We’ve come a long way with Faraday and it’s evolution has been fun to see. The concept of Faraday began at the Rochester Institute of Technology in Rochester, NY in 2010 while Brent and I were Electrical Engineering students. Deeply invested in K2GXT, the RIT amateur radio club, we were unsatisfied with what amateur radio technology was shaping up to be. Digital “modes” tend to focus on voice activities with data as a filler for unused bandwidth. Emergency communications groups enjoyed setting up repurposed Wi-Fi routers for high speed networks. This is fine but they shouldn’t be the only real options regarding the advancement in digital ham radio technology. There was no way to learn the basics of digital communications and subsequently have a digital medium to experiment with.
Fast forward to 2012 while interning at SpaceX, Brent and I purchased the Texas Instruments EM430F6137RF900 development kit to start learning about digital communications in our spare time (very little!). It wasn’t until 2014 Brent designed a simple breakout board for the EM430 boards to start serious development. Hindered by the responsibility to help AMSAT design and build a Maximum Power Point Tracker I didn’t put much time in helping with Faraday until the end of 2015. I did manage to design the hardware for Revision 1 but not much else. It was only after the MPPT when both Brent and I focused on the project together did we manage to cover the huge amount of ground necessary to be where we are today.
Brent and I appreciate everyone who supports Faraday in one way or another but this first group of 25 developers willing to purchase the first batch of hardware deserve some notice. They are investing in the idea of Faraday without the software being mature, without extensive field testing of hardware, and without an infrastructure. We’re doing our best to include additional hardware for them such as free antennas but in the end they simply share the belief that they can help improve ham radio and it’s worth it. It’s hard to describe the joy it brings us to know others believe in our mission. We’re committed to seeing this through. Thank you.
Faraday Revision 0
Revision 0 is where it all started. The Texas Instruments EM430F6137RF900 with a custom breakout board for UART and IO expansion was all that Faraday was. The entire goal of Revision 0 was to give Brent the ability to learn how to use the CC430 and interface with USB serial data using Python. The breakout board provided a linear regulator for power and an FTDI USB to UART IC with some LEDs.
The EM430 kits were expensive at about $299 for the pair back in 2011/2012 which only provide 16mW output power, no USB interfacing, voltage regulators, or peripherals. These boards let us start learning to work with the CC430 and wireless fundamentals without having to first design and build working RF hardware. Our goal with Revision 0 was to figure out if this was even a feasible project, not if we could build the hardware. We knew we could to do that.
Faraday Revision 1
Venturing into OSHPark, we designed a four layer board in 2014 to consolidate our efforts with the Rev 0 boards and provide several peripherals such as an on-board uBlox NEO-6M GPS. Low volume Ublox modules are expensive unless you did what we did and sourced them from Amazon in non-ESD safe packaging. We immediately decided this was not a good idea since we would not be sourcing large volumes of GPS modules anytime soon. Revision 1 Faraday hardware was also Arduino™ Uno R3 shield compatible. At least we thought so at the time.
While Revision 1 helped us make a large amount of progress it also had a huge number of oversights. Most notably the 3.3V switching regulator used an unshielded inductor in its buck converter stage. This was unintentional and had an “oh crap” associated with it when we opened the bag of inductors. As expected, an unshielded inductor switching high dV/dt signals directly over ADC input traces caused noise on the ADC channels. We also forgot to place a VCore capacitor for the CC430 which caused the microcontroller to reboot/corrupt every time we transmitted! That’s the blue capacitor dead-bugged over the IC below. Lastly, while Arduino™ shields could plug into Faraday the voltage level for all IO was still 3.3V making at best most shields not work and at worst completely destroying the CC430 when a shield was plugged in. Even with these errors, we were able to identify, fix, and workaround almost all of them to keep pushing forward with our code.
Faraday Revision A
The workhorse of Faraday has been Rev A. Originally intended to be our first production board it allowed us to start the hardware simplification process. This board solved many of the design issues of Revision 1 such as the noisy power supply and missing VCore capacitor for the CC430. We also changed the GPS module to the Antennova M10478-A2 able to be sourced from Digi-Key. Nearly a full year of coding, learning, and head-scratching was spent working with the three revision A boards we built by hand-placing all components and reflow soldering in our Black & Decker toaster oven. Often, Brent and I would switch the extra board between each other to make a pair depending on which one of us was developing code which required the radio functionality.
This board also had a 5V and 3.3V regulator with ideal diode switching between all power sources and USB power. Faraday Rev A also level converted IO between 3.3V and 5V using TXB0108 bidirectional level converters. This allowed actual Arduino™ shield compatibility. Finally, In August of 2016 one of the three Faraday Rev A boards flew above California and is now somewhere in the Sequoia National Park, maybe Brent and I will find it on one of our many backpacking trips we often take into the park!
Faraday Revision B
Pushing for production in late 2016 the Revision B boards had one mission, simplify the hardware. We wanted to make three revision B boards to prove some large design changes before committing to production quantities. The original power supply buck converter IC’s used on Revision A had reached their end of life which required a complete redesign as one example of a massive change from Revision A. We also moved towards an integrated passive balun which not only was smaller and had less manufacturing variance but helped reduce BOM line items. Less line items means the board is easier to manufacture.
Faraday can still plug into Arduino™ shields as the top board in the stack since the SMA connector obstructs any stacking above. This as semi-intentional to discourage plugging in non-Faraday shields since damage could result. The power connectors and GPS antennas by design do not interfere with normal Arduino™ shields. We removed the 3.3V/5V voltage conversion for GPIOs as in a year of Revision A testing we had not had a good reason to actually use this functionality. Faraday provides a ton of functionality on-board already.
Faraday Revision D1
Our first production board is Faraday Revision D1. While All previous four layer revisions had been manufactured with OSHpark, we decided to work with PCB:NG on this one. They handled fabricating the boards as well as assembling them for us. Building more than 50 boards for the first time is a small test run for established companies, for us it was a huge amount of capitol and risk. Manufacturing Revision D1 boards was a personal achievement for Brent and I.
You hopefully noticed there was no Revision C Faraday. In fact there was, it was simply never built. Rev D1 is a Rev C board with all the features we never really used either removed or not populated. It helped us achieve our price point per board as well as improve manufacturability with fewer components to place. The numeric value in the revision is our way of indicating manufacturers. Since we are trying to keep costs down we want to use the standard PCB stackup from a vendor. This forces us to change the design slightly for every vendor who has a different four layer PCB stackup. D1 means Faraday is designed for manufacturing with PCB:NG, we’d have to tweak a few items to move vendors correctly. Easy to do but you definitely do not want to mix those gerbers! Take note if you use our open hardware designs 🙂
Some of the major improvements over Rev B were the redesign of the power supply and GPS to remove unnecessary components as well as pulling back the ground from RF traces to improve performance of the transmission lines. We also added support for direct radio access via the IO pins for use with an external computing platform such as a Raspberry Pi in the future if we decide to implement that.