Close
0%
0%

Edgerton, A High-Speed LED Flash

Affordable photography tool used to capture images of bullets with no apparent motion blur

Public Chat
Similar projects worth following
Taking pictures of very, very fast things is called High-Speed Photography. It's usually done in a dark room and the action is frozen using a flash which is triggered at just the right moment. A typical camera flash is too slow and creates motion blur, so a high-speed flash is important.This high-speed flash uses LED's which are pulsed for 0.5 - 4 microseconds - fast enough to photograph a bullet with no motion blur. Because the flash is so fast, the LED's need to be very bright. The LED's are very powerful given their price, and the output is further enhanced by using a very high voltage (much higher than the LED's rated voltage). The optimal drive voltage was selected by testing over 100,000 flash cycles.The original and complete write-up can be viewed at www.gerritsendesign.wordpress.com

Complete Description Website ~ Hackaday Post ~ DIY Photography Article ~ Hackster.io Post


What Is A High-Speed Flash?

A high-speed flash is a photography tool used for high-speed photography (unsurprisingly).  Photographs of bullets and other projectiles travelling at incredible speeds can be photographed using any camera (even a cell phone) with the help of a flash.

Glass Cup Hit By Bullet

A normal camera flash is not fast enough to capture these images!  Below is a comparison of a typical camera flash vs this high-speed flash (1-μs duration).  Both air rifle pellets were fired from the same rifle, travelling about 280 m/s.  This demonstrates the advantage of using a high-speed flash for high-speed photography.

DIY High-Speed LED Flash
Nikon SB-900, Minimum Power
DIY High-Speed LED Flash
Egerton, 1 Microsecond

Anybody Can Do It

While typical camera flashes require a trigger system to fire, Edgerton is going to be equipped with a built-in tripwire trigger function.  All you need is a camera, a projectile, and a piece of wire!  The tripwire trigger works by placing a small wire strand in front of the projectile's path (usually in front of a rifle's barrel).  The user enters a delay into Edgerton, and it will flash exactly that long after the projectile has broken the tripwire!

Edgerton can also use the same trigger system as a typical camera flash.  This means it's compatible with DSLR's or wireless trigger systems.


Are There Alternatives?

High-speed photography is no recent invention.  Doc Edgerton was already experimenting with high-speed photography in the 1940's, and has taken some incredible photos.  He was known to use an Air-Gap Flash, which is similar to the Xenon flash tubes found in modern camera equipment.  It unfortunately requires much higher voltage which could easily cause injury (SEVERE injury).  While I didn't completely disregard the option, I eventually opted for a safer solution.

A recent kickstarter campaign (Vela One) offered a high-speed LED flash.  It can produce flashes lasting 0.5 μs (microseconds), was not dangerous, but was priced about $1,750 CAD!  A significant amount of development resources was put into this commercially-available flash.  I've likewise conducted destructive testing with expensive components and put much time into developing the Edgerton. 


Feature Set

  • Simple user interface consists of an encoder, 4-digit LED display, and power switch
  • Tough plastic case for real-world use
  • 3.5mm Jack for multiple types of external triggering
  • Arca-Swiss mounting plate compatible with many tripods
  • 1/4" threated hole for non-Arca-Swiss tripods
  • 12x high-powered Cree LED's are capable of well-exposed images in one microsecond
User Interface
DIY High-Speed LED Flash
High-Power Cree LED's


User Manual 0.2.3.pdf

User manual for current firmware version

Adobe Portable Document Format - 37.90 kB - 06/19/2019 at 02:22

Preview
Download

edgerton 0.2.3.zip

Firmware for ATMega328P microcontroller

x-zip-compressed - 21.84 kB - 06/16/2019 at 19:19

Download

Assembly Manual 1.1.pdf

Complete guide for assembling Edgerton

application/pdf - 4.99 MB - 06/08/2019 at 17:21

Preview
Download

BOM 1.1.csv

Bill Of Materials, everything required for assembly

- 11.27 kB - 06/08/2019 at 03:22

Download

STL 1.1.zip

STL files for 3D-printed components

application/x-zip-compressed - 1.67 MB - 06/08/2019 at 02:39

Download

View all 6 files

  • 12 × Cree CXA2530-0000-U20E3 LED’s White LED's with high output/cost ratio
  • 4 × Vishay MKP1848S 10μf Capacitors To supply the LED's
  • 1 × 45-390V Boost Regulator To supply the Capacitors - no model number, no datasheet, just search the description on eBay
  • 1 × ATMega328P Microcontroller The brains
  • 1 × LM7805 Linear Regulator To supply all the 5V components

View all 18 components

  • Edgerton V1.1 Released

    Tyler Gerritsen06/09/2019 at 15:14 0 comments

    What's New?

    The design for Edgerton has been updated to address several issues in the original design.  The circuit diagram, bill of materials, and 3D print files have been updated.  A new complete assembly manual has been written and made available.  The overall cost has increases slightly (less than $5) because of some new components.  Assembly requires a few more steps, but they are simple and the benefits greatly outweigh the cost and added assembly time.  Below is a brief overview of the changes and their rationale.


    Reduced Power Consumption

    The microcontroller's isolation from the high-voltage boost converter is important to prevent resetting, and this was originally accomplished using a relay board purchased from eBay.  Unfortunately the relay drew several hundred milliamps, about as much as the boost regulator.  

    The relay has been replaced with a P-channel MOSFET, opto-isolator, and a couple resistors.  The new circuit draws about 30 mA - and may be reduced further with some optimization.  Isolation is maintained via the opto-isolator.  The cost has not increased and the circuit is easy to assembly and install.  A small benefit is the removal of the relay mount in the case, which means a reduction in filament usage.


    Ferrite Beads for the Gates

    I previously noticed some significant ringing in the MOSFET gates.  Thanks to advice from a friendly German named Volker and guidelines presented in this application note from ON Semiconductor, I ordered some ferrite beads and installed them at the MOSFET gate leads.  Here's the gate signals before and after the ferrite beads were installed.

    The gate signal is cleaner.  Here's a trace showing how much the ferrite bead cleans up the signal.

    Read more »

  • Major LED Failure!

    Tyler Gerritsen06/03/2019 at 03:41 0 comments

    The Incident

    After days of torture-testing a single LED before it even started to show evidence of damage, I managed to fry all twelve LED's in a single unintentional event.  That's $80 bucks of electronics GONE! 

    I felt like Mark Watney from The Martian (the scene that happened in the novel but was sadly missing from the film...  If you read the book, you know the part).  To quote XKCD:

     I have never seen a work of fiction so perfectly capture the out-of-nowhere shock of discovering that you've just bricked something important because you didn't pay enough attention to a loose wire.

    The event happened while I was experimenting with a switching regulator.  I hoped to decrease the number of AA batteries required, so I grabbed a boost regulator from eBay and set it to 12 volts.  Then powering the boost regulator from a 5V power supply, I started testing current draw during operation.  So the LED's functioned before these tests, but didn't work afterwards.

    Read more »

  • Phosphor LED Turn-Off Time​

    Tyler Gerritsen06/01/2019 at 20:14 0 comments

    The Potential Problem

    I've received a few comments regarding the type of LED that Edgerton uses.  Specifically, the LED's may not turn off fast enough.  This is a problem I didn't really consider previously, given a statement in the Cree application note (pp. 10: The typical turn‑on time for an device is on the order of 10 nanoseconds or less).  Since they turn on in 10 ns, why can't they turn off as fast?

    Well, it turns out that the LED's use a little trick to make white light.  The actual LED modules make a blue light, and a layer of phosphor in front of the LED's turn the blue light into a spectrum of light ('white light').  While the LED's can turn on and off quickly, the phosphor is delayed in turning off (and maybe on - I don't know yet).  Notably, Edgerton's primary competition (Vela One) appears to use a similar type of LED and the manufacturer defends their flash durations on their website.

    Experimental Proof Of A Minor Problem

    Here's an experiment that demonstrates the problem.  It's a photograph of the LED flashing a 1-microsecond pulse.  There are three stages of the flashing cycle observed here:  The bright strip of light is the 1-microsecond flash, the illuminated dots above are the LED modules AFTER the flash, and the darker area below are the LED modules BEFORE the flash.

    There were MANY failed attempts to take this photo.  Ignore the absent LED modules, this was a damaged chip.

    Before explaining how this image works, I'll explain what it tells me.  The LED's above the flash strip show that there is some lag in turning off.  But during the turn-off time, the phosphor doesn't output much light.  In other words, the LED is incredibly bright for one microsecond, but then becomes dimly lit for some time afterward.

    Read more »

  • 150,000 Flashes Later

    Tyler Gerritsen05/29/2019 at 02:25 0 comments

    Strobing the LED's at several times their rated voltage is one thing, but a good flash needs to have a useful lifespan.  I began testing the lifespan of the unit a couple weeks ago.  Testing was done using only one LED (too expensive to risk multiple LED's) strobing at 1 hZ.  I started at 70 V, cycling 10,000 times at 1 microseconds and 5,000 times at 4 microseconds.  The LED was aimed at an 18% grey card 20 cm away and a DSLR photographed the card every 20th flash.  After the LED cycled 15,000 times I incremented the voltage +5 V and started again.  At 95 V, the LED showed no sign of quitting so I increased the flash rate to 2 hZ and photographed every 50th flash.

    Well, it turns out that the LED could handle all the way up to 125 V - and I stopped the test as the circuit board wasn't designed for more than that!  The photos were batch analyzed with ImageJ and no fluctuations in light output was observed.  Above is a photo of the LED with 150,000 flash cycles vs an LED that didn't go through the torture test (0.3 seconds vs >0.001 seconds total illumination time).  Interestingly a single element died at some point, but it still conducted current and light output was barely affected.

    Read more »

  • Voltage-Exposure Testing

    Tyler Gerritsen05/25/2019 at 19:33 0 comments

    Shortly after assembling the device, I tested a bank (3x LED's) by increasing the voltage until the LED's were damaged.  The test began at 45 V and incremented by 10 V following a single flash.  Major damage occurred at 125 V.

    More recently, I started a series of longevity tests with a single LED.  It was strobed over 100,000 times at various voltages.  Surprisingly, the LED reached 125 V and survived a battery of cycles at that voltage.

    Some head-scratching later and I realized that the FET driver was changed since the original destructive testing (TC4420 to TC4452).  The new driver could ground the MOSFET gates faster, shortening the pulse and probably preventing the LED from premature death.

    Read more »

  • Future Improvements Roadmap

    Tyler Gerritsen05/25/2019 at 03:47 0 comments

      Edgerton has been working well, but there is still room for improvement.  Some changes will require testing to implement properly, while others are going to be fairly straightforward to implement.  Here's a list of tests and improvements I intend to complete (ordered by priority).

      1. Lower Supply Voltage.  Using 8x AA batteries is very impractical, especially since the flash doesn't really need much power (LED's are more efficient than a xenon flashtube after all).  Additionally, the batteries are directly supplying the FET drivers, and as such the gate voltage is affected by the battery level.  Finally, the case size could be decreased by using fewer batteries.  I intend to accomplish this by adding a 12-volt boost converter and experimenting with a more efficient high-voltage boost converter.
      2. Better Voltage-Current-Luminous Power testing.  I've done some testing using a DSLR to measure light levels and the controller's on board ADC to calculate voltage current.  Better measurements can be taken with an oscilloscope and a light meter with an integrating sphere.  These measurements would be useful for selecting a drive voltage that maximizes lifespan and light output.
    Read more »

View all 6 project logs

  • 1
    Source All Required Components

    Download the Bill Of Materials from the github repository at https://github.com/td0g/high_speed_flash/blob/master/BOM%201.1.csv

    Purchase all the hardware listed.  Components purchased from eBay / Alibaba / Bangood require several weeks shipping time.  You may want to get started on assembly before all components have arrived.

  • 2
    Prepare the Tools

    Be sure to have the following tools ready:

    • 3D printer with 200mm x 200mm bed
    • Soldering iron & solder
    • Small side cutters
    • Hot glue gun
    • Allen wrenches
    • Dupont crimper kit (optional but recommended)
  • 3
    Follow the Steps in the Edgerton Assembly Manual

    Download the Edgerton Assembly Manual from the Github repository at https://github.com/td0g/high_speed_flash/blob/master/Assembly%20Manual%201.1.pdf

    Follow all the steps outlined in the manual.

View all 3 instructions

Enjoy this project?

Share

Discussions

Tom Nardi wrote 06/05/2019 at 06:14 point

Such an awesome project, glad to see you decided to enter it into the Hackaday Prize this year.

Very interested in seeing the Mark II, and the design changes made in regards to getting ready for larger scale production. I really like your approach there, having a "Classic" version that people can assemble on their own, and then evolve that into a commercial addition intended for the everyday user.

  Are you sure? yes | no

Tyler Gerritsen wrote 06/05/2019 at 14:11 point

Thanks Tom, I appreciate the encouragement!  I really don't have experience with DFM, so starting with a working design seems like a good way to get going.  I'm not even sure if it will even enter production since the market is so small.  Cheers!

  Are you sure? yes | no

Alan Green wrote 05/25/2019 at 21:41 point

Thank you for the very complete write up. I learned a lot from the testing and fault finding. The  MOSFET driver's sub-microsecond asymmetry between on and off times must have been fun for you to find.

  Are you sure? yes | no

Tyler Gerritsen wrote 06/04/2019 at 22:52 point

Thank you Alan!  Yes, I have a fairly rudimentary understanding how MOSFETs behave, and using an oscilloscope on them is always a learning experience.  I've recently applied some newly learned techniques to controlling the gate ringing (ferrite beads and different resistors), the signal has improved quite a bit!  Cheers, Tyler

  Are you sure? yes | no

Michael G wrote 05/25/2019 at 06:30 point

Might be worth considering adding an adjustment to change the brightness with a potentiometer or something similar (just an idea -- not my area of expertise by any means). Very cool project!

  Are you sure? yes | no

Tyler Gerritsen wrote 05/25/2019 at 13:45 point

Thanks for your interest and the like Michael!  The potentiometer idea intuitively makes sense, but for a high-speed flash it's actually counter-productive.  That's because the user ALWAYS wants the flash to be as short as possible, so Edgerton keeps the LED's at max power and just changes the amount of time they're on (too much light? turn down to 1/2 microsecond).  Cheers!

  Are you sure? yes | no

wim wrote 05/22/2019 at 09:36 point

Hi Tyler,

What an inspiring project this is! I was thinking along these lines a couple of months ago (triggered by the velo one indeed), but deemed it's design beyond my electronics skills. With this documentation I think I will give it a try, after the summer. A few questions:

- have you tried to approximate its guide number to get us an idea of the output power?

- would it be easy to change the design to six leds instead of 12 (for use in close-up photography)?

- can the leds also run continuously at low power, to serve as modeling lights?

thanks & good luck,

Wim

  Are you sure? yes | no

Tyler Gerritsen wrote 05/22/2019 at 14:02 point

Thanks for liking and following Wim!  Those are good questions, I'll answer best I can.

-Not yet, I will test that and update when the current endurance tests are complete.  The LED's have a wide viewing angle (110 degrees) so it will have a very low GN compared to its actual energy output.  Lenses or reflectors would improve the GN and I'm looking into adding those.

-Very easy!  The case needs to be redesigned, but no other major modifications needed.

-That's a very good idea, a couple complications that need to be addressed first.  Currently the LED's don't have heatsinks, so they will overheat quickly if run continuously.  Also it would need an adjustable boost converter to switch between high-voltage and low-voltage.

I'm going to keep the modeling light idea in mind :)  Cheers!

Tyler

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates