|
As long as you're within the communication range and in whatever your protocol calls a "reasonable" environment (e.g. not underwater), you're OK. The strength of the signals isn't chosen at random, it's chosen so that it allows devices to function well (within the design constraints -- range, environment conditions etc.) -- and so that it meets the power consumption constraints imposed by size, design, cost and technology. tl;dr when someone decided to use a 4 dBm transceiver, they (should have) made a conscious choice about it and would have verified that it's enough. Quite plainly, if the connection drops when it shouldn't, it's either bad software or bad design. Bad software is either the Bluetooth stack (especially in legacy Bluetooth devices; newer stacks tend to fare better) or the firmware that talks to the Bluetooth stack (if it's BLE, this is the safer bet). Bad design on the hardware level is less excusable than it would seem, because a Bluetooth device is not exactly a long-haul wireless link that has to work during the mother of all thunderstorms. There's a wealth of information and modeling tools that help you with this stuff, too. Bad antenna design, bad filtering, bad casing are all common culprits, but I've seen a lot of electronic engineers who wouldn't call themselves RF engineers get it right by just following common sense and doing the math. Sometimes (but even less often), it's not a design problem, it's a manufacturing problem (e.g. PCB antennas that get damaged or don't get etched right. But this is pretty rare. Certainly, the operating environment plays a role, but that's why you consider it while you're designing the whole gizmo. You don't (or shouldn't) get to shrug and say hey, it's not my fault we're basically walking cucumbers. So if a pair of super high-end Bose speakers and a super trendy iPhone keep dropping connection while they're within range, not inside and, respectively, outside a Farady cage or whatever, the reason isn't the black magic that RF design is, it's Bose's and Apple's profit margin. Edit: I do concur with the other fellow who posted here, Bluetooth really is complex and the band it operates in isn't exactly a charm to work with, but frankly, neither of these reasons account for the huge range of devices that are simply badly designed and/or run bad software. There are a lot of Bluetooth devices that work just fine, a lot of other protocols that operate in the 2.4 GHz band and work just fine, a lot of other protocols that operate in similarly noisy bands and work just fine and a lot of protocols that are more complex and work just fine. |