It depends on how much time in the test chamber you need and how many retests. It can be many $10k-$100K.
But since you can't really test until you have a production PCB layout and then you might have to move tracks or add caps/inductors/ground planes - the costs of a redesign are normally much much higher than the testing
This is why you sometimes see, for example, a device that is designed to accommodate RF shielding but the shield is not installed. It's cheaper to design the shield and then omit it in production if the tests happen to pass without it, than it is to risk failing a test and have to go back and redesign with better shielding.
I've always wondered what sort of devices would actually break the FCC regulations on interference. I'd imagine the only things that might do that are radio devices and jammers. Does anyone have experience with such devices?
I presume those regulations exist for a few reasons:
1) In any wireless device, it is not hard to imagine a bug causing problems
2) In any electronic device carrying oscillating currents, there is EM emission, therefore it is possible for unintended leakage on bad frequencies (e.g. this trace is just the right length to be a broadcast antenna at frequency Y, given normal modes of operation.
3) Some devices may individually not cause problems, even though they are low power emissions on certain bands, but cumulatively they may cause signal degradation in those bands.
4) Having and enforcing them forces engineers to ensure they are met up front, resulting in better, nicer playing devices to begin with, as it is cheaper to engineer it right, than to deal with the consequences of doing it poorly.
It's also something of a hold over from larger electric devices. A poorly designed laptop can mess with your local radio reception and networking equipment, but a poorly designed 50KW electric motor can disrupt radio transmission over a much larger area.
Find someone with a running TRS-80. Bring your favorite portable AM radio with you. Have your friend run a program, and while that's occurring, put your radio next to the machine and start tuning it around. It won't take long before you find one of the (many) harmonics emitted by the computer.
Perfect example of such devices are cheap Chinese replacement chargers for just about anything. Anything that contains oscillator radiates RF and when the device switches relatively large currents periodically (like for example switched mode power supply) it becomes pretty good transmitter unless adequately shielded and filtered. Another significant source of EMI is almost any kind of digital sequential logic which contains signals with very sharp edges, which are mostly synchronized with each other.
The Raspberry Pi had to pass radiated and conducted emissions and immunity tests in a variety of configurations (a single run can take hours), and was subjected to electrostatic discharge (ESD) testing to establish its robustness to being rubbed on a cat. It’s a long process, involving a scary padded room full of blue cones, turntables that rise and fall on demand, and a thing that looks a lot like a television aerial crossed with Cthulhu.
As someone who's looked into getting a substantially similar device CE-marked, getting it all done inside a week is pretty heroic.
15 comments
[ 74.5 ms ] story [ 120 ms ] threadI didn't see anyone mention how much the process costs. Does anyone have any experience with this?
1) In any wireless device, it is not hard to imagine a bug causing problems
2) In any electronic device carrying oscillating currents, there is EM emission, therefore it is possible for unintended leakage on bad frequencies (e.g. this trace is just the right length to be a broadcast antenna at frequency Y, given normal modes of operation.
3) Some devices may individually not cause problems, even though they are low power emissions on certain bands, but cumulatively they may cause signal degradation in those bands.
4) Having and enforcing them forces engineers to ensure they are met up front, resulting in better, nicer playing devices to begin with, as it is cheaper to engineer it right, than to deal with the consequences of doing it poorly.
It turned out I wasn't going mad, the thing was picking up local CB radio.
The Raspberry Pi had to pass radiated and conducted emissions and immunity tests in a variety of configurations (a single run can take hours), and was subjected to electrostatic discharge (ESD) testing to establish its robustness to being rubbed on a cat. It’s a long process, involving a scary padded room full of blue cones, turntables that rise and fall on demand, and a thing that looks a lot like a television aerial crossed with Cthulhu.
As someone who's looked into getting a substantially similar device CE-marked, getting it all done inside a week is pretty heroic.
[1] http://www.raspberrypi.org/archives/978