I have an application where I want to have a PCB magnetically attached to a another pcb board with a black cloth in between. The outside board is powered from, and can send/receive communication from, the board on the other side of the cloth. I do not want to puncture the cloth.
Powering the outside unit is a no-brainer using a PCB planar transformer on both boards (similar to wireless charging). The communication is where I could use some advice. Any ideas? I was kind of thinking I could have two secondary planar transformers for communication, but I'm sure there are also some microcontrollers out there that might have a built-in low power rf UART. Optical UART is not an option. To further complicate this question, I have some clients in very sensitive government facilities that don't allow anything that can transmit. So how would one transmit from one board to another without a "transmitter"? Somehow I'm thinking a magnetically coupled very low frequency very low duty factor signal transmission might be an acceptable method. Or perhaps ultrasonic? Heck, I guess I could revive DTMF tones as a transmission means to get past the "no radios" requirement? Thanks.
I have some clients in very sensitive government facilities that don't allow anything that can transmit.
You absolutely need to work
with those clients to establish what "transmission" means and whether any proposed work-around is okay. If you're proactive and open about what you're trying to do and why, you probably won't have anything worse happen than losing the client (or just the sale, because they simply can't use what you're trying to make). If it looks like you're trying to work around those sorts of restrictions surreptitiously, you could end up
in jail, not to mention banned from working with the government in any capacity.
Check NFC (near field communication). As opposed to RFID, it's bidirectional and there are plenty of IC solutions available. Problem is, that data sheets are often confidential and under NDA.
I'm quite sure audible DTMF tone production is not considered a transmitter in any government facility, so will likely just go with that, or local coils that are magnetized and read out by a few Hall effect sensors. I have never had to provide my schematics to any customer, and if a customer were to ask, I would not sell to them.
You don't have to disclose your schematics, but you should provide information that would be relevant, and volunteer it you know it is important. If your client asks "does your device have a radio or transmit data" and you say "no" but omit "but it uses near field EM communication", that's where you get in trouble if somewhere down the line someone decides that is a problem and you withheld information that they feel you should have volunteered. And if we are talking about agencies like the NSA, you can bet that in the very unlikely event that it got to that point, they would track down your posting history here and use that as evidence that you were worried there might be a problem but didn't tell them. Seriously, don't ask us, ask your customers or their government partners.
What is OK depends on who your customer is, where they are using the product, and how they are using it. If the device will be explicitly handling sensitive information vs. just being in the secure facility that is where you really get into trouble. Audio / ultrasonic is not necessarily a free pass either -- plenty of places have rules against any device with a microphone.
It's also certainly possible it's no big deal -- near field coupling may be perfectly fine. But nobody here is in a position to tell you what is OK.
Isolated data transfer through magnetic coupling is easy. Bench multimeters often use it.
Starting with a pulse transformer, one way is to differentiate the digital signal and use the pulses to drive the primary. Then on the secondary side, the positive and negative pulses are turned back into the digital waveform using a set/reset flip-flop.
Regarding: clients in very sensitive government facilities that don't allow anything that can transmit
You should also expect (at least, prepare an answer for) push-back even if you use NFC or any inductive coupling, because people will automatically switch off to the idea even if you can demonstrate it can only operate millimeters away. That's not the point as far as non-tech people are concerned. There's an extremely good chance that at least some will consider inductive coupling to be no different to radio comms, and not want such an implementation.
NFC for example has been demonstrated over significant distances (~10m), albeit with a pretty sensitive (read: sizable) antenna.
I'm more curious how you or the customer thinks an electronic device is usable at all in such an area. Knowingly, anyway. Perhaps that's the point, and it is literal spy equipment; which, AFAIK, is not at all illegal to make and sell, by itself, but don't be shocked if you have suits asking for your sales history at some point, I guess? (It's safe to assume they have identifying information; there are legal processes to discover your forum email, IP addresses, VPN accounts, etc. Maybe not if you've managed to do all that via Tor or something (does Tor interface meaningfully with the rest of the internet, I don't actually recall?!), but if someone like the NSA took interest, they're able to piece together network logs and deanonymize across that.)
In any case, the crime would most likely be the customer's; you would only be culpable I think if there's no obvious or common lawful use of the device. Though they might still get you involved legally, and then you'd have to prove as much to a judge, or jury, at your expense. Maybe factor that consideration into the design/sale price (which is perfectly reasonable: legal costs, criminal or otherwise, are just another business expense), and perhaps consider business insurance.
Likewise, it would be the customer's burden to prove that, for their purposes, your device is adequate for their task, whatever that may be; legally speaking, your intent isn't to enable unlawful acts (...one would hope), and if your device just happens to radiate detectable signals, that's their problem. One must assume spy agencies know enough about things and stuff to test their equipment in detail, and will select from many possible devices, or perhaps modify them to their own ends. If your device is accidentally undetectable, that's not necessarily your problem... but again, proving that in court might not be something you want to risk.
Or something like that. IANAL, not legal advice, etc. etc. Definitely sounds like something you might want to check with a lawyer first.
Tim
I once suggested introducing a form of up-close "configurability" to a product, and it was immediately a showstopper as far as salespeople were concerned because certain customers will always ask if there are any wireless comms, and it's not a great starting point to have to caveat that on a technicality (to the average person, inductive and RF is completely identical, and they are completely right from their perspective the only difference is range, plus, in their eyes, pedanticness from engineers about the underlying physics).
It's then an uphill struggle to persuade differently because then they need to prove it is OK to their management. The customer will simply go for an alternative, even if it's more expensive or not as good otherwise. Anything that needs a slight caveat could take them ages to get approved, and that's assuming the customer wants to make that effort internally (most will go for an alternative if they can).
Capacitively couple the data?
(It's safe to assume they have identifying information; there are legal processes to discover your forum email, IP addresses, VPN accounts, etc. Maybe not if you've managed to do all that via Tor or something (does Tor interface meaningfully with the rest of the internet, I don't actually recall?!), but if someone like the NSA took interest, they're able to piece together network logs and deanonymize across that.)
Wow man! Paranoia will destroy you. I sell devices all over the country. I haven't mentioned my business, my name, or the product here. It is my risk, and I'm quite sure nobody is going to be looking up my IP transmissions way back during my prototyping efforts just because I build a device that can ultrasonically or magnetically couple data 0.1 inches across a cloth. Every piece of electronics that has a microcontroller is likely coupling data that far via EMI on circuit board traces. I'm going to make it a simple answer. No, there are no rf transmitters in the product I produce. That is the question they ask, and they are required to ask it based on the telecommunications FAR chapters. I too have to answer those questions in my government SAM certifications.
NFC would look like the most reasonable solution in your case. ams used to have nice and cheap ICs for custom NFC links, but they stopped selling that when they got bought by Osram. A shame. But ST does have a whole range of NFC interface chips with public datasheets.
For power, you could use Qi, which is also standardized. Heck, if the communication is a small amount of data, you could probably use Qi to transmit this data as well, I think it can transfer some data too, but I don't remember if it's bidirectional. To be checked.
If there's no need for anything standard or very secure, you could just implement your own inductive power/data transfer - if the distance is just a few mm, the power to transfer is limited, and efficiency is not a real concern, it's a relatively simple endeavor too.