8 hours RGB Blend,
6. 4 mA = 51. 2 mA
8 hours Color Organ,
5. 75 mA = 46 mA
8 hours Sleep,
1.2 mA = 9. 6 mA
Total current consumption
per day = 106.8 mA
CR2032 typical current capacity
(down to 2V cell voltage) = 225 mAh
Estimated battery life = 225 mAh /
106.8 mA = 2.1 days
Real-world use seemed very close to this estimation.
Those who registered earlier in the weekend would notice
the blue diode on their RGB LED getting dimmer and, as
the battery voltage fell well below the diode’s forward
voltage, failing to illuminate at all. Green would go next,
followed by red. The microcontroller and microphone could
operate down to 1.8V and 1.5V, respectively, so they’d
remain alive and well (but somewhat useless without the
badge’s visual indicator).
Development Timeline
to rewrite large chunks of code to squash them. By the end
of April, the firmware was finally completed and
programmed into the microcontrollers.
By May, I started shipping all the components to e-Teknet — our manufacturer in China. We were so far ahead
of previous years’ schedules that the organizers and I
thought “There’s no way anything can go wrong. We have
so much time.” The month of June was spent waiting for
parts to arrive in China and pass through Customs.
Just when I thought we were home free, I noticed that
a single box I shipped on May 21 had been sitting in
Chinese Customs for two weeks with no progress. As
Murphy’s Law would dictate, this box contained all of the
programmed microcontrollers and custom-built LEDs —
both integral to the design. We still had a few weeks to take
corrective action and, while on a trip to the Middle East
with little Internet access and poor cell phone reception, I
placed orders for new components just in case the original
box didn’t get released in time. An added wrinkle was that
— due to manufacturing leadtimes — I could only order
partial quantities of components which meant fewer people
would get badges if we had to use this option. Not good.
After daily communication with UPS and e-Teknet,
attempts to contact Chinese Customs, and nervous,
sleepless nights, the package in question was finally
released without explanation and delivered to the factory
on July 22, only nine days before the start of DEFCON (last
year’s packages that were stuck in Customs were delivered
to the factory four days before the start of DEFCON). e-Teknet scrambled into action and, each day, they sent a few
thousand pieces directly to the conference venue as they
came off the assembly line.
Despite our early start and extra precautions, the final
batch of badges arrived only one day earlier than last year
(on Saturday instead of Sunday). Long lines formed in front
of the DEFCON registration area for people to swap out
their temporary plastic badge for the real deal. Attendees
missed talks and opportunities to hang out with friends in
order to stand in line for hours. Once again, the badge
delay was a topic of conversation throughout the weekend.
I certainly don’t want this becoming the norm. Having run
into problems two years in a row and knowing our luck will
eventually run out, we are considering having our PCBs
fabricated by e-Teknet in China, but then assemble the
badges in the United States. The approx. 300% increase in
Prototype hardware using the MC56F8006DEMO reference board.
July 2010 49
After the too-close-for-comfort, last minute arrival of
DEFCON 16 badges to the conference last year, I vowed to
start this project much earlier and to add enough cushion
into the engineering schedule to make sure the badges
were completed well ahead of time. Also, my son,
Benjamin, had recently been born, adding unfamiliar, new
parent challenges — like extreme sleep deprivation, non-maskable-interrupting diaper changes, and lack of quiet
hours dedicated to work. The project officially got
underway in December 2008. The initial design and parts
selection were completed relatively quickly with few issues.
Because of the parts sourcing and supply chain problems I
encountered last year, I took extra steps to ensure that the
components I selected for the design were immediately
available or had a short leadtime that I could work directly
with the manufacturer to manage.
By January 2009, I already had prototype hardware
completed using a Fresscale MC56F8006DEMO board with
the additional LED and microphone components soldered
onto it. At this point, even though the final firmware wasn’t
close to being done, I had written enough low-level code to
prove out the basic constructs and the prototype hardware
verified that my physical interconnects were correct. Since
the hardware design was simple and straightforward, I
decided to jump directly to the true-to-form circuit board
design without any interim layouts. The PCB design was
completed in February.
In March, I ordered all the production quantities of
components and focused on finishing the firmware. The
firmware design was a bit tricky, as I thought I had
functionality working great. But, when I demonstrated the
badge to the DEFCON organizers in person, they
discovered a few critical bugs. Much to my chagrin, I had