brainstorm

I didn't come into the course with a concept in mind, so this is somewhat an exercise in backreasoning what will fit into its constraints.

I lean towards something supporting embedded development, as I experience a lot of pain points that I feel are unnecessary. Considerations:

• Every embedded bringup inevitably results in a huge tangle of wires that consumes my desk
  • Mutually exclusive with other projects
  • Bringup burden + switching cost
• Wired connections to power and instrumentation make debugging fragile re: physical disturbances
• I always want an SWD programmer
• I always want a bus + serial sniffer (SPI, I2C, UART mostly)
• I want workholding for my board
• I want to probe signals on my board

The full, simultatneous realization of these desires in a single product would be a weighted PCB clamp with attached PCBite goosenecks, integrated scope, SWD programmer, and logic analyzer, running on battery power, all of which you can connect to over WiFi. Clearly, this would be overdesigned.

However, I think parts of this could make sense. A WiFi-enabled bus sniffer (à la BusPirate) or SWD programmer would be great. The idea is that you would plug this into a device you're bringing up with short leads and be able to eliminate a lot of the wires between the device and its power/instrumentation. Ideally, you'd be able to move the DUT and the sniffer/debug probe around as a single unit.

Possible features:

• Host-side client libraries

  • As opposed to BusPirate, which has a custom text protocol

• Host-side GUI application

  • Configuration and linear log display are straightforward
  • Graphical timeline viewer is probably a lot of work

• Portable +3.3V and +5V for powering debugged project

  • Sniffer/probe would need a battery anyway
  • Does come with additional electrical design complexity
    • Fusing
    • RP, ESD protection
    • Increased requirements on regulators
    • Boost for +5V

• ADC lines for monitoring power nets

• Some way to physically couple the probe to a board / treat them as one unit (for the sake of desk-clearing / context-switching). Possibilities:

  • Housing integrates a clamp
  • Housing can easily be attached to a clamp
  • Design housing with intent to be VHBed
  • Design housing with intent to be ziptied (probably not this one)

what exists?

(Only considering wireless-enabled devices.)

The Flipper Zero is a potential extant alternative, but it doesn't have WiFi capabilities. It can speak Bluetooth and supply portable power, but I believe WiFi specifically is important for the ability to walk away and still access the device with decent link speed. I would also prefer a stripped-down, small, lower-cost option that's easier to work with.

SEGGER produces the JLink WiFi, which performs the debug probe function, but it's very expensive ($1280) for what it is. A proof of concept alternative would be a picoprobe made out of a Pico W with tweaked firmware and an SLA-printed enclosure. Accounting for engineer time, I'd estimate the cost at $1-2k. Seems on the same order of magnitude as the JLink board for single quantities. This implies that a significant part of the cost of the SEGGER product is software and intangibles (product support, stability, "it-just-works" factor), but I believe there's a meaningful segment of the market that would trade this off for lower cost.