So I just recently bought a high speed camera because I have been itching for one since 2015. I now have a camera that would make any of the can grabber races of 2015 look like Tai Chi performed in molasses. My mind has been in overdrive thinking of things, that as FRC teams we could get valuable data from filming in high speed. Most of what I have been focusing my thoughts around has been using it to examine our prototypes during build season to see all the little things that happen so fast we can only guess at what is happening. Stuff like the following.
How the balls in 2014 rolled out of catapults
What in the heck is actually going on with a can grabber while impacting a can
How the wiffle balls get stuck as they are fed into a shooting mechanism
How compression of the 2016 boulders affects the slip of a shooter wheel
That is just stuff I came up with off the top of my head in a few minutes, there is so much more. A lot of people haven’t realized that serious high speed cameras have become affordable in the $3,000-$5,000 range for total cost of entry for something that can give you access to a tremendous diagnostic tool as well as the ability to make cool looking slo-mo shots.
Here are a couple shots of my CNC Haas at work going to town on a piece of aluminum.
The best I can think of this year is how well intakes work, and seeing how variance between tests can influence how the intakes hold the cubes, but im not sure it would be that effective.
You got me curious, what camera did you spring for?
Cronos 1.4 color 16GB w/ Computar zoom lens, it all lives in a lime green 1450 Pelican Case.
You are not wrong, I need not a photon cannon, but a a photon bazooka. It is downright nuts the amount of light you need for filming higher than 3 or 4 thousand fps. Every photon counts.
Actually, I would not go with LEDs, unless they have a drive circuit with very good smoothing, for the reason seen in the above videos (I bet that pulsation is at 60 Hz…)
My description of a 1000W LED was slightly tongue in cheek. My point was, you almost need a portable Sun to provide proper lighting.
If you can synch the LEDs to the camera, or just use a pure DC source to drive them, you might get away with LED. Otherwise a small thermonuclear reactor might work .
Commonly, cheaper 120 VAC LED lights just use a half- or full-wave rectifier and current-limiting resistors. This leads to a 60 Hz or 120 Hz flicker, depending on the rectification (you can easily see this effect in most LED christmas lights). If you have an LED driver with a real AC-DC converter and smoothing caps/inductors, you should be fine with them fully on. Nearly all LED dimming systems implement PWM, which will introduce flicker unless the PWM frequency is significantly higher than the camera frame rate. Unfortunately, if you’re buying from the normal sources (amazon, ebay, etc.), detailed specs might not be available.
The Sony RX100 and RX10 (V version I think?) both provide ~1000fps shooting but the resolution is lower than the cameras mentioned here. That being said, the RX100 has been an invaluable tool for The Zebracorns in the last few seasons so this thread caught my attention. We use ours with a couple of LED flashlights and it works well enough.
I don’t have any great videos uploaded but this one was captured as a side view of a bit of fuel exiting the robot in 2017 so we could figure out a jamming problem IIRC: https://m.youtube.com/watch?v=GnKT-gYdLrk
Ok, I’m going back to my cave now… my students have started looking for a helmet sponsor for the team or at least for me.
All the actual bulbs I’ve seen tend to incorporate real LED drivers. Christmas lights don’t because you’re talking about a whole ton of LEDs in series, and the heat made by each LED is very small.
Especially if you’re talking about a 100W+ photon cannon, that needs an LED driver in order to not cook anything within 5ft of it.
I’ll be bringing my Cronos 1.4 to Chezy Champs to film robots and the competition. If you have requests for stuff you would like to see filmed, let me know.
P.S. someone should try and bring some crazy bright flashlights if you want to see anything filmed faster than about 2k fps.
I picked up a Nikon J1 in ~2013 for about $300 that was able to do 1000fps in burst mode (at a dismal resolution of ~100x200). Aside from being a halfway decent personal camera, I used it along with some black stripes on the frisbee and shooter wheels to look at slip as well as how the frisbee was deforming as it went through the chute.
In 2014, having the high speed camera was nice on our catapult because you can learn a lot about how strong/weak a structure is by how it deforms under load. It happens in a blink of an eye, but a high speed video of a robot doing something like releasing a catapult showed significant movement in the structure.
Nothing really in 2015 for us, or 2016.
In 2017, we prototyped a 2 chute side by side fuel launcher. The camera helped diagnose the shot inaccuracy that was due to balls coming out the launcher in various orientations or dragging against one side of the chute or the other.
In 2018, definitely useful for watching how the intake interacted with the cubes at various angles and where it bottomed out.
I did just pick up a new camera for personal/general use (the Sony RX10), and specifically picked it over others for the 1000fps at a higher resolution than what I have now. I’ll be interested to try it out as well in the 2019 season.
.