Voting resources, early voting, and poll worker information - VOTE. ... Adafruit is open and shipping.

Production vs Dev boards
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

Production vs Dev boards

by donniekerr on Wed Aug 12, 2020 3:34 pm

I'm new to Adafruit and I'm considering buying some products. I'm building a Lora device that would probably use the Feather lora module. In my production product design, I need the actual sensors, like a light sensor, to be away from the actual board and exposed to the outside world to detect the light. It doesn't makes sense in my production product to have the tiny light sensor embedded into the board, like you would if you were doing it as a hobby or dev kit. I need the light sensor at the end of a long wire that connects to the board on the other end.

I've never built a production device before, so my general question is do you use these dev like board in production? or do you buy a different version of the device for use in production? Or do you just pay Adafruit to customize the board for you?

When I've successfully tested my sensor operations with the Lora module in dev, what is the process to build a production version of it? Hope this makes sense. I'm new to manufacturing side of the IoT business and it confuses me that all the demo are people playing with these products like they are toys or just tests.

I'd love some advice. Thanks!

Posts: 2
Joined: Wed Aug 12, 2020 3:10 pm

Re: Production vs Dev boards

by adafruit_support_mike on Thu Aug 13, 2020 2:23 am

The whole subject is pretty open-ended, really.

At one end of the scale, a tightly integrated production board is a stripped-down version of the hardware used for development. It doesn't have any parts that aren't necessary to make the production device work, the PCB is designed to fit any physical requirements, parts in the Bill Of Materials (BOM) are optimized for size, cost, or whatever is most important to the project, firmware is tweaked to get the best use from features the device needs, etc.

At the other end of the scale, development boards are built for convenience on the workbench. If you use three sensor breakouts that communicate using the I2C protocol, each board will have its own voltage regulator, pull-up resistors, level-shifting circuits so you can connect them to 3.3V or 5V, and so on. Connecting them to a microcontroller development board is just a matter of plugging in some jumpers. The microcontroller development boards will have their own set of convenience features: USB connections and a bootloader so you can program them from an IDE, another voltage regulator, maybe some signal level shifting, connections to GPIO and protocol pins routed to the side of the board for easy access, etc.

A whole lot of production devices live between those two poles though. We've sold a lot of boards to groups working to build medical devices over the last few months. To them, something like a Feather microcontroller dev board is a component.. something that plugs into a slot and handles the user interface on a display or something.

donniekerr wrote:Or do you just pay Adafruit to customize the board for you?

We don't do any custom fabrication or modification, but our designs are Open Hardware and the design docs are published on Github for anyone who wants to modify or build their own version of a particular board.

Doing your own production is a good call for anyone who needs to control things like provisioning and timing. If you need X-many boards every month, and not getting a batch would cause problems for you, we openly suggest you take our design files to an assembly house and coordinate your schedule with them. That way you can manage your own inventory of key components, get better information about potential timing problems, etc.

donniekerr wrote:When I've successfully tested my sensor operations with the Lora module in dev, what is the process to build a production version of it?

You have the first two steps right: build a working version with dev boards, then adapt that design to your own project.

At minimum, you'll know that the circuit design and components on the dev board do what you want. Even that much is tremendously helpful, since it saves you from having to consider twenty other ways of doing the same thing. It also gives you a performance baseline that you can use to decide whether some other device will do the same thing in a way that fits your project requirements better.

The rest of it is basic design: weighing options and making tradeoffs to fit the competing requirements of your spec. There will always be something smaller, something that does X better, and something cheaper. In all probability, they'll be three different things, each being really good in one category by being bad in the other two.

Physical size, cost of materials and construction, and energy budget are probably the three most common constraints people think about during design. In practice, maintenance and support will probably consume 60% to 80% of your total investment over time. Rework is a huge time-and-money sink, so designing to prevent or minimize it is almost always cost effective, even when it demands extra components or space you'd rather not spend.

Once you have a working system with all the features you want, make a list of things you consider design priorities for the final device. For each priority, do a "to heck with everything else" design that you consider the ideal optimization of that single item. Then start making compromises to bring all of those designs and the working prototype to a point where they can coexist.

Posts: 61422
Joined: Thu Feb 11, 2010 2:51 pm

Re: Production vs Dev boards

by donniekerr on Thu Aug 13, 2020 10:40 am

Thank you! Very helpful. I like the idea of a hybrid between the two in our greenhouse use case. It will be very important to be able to easily replace a failed sensor that in connected to the Feather Lora Module, which seems fine for production.

I appreciate the help! I'll be purchasing some parts soon!
I want the feature lora module for sure and a few sensors to test with. Have any recommendations on an ambient temperature sensor, a light sensor, and a soil sensor that does moisture and EC? I prefer a temp and light ones that could be on the end of wire (what do you call that kind?) vs embedded. I want to run them up a 2-4ft pole inside the hemp plant pot. The soil sensor will stick into the soil and hold up the pole.


Posts: 2
Joined: Wed Aug 12, 2020 3:10 pm

Please be positive and constructive with your questions and comments.