Moving from Development Board to Custom Hardware

[czuvich]

Hello,

I have developed a prototype that uses the nRF52840 Adafruit Feather Express, a few sensors, and some other components.  I am somewhat new to hardware engineering; however, I was able to get some schematics designed and I have a custom PCB that uses the Feather Express.  I currently use the Arduino IDE to do development, and I'd like to stay there.

Having said that, I am trying to reduce some of my costs in my prototype.  Purchasing pre-made dev kits and components is very easy to get prototyping, but when I go to Mouser.com and see the raw components the price difference is fairly significant.  For instance, the raw nRF52840 is $7 whereas the Adafruit Express is $25.  I have been using my prototype for several weeks, and I've got some interest from other people.  What are the next steps to go from dev kits to a more targeted/customized PCB?  I do all of the soldering myself; however, I was looking at the raw chip, and I'm not really sure how to solder something like that (maybe the USB Dongle is a better choice).  

I know this post is long-winded.  I've attached a picture of one of my prototypes to get an idea of where I'm at.

[RK]

You aren't likely to be able to solder the raw chip by hand, unless you're very good. The older QFNs were fairly tractable, but the newer chips have completely hidden pads, they're pretty much LGAs. It's possible to do them with a hot air gun if you have lots of practice but it's hard. So you'd probably need at least a homemade reflow oven, there's tons of information out there about how to turn a toaster oven into one. So a better bet may be to buy a module which has the chip and crystal and antenna already on it and has certification, if that matters. You can buy the Rigado one for about $10. However that too has pads underneath, although much more spaced and so you'll still need hot air or an oven, those are a lot easier to do than the raw chips however as the pads are pretty well-spaced. 

Your next problem is that adafruit has done the work to make the whole thing Arduino compatible, I guess with custom firmware and I don't know what else. So if you want to continue with the Arduino IDE you'll have to work out if that firmware is something you can download and use and put on your own custom board, and whether you can do that legally. Personally I'd run as far away from Arduino as possible but I see you want to stay with it, so you may need to ask Adafruit what they'll make available to you; they tend to be fairly 'share-y'. 

Good luck - looks like a fun project. Are you intending to go commercial with this or just make it available as a hobby project? If commercial you have some vile licensing fees in your future from the Bluetooth SIG which may ruin your whole day. 

[czuvich]

Thank you for the quick response and the great feedback!  If I need to move away from Arduino IDE, then I'm definitely not opposed to it if that's what makes sense.  I think Arduino IDE was great for the prototype, but a targeted product sounds like needs a better IDE/environment (which was my hunch).

I would love to go commercial with my product if I can get enough interest, but I wasn't aware of the licensing fees.  Do you mind elaborating on that? (or pointing to a link).  I would hate to abandon BLE all together due to that, but BLE is such low-power and great range that it's a great fit!  Again, thank you so much!

[RK]

There's threads on here occasionally about licensing but I don't know if I've ever seen a complete answer and I don't know all the intricate details. It used to be (5 years ago?) that if you had a certified module or even a design which was certified and you followed it, the cost of registering your product was somewhere between low and free, but the Bluetooth SIG changed the rules so every product has to be registered and you have to pay. Personally I think this misses the whole original point of BT which was to be a technology with low barriers to adoption and which encouraged 'man in garage' to build something but the SIG clearly decided they weren't getting enough money in and made it more costly. 

So the cost of registering is something like $8000 if you're a full member, 1/2 that if you're an associate member and they have an exemption for entities creating their first one or two products which have low revenues cutting it to $2500. 

Purchasing a pre-made module will help here because they already have hardware registration with the SIG (which is another fee) plus testing reports and other certifications. You probably still have to do your own testing on the finished product however at least you know the module itself was designed to be compliant. I suggested Rigado because I've had some success with their modules and they're nice people, there are many others, Nordic has a list of them on their site somewhere. I wouldn't suggest buying the cheap ones from Aliexpress, they are likely to be poorly assembled, have no certifications and may well have terrible RF performance which would make a final product un-certifiable. 

I assume any product would need to pass FCC certification when built and possibly UL or CE as well depending on where you want to sell it. I literally have no idea what those tests cost but I've always guessed a few thousand dollars. 

You might want to post a separate question asking "If I was to build a product from scratch today, what would all the licensing and approvals cost" because I'd love to know the answer myself. My complete handwaving guess would be 5-10K but I could be wrong in either direction by an order of magnitude. 

[czuvich]

Thanks for that information!  I'll definitely post on here about licensing.