How to validate a market with development boards and SD cards

Market forces alone didn't work. It's an externality, a cost paid by others not involved in the transaction. Market forces don't handle that. A 1970s Milton Bradley Big Trak and a Radio Shack TRS-80, both popular products in their day, will, if brought near to each other, both crash. Without fairly strict regulation of unwanted RF emissions, there would be many incompatible devices. There were before the FCC started requiring more testing in the 1970s. A world with a huge number of consumer devices emitting RF noise would have prevented low-power cellular phone and WiFi deployment.

It's not that hard. This is an "unintentional emitter" (it's not trying to send a radio signal). The rules for that are not too bad. Testing costs about $3000 to $5000.

You want to have some ability to pre-test. You might find something. Attaching a wire to something can give it an antenna and make it emit much more RF, so you do need to test. It's not too hard.[1] Actual FCC certification is $3000 to $5000, assuming you pre-tested and fixed any problems before getting a certification run.

From the project's FAQ:

"Given that this will initially be a niche product, the price will be quite high. I was once taught to ask myself the following question: Who is your rich customer? The type of person whom I have in mind has a high discretionary budget for personal electronics and willingness to pay a premium for novel ideas."

[1] https://www.nutsvolts.com/magazine/article/low-cost-emi-pre-...

I worked on a equipment project for a large restaurant chain about a decade ago. The core application and related assets/recipes/files were all on an SD card. When it was time to upgrade the app or release new seasonal recipes, every store got a new SD card in the mail with instructions to wait for a certain date, power down, swap cards, power back on, dispose of the old card.

It was way cheaper to send updates that way than bother with encryption, networking, corrupted disks, etc. A bricked machine lost a hundred dollars or more per hour. If the new card failed, the operator could continue with the old one until a replacement could be sent.

One major problem was suppliers always trying to swap to lower cost SD cards, even counterfeit ones (c.f. Bunnie), and things would go south really fast. The Linux system and hardware were both pretty old and had MMC stack issues when the cards showed shaky margins on the timing. Or, capacity wasn't what was advertised (c.f Bunnie). We had to spend a cycle or two qualifying each mailing release to make sure a shitty batch of cards didn't make its way into the stream.

SD has its uses, although I still prefer a read-only eMMC partition to hold the bootloader and O/S. I don't get why RPi users put themselves through such misery to save $20 on their SBC.

The various test labs I’ve worked with haven’t been “too busy” to respond. However, they are generally hesitant to work with people who don’t really know what they’re doing.

If you are an engineer with knowledge about the process and who needs a lab to partner with, it’s not hard to get in somewhere.

However, if you don’t have the knowledge or experience, the lab might sense that you’re looking for someone to hold your hand heavily through the process. They may be less than enthusiastic to take on a one-off customer who might require an abnormally high amount of communication and hand-holding when they can fill that same spot with a repeat customer who needs nothing more than to book the time at the lab and can show up prepared and ready to go.

I suggest teaming up with a local consultant for your first round. Not only will they help you through the process, they’ll have connections and reputation to get you into the labs.

The lab fees aren’t extraordinary high for a hardware startup, really. It’s not free, but it’s not much relative to the up front costs of building hardware inventory.

Here is a pretty bad scenario for apparently willful unintentional radiator violations, where an ultrasonic foot bath company was at best unorganized and slow to comply with testing and labeling requirements: https://docs.fcc.gov/public/attachments/FCC-18-67A1.pdf

Here is a better scenario, where an LED sign manufacturer was a bit more on the ball: https://docs.fcc.gov/public/attachments/DA-22-1136A1.pdf

Note in both cases there is no mention that these devices emitted RF above allowed limits for unintentional radiators. These companies simply didn't test and didn't label their devices appropriately.

Here's one for Asus, where they got WiFi products certified, and then changed something, probably firmware, that allowed those devices to transmit more power than allowed: https://docs.fcc.gov/public/attachments/FCC-24-69A1.pdf

[0] https://flyingcarcomputer.com/posts/a-new-personal-computer/

Market forces naturally determined this outcome though. If you're big companies you naturally want to limit the threat of new competition. Making compliance more costly achieves this.

The primary problem with mystery-parts is they tend to have issues with RoHS documentation, complex customs clearance requirements, and unknown specifications.

DIY evaluation kits people assemble do fall under a sort of gray area, but if your hardware does splatter the RF spectrum it is a $1m fine in the US, and a $5k fine + up to 5 years in jail in Canada.

Unshielded RAM, USB/PCI to Ethernet, and Video GPU chips will often just barely pass EMI testing under ideal circumstances. Cheap stuff from the mystery bins will usually just glean the FCC id off a refrigerator to get through customs.

Have a nice day, =3

Basically everything on their blog/faq ranges from inept to dangerously misleading.

And that for good reason. Any bad actor on the RF spectrum can be an actual, significant and direct threat to people's lives - particularly the EMS bands as well as the rail, marine and flight safety/coordination channels are absolutely vital. Up next is stuff like GPS, radio and television where disturbances affect a lot of people, and then there's local stuff like wifi, Bluetooth, Zigbee, door openers and whatnot that only affects very few people.

Unfortunately it is very, very easy to be a bad actor on the airwaves. Powerline/PLC is hated by radio amateurs for a reason, and that one is actually even licensed. The other stuff is much, much worse.

https://www.pollin.de/p/bausatz-led-wechselblinker-810051

German company selling a German-made electronics kits in Germany without CE certification. And they have lots of them:

https://www.pollin.de/bauelemente/bausaetze-module/bausaetze...

As long as you don't connect to mains power and you don't ship a finished product, you're exempt from CE certification. So use an USB plug as your power supply and sell it as DIY kit to be assembled by the customer and you're good to go.

This is across large and small companies, so I'm going to take a guess and say that in the AliExpress and Temu age, simply mailing the device from China will solve all of your problems.

Do you wonder why all of your new electronics are made in China? One big reason is that China has its own regulatory labs (which may or may not do testing - who knows?) that are literally at least an order of magnitude cheaper than getting certification done in the US or Europe.

I'm working on two client products now that I and the clients would prefer to have made here in the US, but both will be made in China because the companies literally cannot afford the rapacious cost of getting them certified here. (And China mfg is way cheaper, too - partly because of parts distribution models: It's literally cheaper to buy a finished product from China than to buy the components here to assemble the same product!)

We're using ESP32s, and are currently going through ethics approval, which, from what I understand, means we can use the device prior to sale, but maybe we've got that wrong. I can't imagine having to have each hardware iteration certified by the FCC.

What struck me more about this article is the subject of marketing.

For companies that are doing pre-sales, and are still in development, and likely haven't been certified yet. Isn't that considered marketing? How are other companies handling this? We're looking to run a marketing trial in a few months, and marketing is part of the recruitment process for a trial.

Parasite of the economy, right there

Disclaimer: Nothing below is meant as legally relevant compliance advice. This is just my opinion on the matter.

Going to snark:

"The testing and certification industry is odd"

Except, outside the software world, the real world, where there are real consequences, it's not really.

"The line about CES, in particular, made my hair stand up."

Why? Absolutely the unauthorised device at CES is should NOT be allowed. What if said device caused too much interference on cell phone frequencies and suddenly nobody at CES can dial the local emergency number?

If that made "your hair stand up", here's one from personal experience that will freeze your blood:

I worked as a teen for a certain electronics chain. Said chain was selling a wireless weather station imported from China. A government department that monitored the country for Earthquakes noticed that this device impinged on their frequencies. After the spectrum regulator confirmed the finding, a nice gentleman from them visited us a told us the following:

1) As of this moment this device can no longer be sold. Move it off the floor immediately (he stayed and made sure we did exactly that).

2) That we will immediately issue a recall of said device at your own cost and issue full refunds to the customers.

3) He will return when we decide on further enforcement action which may include punitive fines and recommendations for further remedial action you will need to undertake.

"In theory, it exists to serve the public good and uphold consumer protection laws."

Well here's a (very simplistic) tidbit for the author: In the US, part of the gestation and formation of standards bodies and testing was "market forces", not for the public good. It was to help protect companies from litigation. If you followed the standards, tested and certed to them, paid the fees etc you then had the standards entity bat for you in court (UL is short for Underwriters Laboratory. That name was not chosen for funsies).

"However, in reality, the labs are “too busy” to respond or reply very late and generally sound less than eager to work with you."

Well you don't sound like a serious customer. AND the Labs are not there to give you advice. They're there to do INDEPENDENT testing.

"Variations of the FCC exist in pretty much every developed economy. Putting a poorly tested hardware product on the market immediately puts a target on your back. Maybe you’ll get lucky, but chances are that someone somewhere will report you. And, unless you are operating entirely out of China, it will hurt. A lot. Both your company and maybe even you, personally."

As it should. The electromagnetic spectrum is a very precious and very limited commodity and IMO, the best regulated "commons" in human civilisation (though still not perfect). So no, you are not welcome to just urinate in it willy nilly with your hustler start up product.

"I did not want to spend so much money on testing before I validated the market or gathered a community of believers."

And there it is.

"This way, the electronic device liability will fall on the manufacturer, and the magic of friendship EULA should afford me enough protection to make this a pure software play."

No. That's not how this works.

1) I assume the author is from the US (as they speak about the FCC). I had a 30 second look at these dev boards and their instructions. There is no FCC conformity declarations or markings, so US customers can't use it.

2) It has CE and UKCA though, so customers from EU+UK (and some other countries) can buy them but the certs only cover the dev boards AS SOLD. (i.e without the authors software)

3) Author is modifying the product behavior with their software. So yes author. You are still liable. Technically, your customers are first in the line of fire. But the likely sequence of steps is: Friendly Spectrum Representative will visit them first, have a chat, ask them to stop using the device, then leave them a lone and then come for YOU.

4) What the author has actually done is "buy down" their risk. It is simply less likely that the product will become non-compliant when their software is loaded. But it is still possible. At second glance, those dev boards don't come with a power supply. What is your recommended power supply to use Author? Have you tested your setup with said power supply and have test reports at the ready for when Friendly Spectrum Person comes knocking?

5) Sure it seems clever but Friendly Spectrum Agencies actually have quite far reaching and scary powers. Don't think that your little sleight of hand here is clever and protects you. Fundamentally: You are repackaging + modifying an existing product. The steps you are taking in between to "launder" your liability are irrelevant.

Frankly, it's shit like this, that makes it harder for everyone else playing by the rules. It did actually used to be easier. There used to be exemptions for "low volume" products. But all of those were seen as loopholes and HEAVILY abused. Now these toys have been taken away, with more to follow.