Before I cover early reflow oven results, I want to outline a few of the accessories I've fabricated that will help assemble individual Neurons into functional networks. Pretty basic stuff; lots of mindless mass production here. I'm getting better with the Engineer PA-09, especially given that the ribbon cable I'm using is a bit oversized for the connectors (the terminals require a bit of encouragement to snap together).
Okay, I couldn't think of a better term for these. Membrane potential modifiers? These are the simplest accessory--just a connector with a loop of wire shorting VCC to Signal. Provided a Neuron is powered through one of the other six ports, these are used to increase (excitatory) or decrease (inhibitory) the resting membrane potential of the device. You'll note I snipped the locking tab off the connectors--I do this with all of plugs to make 'em easier to quickly disconnect. Friction seems to hold them together securely enough for my purposes.
Identical to Exciters, but they use a momentary switch to short VCC to Signal. Allows one to locally modify Neuron potential on the fly--ideal for injecting signals and starting cycles. I had a few kinds of switches lying around so I made two types; the snap action version has a ~3" lead, while the tiny pushbutton units are just a bit bigger than Exciters and fit nicely on a crowded Neuron.
AxonsThese are designed to chain Neurons together, carrying both signals and power; as such, they use all three connections (VCC, Signal, Gnd). Various lengths allow for some network flexibility. Fun fact--we made a bunch of these with half of the connectors on backwards, just like a crossover Ethernet cable. That made for some cursing and a few rework sessions. Fortunately, getting Gnd and Signal reversed just makes stuff not work--no magic smoke.
Neurons often link to multiple downstream devices. In an effort to save board real estate, we only equipped each one with a single output connector; these adapters allow one to use Axons to connect to three downstream Neurons. We couldn't do this on the Dendrite (input) side since we wanted to be able to use Exciters; also, I was afraid our software debounce routine would cause us to miss fast sequences of signals. Hence, six inputs.
I made a few versions, some of which hook in to my bench supply. This one is USB powered (I've got a big hub on my bench, along with a bunch of wall warts), made out of an old ribbon cable, and features a current measurement tap. Highly useful.
Covered earlier but deserves another mention here. This is the most important accessory--without it I can't burn Neuron firmware. Built from a few bits of wire, three plugs, a clippy-do (for the Reset loop), and an Adafruit ISP breakout board I had lying around. Generally spends most of its life attached to my USBtinyISP.