buy a Thermal Imager – The imager will be useful for multiple tests – from testing which mode the cavity is resonating in (the pattern of hot spots) to seeing if there is a thermal gradient across the cavity that might be the mundane explanation for the movement. I will be getting an Extech I5 which has a 2% accuracy, weighs only 16oz and takes JPGs on an SD card.  There is an auction for one ending today at 5:30pm and I will be bidding at the high end of the range for second-hand I5s ($1600) .  I suspect it will cost me about $1700CND shipped.

I also got all the parts I needed to connect the vacuum pump to the chamber, and after I get a few more blanks, we should be ready to test the chamber for the first time!  The picture below shows the vacuum chamber connection at the bottom, the 3/8″ vacuum pump connection to the left and the vacuum gauge connection on top.  It is important to keep the vacuum gauge vertical to measure pressure correctly.