I am helping out with development advice for a custom IoT sensor initiative slated for low volume production (<1k units) within a year if proven successful. I am trying to understand where the cost/benefit line lies for modular and rapid prototyping systems such as WisBlock < WisBlock module | Easy-to-build IoT solutions | Available now on RAK – RAKwireless Store > from RAKwireless, Mikroe Click Boards < Click Boards - MikroElektronika >, and the Qwic system from Sparkfun < https://www.sparkfun.com/qwiic > so that I can decide whether or not to recommend these as a great starting point.
I appreciate any feedback!
The deciding factor for me on this entire category is, does it get me closer to a fieldable/testable product. There are two aspects of that:
First, are the modules such that I can get them into a package that is a reasonable size for the application. I once tried to use Sparkfun (because they had a specific module), but the functions were broken up so discretely that I had a dozen boards to deal with in the end. The link above makes it look like they still use that concept.
Secondly, is the issue of getting me closer to product. If the boards do not come with commercial grade pcb files that I can lift in the end and layout onto a single board, ditching all the connectors AND/OR there are ambiguous licensing or ownership arrangements it is no good to me. I might as well start from eval kits and reference designs. They may cost me $300 a board in the beginning, but it well pays for itself going into production.
A couple technical points. The products closer to the science fair end of the category often have no temperature range, so field testing is a problem if is not a living room product. Secondly, I want the more complicated modules (uP, WiFi, BT…) to at least appear like they might pass part 15 or CE or even part 90 if laid out properly, without a lot of work.
Ublox does a great job in this space.
Just a few thoughts. Hope it helps.
All these boards mentioned can help you to get started relatively quickly and understand what exactly do you need down the road.
Would also add adafruit.com as a good starting point, however, to get much better efficiency and high level of customization you’ll eventually need to develop your own hardware and maintain your own copy of firmware package (toolchain, drivers, libraries and app logic, …).
Would suggest to start based on prototypes board in order to validate the idea, if possible, then develop your own board. This of course depends on the exact project needs and the target price.
One more point to note, is that sometimes customizing the prototype platform and the software framework to your needs, even for small stuff, may be not that straightforward, so before selecting any platform would suggest validating at least on “paper” if all the for planned functionality could be implemented.
I am 100% in line with the points raised by @Reuven_Man.
Sparkfun and Adafruit boards are a perfect starting point (this is how started as well), but do not fit to the production-ready hardware. The complexity, size and cost for building a custom board based on them will be eventually prohibitive for high-volume projects. For up to a few tens of units I’d say go with them, but for >100 I am not sure this is a good idea.
I also like the idea of rapid prototyping systems. Here is the process I have used over time that does not require too much skill / investment:
The metric here is not total cost, it is at each step, to reduce the cost to next thing learned:
- Prototype on a breadboard (using SMD to through hole adapters if needed). This is where Qwiic and Stemma play - but they are limited in IOT as they are only 4 wire and don’t include the all important interrupt line. I feel like they may have set the IOT market back by making it easy for hobbyists / enthusiasts to get to “Blinky” but not supporting the low-power use case that is at the heart of many IOT projects.
- Once I have the schematic, lay out a simple board and have it made at OSHPark.
- Once the board is validated, refine it with house parts at your favorite assembly house - mine is Macrofab (10-200 qty)
- If you need more, find a local contract manufacturer (ideally one within 1 hour of your home / office). If you get a good one, you will learn a lot in the exchange with them that you don’t get with the web-based tools like in step 3
- If you need more, then consider moving off-shore.
I sell business to business products and am still iterating so I have not got to step 5 yet, but would be interested in hearing from those who have.