Prototype to Production Help

[j5th]

Hey guys, I am looking for someone to help me make my Arduino built project into a more refined product. It is built using the ethernet shield and I have the basic proof of concept working but I don't know where to go from here. My intention is to get it working with WiFi and then make a nice web interface to control functions. If anyone has experience making a real product based on one of their Arduino projects please give me some advice on what my options are. Thanks!

[franklin97355]

My intention is to get it working with WiFi and then make a nice web interface to control functions.

I would say this is where you should go next. Don't worry about making a finished product until you have a finished design and working model.

[Alan Chatham]

Well, I think for one, I would look into starting to play around with the raw chips. Get yourself some sort of AVR programmer (or use the ISP programming ability of the Arduino) and a couple of straight-up ATmegas or ATtiny's and start converting your code to run on whatever chip makes the most sense to put your application on. Of course, you could always design your thing as a shield, and ignore that, but if you want to sell a whole product, having the $3 chip running it all cost you $30 since you have to buy it in Arduino form is potentially a poor business decision.

I would also be interested as to why you want to have the device WiFi enabled. I mean, there's a number of good reasons to have something controlled via a wireless network, and I have weird personal hang-ups about making things wireless. But, WiFi is kinda an expensive option, and making something that effortlessly connects to WiFi networks in customers' homes or businesses sounds like a lot of engineering. But I'm a big wireless sissy, so take that advice with a grain of salt.

[stinkbutt]

Alan's right. Next step after getting the system working is converting from Arduino to AVR. How much memory does your sketch take up? You might be able to downgrade to a cheaper AVR processor, down from an ATMega328 to an ATMega168 to an ATMega88, etc. etc. etc...

It probably won't take more memory if you compile in AVR-GCC, but you'll need to rewrite the code, going from Arduino to C.

[engblaze]

Alan's right. Next step after getting the system working is converting from Arduino to AVR. How much memory does your sketch take up? You might be able to downgrade to a cheaper AVR processor, down from an ATMega328 to an ATMega168 to an ATMega88, etc. etc. etc...

I would actually argue that this depends on how much experience you have with C programming, and where you see the project headed. Porting an Arduino sketch + associated libraries to C and mastering chip programming can be an intensive process for newcomers. It's a great thing to learn and probably necessary if you wanted to scale up. For now though, it seems like you should polish your whole implementation first, and see if it's something you're truly serious about.

Plus, the chips may be $3, but what about peripherals and necessary components? Unless your project is conducive to an extreme barebones setup, by the time you throw in necessary caps, voltage regulator, and whatever else into your circuit you're looking at more like $10 per unit in small quantities. Considering the Sparkfun Arduino Pro or Pro Mini are $19, that's not a huge gap until you start producing large numbers. We recently took a prototype to small scale production using Pro Mini's, and at 40-50 units it's not worth the extra design and testing effort to go to pure AVR yet.

[cstratton]

Alan's right. Next step after getting the system working is converting from Arduino to AVR.

It's should not usually be necessary to do such a conversion. You don't have to have an 'arduino' or even a full clone in order to run an arduino sketch on a compatible atmega part. There's very little on the arduino style boards unique to supporting the arduino style of code, vs. what is necessary to just make an atmega chip operate.

How much memory does your sketch take up? You might be able to downgrade to a cheaper AVR processor, down from an ATMega328 to an ATMega168 to an ATMega88, etc. etc. etc...

You might save a dollar immediately there, but though the real savings on smaller capacity parts may only come in on larger quantities (for something I'm doing attiny25/45 were actually going to cost me more than 85's, probably because they are less popular, though there is some savings going from atmega328 to atmega168). Actually, the first thing you can do is shop around - I've seen differences of over a dollar between major suppliers.

Then you can decide if you can eliminate things like the external clock or resonator, the usb-to-rs232 or level shifters, etc.

But for an ethernet application, and atmega is most likely not the best platform compared to something that has more of the ethernet functionality on chip. There are likely more cost effective combinations for wifi too.

[spacecowboy]

Some things I've learned trying to get a board as close to production-ready as possible

1) Form factor is important. Select your case and do your physical design before you even start laying out the board.
2) Surface mount, surface mount, and more surface mount. SMT passives are -very- inexpensive.
3) Find a book on PCB layout guidelines. Devour it.

Since you're using WIFI, or want to use WIFI in a finished product, I suggest that you shop around for an FCC-approved wireless Ethernet modules that are approved for modular use in 802.11 systems so you don't have to do testing to get part 15 approval (provided we're talking about an end-user product).

While there are complete chips like the MAX2830 that are a complete ISM transceiver, finding and adapting a module that does most of the heavy lifting for you will be the easiest route. Even laptop makers like to do this. For small quantities the development costs of rolling your own Wifi transceiver are going to be huge.

[bigmessowires]

Since you're using WIFI, or want to use WIFI in a finished product, I suggest that you shop around for an FCC-approved wireless Ethernet modules that are approved for modular use in 802.11 systems so you don't have to do testing to get part 15 approval (provided we're talking about an end-user product).

That raises a question I've often wondered about: do all the gadgets sold by Adafruit, Sparkfun, Dangerous Prototypes, etc need FCC approval? I asked Ian at DP about this, and he said none of their parts are FCC approved. Whether this is because they're exempt, or because everyone's just expecting the FCC to look the other way for low-volume products like these, I don't know.

I'm hoping to sell a product of my own soon that includes a 0.96" OLED screen. When I ordered the OLEDs from China, they were held by US customs, and I had to fill out some FCC paperwork saying they were exempt from FCC licensing requirements because they were engineering samples and not for re-sale in the USA. According to the guy at FedEx who helped me with the customs forms, if I eventually sell boards containing these OLED screens, I will need to get an FCC license. Anyone know if that's true?

[adafruit]

all depends on the product(s) whoever is making and/or selling, what/if it emits, etc. in the kit biz section there are some threads, links and resources about this.

cheers,
adafruit

[bigmessowires]

Funny, I never looked in that forum before, because I read its name as "Kibbitz" rather than "Kitbiz".

[spacecowboy]

That raises a question I've often wondered about: do all the gadgets sold by Adafruit, Sparkfun, Dangerous Prototypes, etc need FCC approval? I asked Ian at DP about this, and he said none of their parts are FCC approved. Whether this is because they're exempt, or because everyone's just expecting the FCC to look the other way for low-volume products like these, I don't know.

The issue of kits and the FCC comes up from time to time in Ham Radio circles. Certain Hams have gotten in trouble before with the FCC and have been issued fines for buying and selling assembled "kits."

Generally kits do not require equipment authorization to be marketed (sold) but they would still need to comply with the limits set forth in corresponding sections of part 15, hence one of the many reasons you'd get in trouble for selling a WaveBubble kit. The responsibility of compliance lies on the user, or the person who assembles the parts together, not on the manufacturer of the kit. The Hams that got in trouble for selling assembled kits got in trouble because by performing final assembly, they were manufacturing for sale and that's a different animal then selling kits that require some additional assembly.

As for your import concerns, I'm going to venture an assumption that the paperwork was a form 740 and that you checked box III, declaring that you won't offer for resale. You may want to ask the FCC because it sounds like the OLED displays qualify as unintentional radiators/sub assemblies sold to equipment manufacturer (you) for further fabrication.

An OLED display sounds like an unintentional radiator, and thus equipment authorization would be subject only to verification to $15.101 under "all other devices" meaning the manufacturer (you) self-approves via necessary tests to determine compliance. However since you're essentially performing further fabrication, note (1) indicates that no equipment authorization is necessary hence you could probably check box II on the customs declaration form (IANAL) if you wanted to incorporate the OLEDs into a Widget for sale.

FCC Part 15.101(1): No authorization is required for a peripheral device or a sub-assembly that is sold to an equipment manufacturer for further fabrication; that manufacturer is responsible for obtaining the necessary authorization prior to further marketing to a vendor or to a user. So it doesn't seem like you'd need an FCC ID to import the OLEDs.

[paul66055]

I have used WhizNets WiFi starter kit and it is really easy to get started with the kit.They provide all reference schematics and other related document. With the help of the starter kit you can easily develop your application on top of it.