Assembled the Near Eye Display 3:12 AOM Breakout Multiplexer:

Dump of Switched AOM Breakout Board notes here.

Switched AOM Breakout board design finished and sent to OshPark:

Code pushed to GitLab.

Most important developments on the Near Eye Display (NED, also for flatscreen display):

• Low frequency repetitions are most likely from internal reflections, but shortening the laser pulse down to 50 ns helps also
• Smooth Focus demonstrated in two ways
• Two-Channel operation demonstrated
• Need a cylindrical lens to vertically collapse streaks to points

Manually tracking the camera focus to match the moving computer focus NED output:

Two Channels & Cylindrical Lens:

The above two videos come from videos 3 and videos 4/5 respectively, from the following list:

1. https://youtu.be/4oTfn4wY0Z0
2. https://youtu.be/WOG0x4QV444
3. https://youtu.be/XLZjmIMpuYE
4. https://youtu.be/SeAKL0SUPkY
5. https://youtu.be/K0d84rrv0gc

Excellent frequency response from the new DVI Breakout (Rev. 28 Mar 2017) and from Jesse's RF chains:

Multiple waveguides (about a dozen) with light coupled in:

Near-Eye: v1.1 report submitted

Jesse's boards work fantastic!

Cleaning Near-Eye setup and using our Mixer-Filter-Preamp-Preamp boards.

Setting up dual-channel Near-Eye… mwahahaha…

If you want to get results you've never achieved, do something that you've never done before.

Near-Eye:

• A - Camera image of continuous input frequency
• B1 - RGB components of camera image
• B2 - HSL components of camera image
• C1 - Lightness component of HSL camera image
• C2 - Frequency domain of Lightness component
• D - Matching prominent frequencies with features
• E1 - Low pass (purple) / High pass (cyan)
• E2 - Low pass (purple) / unfiltered (purple+red)
• E3 - High Pass + envelope
• E4 - Envelope reimposed

Results:

• High-frequency ripples have period of about 268/41 = 6.5 pixels
• Sync envelope isn't a perfect sync, but it has a period of about 268/7.75 = 34.6 pixels

New Questions:

• Why is there a near zero at Frequency Bin 42 (the chosen LP/HP corner frequency)?
• Why is the camera seeing light coming from different directions? These side lobes are visible at all nominal horizontal positions, which, if they were coming from RF sources, would be different tones…
• Could the output grating be causing this?

Next Action:

• Implement + verify one or two simple window functions in MATLAB.
• Edit Near Eye code to accept amplitude arrays imported from MATLAB to test apodization filtering.
• Test code.

Interesting Links:

• https://en.wikipedia.org/wiki/Fiber_Bragg_grating
• https://en.wikipedia.org/wiki/Hann_function
• https://en.wikipedia.org/wiki/Window_function

Near-Eye:

• 0.6 Amplitude sharpens center lobe
• Near camera focus shows different image
• sync envelope from rectangle window
• Apodization (Hamming window?) to remove sync envelope side lobes
• Spectral Leakage

Other:

• Murata RF Switch Connectors
• SMA Male End Launch Connectors

Worked on the Near Eye Display (with a slanted interface between the AOM and the output grating) and Showed Madison its output. Unfortunately neither a polarization filtering effect nor the needed streak on the output was seen. The following video shows what we saw. The output grating Steve made for this works great. We'll use the flat 90 degree face on its other end next time.

Worked on the public HoloMonitor website, revising div's organization, learning CSS…

This video demonstrates focusing a chirp in the Near Eye Display and a video camera independently to different depths, demonstrating accommodation. JESSE helped me to film the video.

Note that this is not 100% convincing to me that this system is working as intended. Sometimes the chirp appears in focus even when I consider it should be out of focus. There wasn't a good example given showing different rates of motion parallax when they're focused to different depth. The chirp is not as crisp at ~1.5m as it is at ~3m; it would be expected to be crisp at any distance when it and the camera are focused on the same depth.

- post a video showing how the software works

INCOMING HI-FIDELITY DATA AND ANALYSIS ON SCANNER SHIELD BEHAVIOR

Wrote a new Wiki page on custom polygon mirrors.

Wrote a new Wiki page on the ECEn Shop.

Wrote a new Wiki page on the big laser cutter.

Steve's grating works! (as far as we can tell)

I designed and sent off for fab the SCA-3-11 splitter board and I updated the overview section in the mixer-amplifier_segments page.

I've written specifications for a protocol to stream commads to the Leia Display: leia_display_protocol.

Wrote some Matlab code that takes averages over a couple of video frames and isolates the AOM's output:

I tried aligning chirps using either the top or bottom output dots (as seen below) to fixed locations in the camera feed. I would take a whiteboard marker and place a mark on the computer monitor connected to the camera and move the dot of each segment separately until they were more or less centered with the mark I had placed. Often times, the dots were very dim or it was difficult to see their boundaries. Sometimes there was ambiguity as to which exact frequency would center the dots with the marker. Sometimes these frequency measurements were uncomfortably speculative. The bottom dot seems to behave more linearly than the top dot in most tests. The individual tests focused only on one dot.

Demonstration of Near Eye Segments:

Screen grab recorded with OBS Studio and combined with camera footage in Adobe Premiere.

I also tested (with Scott's help) solder-joining two RF PCBs together instead of soldering on SMA adapters.
I took a pass-through board (just a waveguide, no components) and measured its S21 transmission characteristics on the network analyzer from 50-990MHz.
I then chopped it in half in the ECEn Shop, placed it in a PCB vice (which perfectly aligned the two planes with each other), soldered a straight strip of PCB to its back, solder-bridged the ground planes on both sides, and soldered a 0.0 Ohm resistor across the signal trace (I used a 0.0 Ohm resistor for consistency, instead of a solder blob).
I took it back to the network analyzer and found near-exact-same characteristics as before.
I found that this also had near-exact-same characteristics as well to a Female-Female SMA adapter, in a frequency sweep test measured by the spectrum analyzer.
Data was gathered with GPIB commands (see

• Test different boxcars of frequencies with gaps between to account for blur
  • have frequencies of individually-controllable offsets
• Do software tests first before moving onto the thick samples
• Use Photodiode/Photoresistor to measure laser pulse width?
• Use a new single-mosfet driver?
• Test with Thick Samples
• Retest grating mirrors

I tested triggering offsets and found somewhat expected results:

Near Eye Videos

See MVI_9579.MOV; skip to 0:40 for frequency sweeps, 1:00 for offset sweeps.

Cleaned up the Monitor top page and made a new electrical block diagram.

I tested 6 samples in two orientations each.

I am revising the system block diagram so that the Senior Project people know how to assemble the components.

A simpler test apparatus was set up to look for expected artifacts in the Near-Eye output mirror gratings. No expected artifacts have been found yet (in either the 60° sample or the known working sample).

I tested the 1-side-roughed sample with the AOM and observed blobs moving. It's possible that the blob travel range was limited by the Wed 8 Feb 2017 Bandpass Filter's upper cutoff frequency (which starts around 475MHz).

I drafted and sent off to OSHPark a board for the Monitor's VCO (for a ROS-535). For the MIT Mark V HoloMonitor, this board was custom-made, so I needed to make this board anyways. 6 copies were requested ($5/in * 2 sets of 3); board fab time was expedited (an extra $5/in for “Super Swift” service; 5 day fab time) and shipping time was expedited ($5 for “USPS - Priority Mail”; 2-3 business days).

The board is under J:\groups\holography\~Monitor Development\~Mixer-Amplifier CCA\~Mixer-Amplifier Sections\Local Oscillator CK605 28feb2017.

Wrote up the initial documentation for the Monitor's VCO.

We got a good filter working today!

It was originally designed for 350-600MHz (green plot), but because of the limited parts available, it was expected to go between 400-600MHz (blue plot). Jesse scouted out what parts we had.

Initially it was soldered with the reflow oven, but several parts drifted around in the molten solder. I moved them back to their intended locations with the shop's excellent soldering tweezers next to the reflow oven / refrigerator. It was surprising to find that the filter worked mostly between 350-500MHz, but I'm happy for it because of the sharp drop-offs and relatively-flat pass-band range.

Thanks also to Scott for the advice and encouragement.

Submitted my algorithm recommendations for the SMFoLD project.

Explained to Kamran principles of HoloMonitor operation & timing (see picture).