I've been designing/building a project for a non-profit I volunteer with, and a couple other similar agencies have expressed interest in the project. I'm currently prototyping on an Arduino UNO R4 WiFi.
It's a fairly simple project employing with a DC H-Bridge motor driver and a 2 switch relays. My question is, at what quantity should I start thinking about moving away from the R4 and to a custom PCB? I know the schematics are readily available so I wouldn't have to reinvent the wheel, but I realize it's not a trivial task either.
One consideration is that I am doing this pro bono for a 501(c)(3). I realize this could turn into a multi-billion-dollar project (lmao) but for now, it's all out of pocket for me.
"If you're planning on making more than X units, design a custom PCB."
What are your thoughts on the value of "X"?
That doesn't mean you need R4 wifi for motor driver and relays.
Maybe some inexpensive Esp module could be next step in development (if wireless needed).
Thanks Adrian, that's helpful. I thought (was hoping lol) the number would have been a bit higher. I know I have at least five agencies (non-profit animal rescues) interested already.
I have designed a few simple PCBs (atTiny85 based) using KiCAD so I am familiar with that, but I've never used EasyEda.
kmin, this is an IoT project, so I've been playing around with the ESP32 and NANO IoT 33 along with the UNO R4 WiFi. I really haven't decided on a base for the project yet. Actually, I've been using the UNO R3 but have a couple R4s coming in a day or two.
I had the project running on the Arduino IoT cloud on the ESP32 but can't figure out how to use the Arduino cloud in a multi-agency scenario. I moved the project to Blynk, but somehow (I know it was my error) used up all 30,000 messages in a couple hours and was locked out of Blynk until I fork over $100. That's just not in my "proof of concept" budget.
Jim, I've been watching a YouTube video titled, "Design a Custom Arduino Board for Mass Production in 3 Hours". I'm hoping he's got it figured out. Plus, he has downloadable files for KiCad which I am somewhat familiar with.
I like to use an Arduino Nano Every with headers. It plugs into two rows of sockets on the PCB. That way I don't have to recreate the wheel for the Arduino circuitry. It's cheaper and simpler than a shield for the Uno.
I think that really depends on how familiar you are on designing and building PCBs. If you ordering small quantities of parts, you need to deal with MOQ, shipping etc.
An intermediate is to order assembly through JLC or PCBWay who will do surface mount assembly for small Q's. Of course you can use KiCad to create Gerbers.
Either way I would design a PCB around a CPU module, and add interfaces to other parts. PCBs can be made so cheaply you can just use them as a mounting baseboard, front panels etc.
I make PCB's when I think I will make a few more of the same circuit. If possible I design them with SMD components that are big enough to be soldered by hand. That way the PCB has proven itself already when I get additional demand and may even improve the design.
And if I make a mistake in the design I prefer to toss away 5 pieces instead of a lot more.
If the alternative is to use existing "turnkey" hardware (e.g., an Arduino board and a shield) then X will be higher than if the alternative is labor intensive creation of complex circuits on stripboard.
And of course, regardless of economic considerations, you might feel that creating your own PCB is worth doing simply for the educational or entertainment value alone.
Good.
But it doesn't mean you should try to build R4 wifi on custom pcb.
There's difference between proto board, dev board and plain Esp module.
Count what you need and find the minimum resources for that.
I try to avoid making PC boards. Sure they're dirt cheap now, but my time is worth something to me. Especially if I don't get it right the first time.
So, the first choice, if I have to make more than 1 of something, is to head to google. It's a big world and there's a very high chance that someone else had exactly the same problem I have (or close enough) and chose to commercialize it. I'd say that 8 times out of 10 I can find a fairly cheap off the shelf arduino-based board that I can use as is.
And there are things like Qwiic or Grove where you can connect various board together very easily with minimal wiring.
As a last resort, I'll make a PC board if I have the time, but it's telling that it's been about 3-4 years since I last had to do it.
EasyEda is part of the "JLC network", i.e. owned by JiaLiChuang (Hong Kong) Co. Ltd. They've got a nice "one stop shop" with PCB, assembly, components (LCSC) and EDA software.
Although the EasyEda website claims they looked for a free multi-platform EDA software and didn't find one (KiCad?), I suspect the main reason for EasyEda is vertical integration with other JLC companies. But maybe JLC bought them out, I don't know.
I might try a PCB through this route, I know KiCad quite well but there is a lot of knife and forking required to get the files into JLC in the right format, with part nos for components from LCSC etc.
