Hyperlapse style video of a Quadcopter flight recorded with only a mobius action-cam.
The video is 6x faster than the original footage and has been De-fisheyed. To keep the video smooth it was run through through “deshaker” plugin for virtual dub before speeding it up. The output was rendered at 1080P 60FPS resulting in a 33GB uncompressed file with 360MB/s data rate.
The artifacts and borders at points in the video are a result of the warping that occurs during smoothing of the original video. some of the borders are then filled with image data from past and future frames by deshaker.
Due to the increased speed the original audio could not be used so was removed. I replaced this with YouTube music but its not great so feel free to mute.
Built a retinal projector last night just for fun (and because i was curious what kind of image quality could be achieved).
WARNING : DONT DO THIS UNLESS YOU KNOW WHAT YOU ARE DOING. SHINING LIGHT INTO YOUR EYES CAN BE SERIOUSLY DANGEROUS. DO YOUR RESEARCH BEFORE TRYING ANYTHING LIKE THIS.
Works Well, Produces a crystal clear image straight into your eyeball. unfortunately it seems that the alignment of everything and positioning of the eye needs to be to accurate for any practical use. The plan was to come up with a setup allowing us to make some kind of FPV goggles with this however without a lot of work this will not be stable enough on your head to remain reliable. the focal point must pass through your iris so a few mm of movement will cause you to loose the picture.
How It Works:
Image is produced by the Pico Projector. it passes through a variable ND Filter made of 2 linear polarising lenses, passes through the glass, reflects off the concave internal side of the magnifier (for refocusing) before finally reflecting off the glass slide and into the users eye (placed above).
after recently building my first 250 sized FPV quadcopter i decided it was time to add some FPV gear. As i was rebuilding, i might as well upgrade the frame and have a tidy at the same time. Results below.
At this point I went for a full system power up test, the first since completing wireing ESC’s. At this point every single component is connected to the main battery connector for power.
The keen eyed amongst you will have noticed that the XT60 in this picture is wired backwards. I quickly realised this as I connected my 2200mAh nanotech. There was a huge spark and a lot of smoke. Suddenly that sinking feeling.
At this point I realised that there was a large possibility that I may have KILLED EVERYTHING. I quickly unplugged everything and reversed the connector. I then began plugging components in one by one to see what was left. I have never been so happy to hear my ESC’s arming. Same with Naze connecting to cleanflight, then video came back.
For some reason it seams that the gods took pity on me and remarkably everything survived. I don’t know how but im not complaining. Could easily have been a very expensive mistake and all the previous time spent rebuilding would have been waited.
The lesson here, DOUBLE CHECK EVERYTHING, also TEST COMPONENTS ONE AT A TIME !
425 grams all in without battery as before. 50 grams lighter and a lot more power. To reduce weight more in future I may swap to a smaller FPV cam and reduce the red stand-off’s bringing the top plate down to just above the receiver. could also swap to nylon hardware but im not sure its necessary.
So the new Hexapod hardware has been in use for a few weeks now. Thought you guys deserved some pictures of it assembled.
For a future project I need to build a completely waterproof brushless motor. Before I start at the deep-end I decided it would be best to make a smaller motor to get a better understanding of how brushless motors work. I ordered a few stators and a kit from gobrushless.com.
The motor has a 9 pole stator (3 poles per phase, 3 phases) and 12 neodymium magnets in the outer ring.
I wound each phase a different colour so that it was easier to follow what I was doing. the phases are joined together in the wye winding style.
Now to test it out:
So this morning I was casually browsing the Epilog (laser cutter) website when I stumbled accost a testimonial that said “We have engraved everything from denim to acrylic with fantastic results”.
I was instantly intrigued. The idea of being able to easily make professional looking custom clothing is amazing. I quickly grabbed the closest pair of good jeans plus an old ripped pair to test on.
Whilst looking for a simple pattern to etch I remembered these designs by mattcantdraw.
With a little rearranging these designs were perfect. I did some tests on an old pair of jeans to find the right power level as I could not find any settings for this online. I started using the same setting that would be used to etch leather however quickly found that this was to powerful and made the material break apart easily. I finally settled on 100% speed 15% power on a 75Watt laser.
I had some trouble fitting the jeans into the cutting bed and found that things had to sit flat so that I did not obstruct the gantry or cutting head.
I’m actually very pleased with how this turned out, the effect produced looks like natural fade, but only in specific areas. I love that these jeans have gone from boring ‘off the shelf’ to something interesting and individual. From now on I will be keeping a look out for other designs that would work like this.
Recently iv been working on a 2nd revision of the chassis used for my hexapod. So far I have redesigned the legs and have a clear idea of how the new body will look.
The new design focuses on increasing the range of motion, functionality and reliability. I have also been trying to get rid of the ‘prototype’ look of the old design in favour of a sleeker and more stylised appearance.
These parts have been laser cut and are now awaiting a sunny day to get painted. I have also painted a spare leg in the planned colour scheme as a tester. This is slightly messy due to the fact that it was painted whilst assembled, the new legs will be painted before assembly so this should not be an issue.
The long-term plan is to eventually release all of the designs for this hardware as open-source. This will probably happen once I have finished the hexapod and am happy with all parts of the design. Hopefully that will not be to far away but development is slow at the moment with many coursework deadlines approaching.