More wireless communications musings

So while I was working on my car Saturday, I had my laptop in the garage, connected to my wifi, playing music, etc. (DNLA/UPnP is just the awesome.)

Or, rather, I tried to; It took me a while to get working. Despite the seemingly high power output from my router, my laptop simply couldn't reciprocate at the same level; communications between it and my laptop got dodgy. I got through a couple songs, and then I lost signal, and wasn't even able to get enough signal quality to get a DHCP response. Not terribly surprising, as the signal had to get through two formerly-exterior-facing walls. I didn't have one of my 330gEs handy, so I couldn't set it up as a repeater or adapter, and had to get the laptop's internal antenna (wherever it is) oriented and positioned.

And, honestly, 802.11g hasn't been very kind for streaming video; When I watch the bitrate on my PS3, it'll typically hover between 300kpbs and 700kbs in low-activity scenes, bounce up to 1.5-3Mbps for moderate and high activity scenes or "grainy" film effects, and shoot up to 15Mbps for extremely tough scenes (such as sustained fullscreen fuzz).

I noticed that there's an ISM band at 24GHz. 802.11b and 802.11g operates in an ISM band at 2.4GHz, and 802.11a operates in an ISM band at 5GHz. 802.11n operates in both the 2.4GHz and 5GHz ISM bands. But nothing uses 24GHz, as far as I know.

And that makes sense; 802.11b has a hard enough time penetrating using the 2.4GHz band, 802.11a has an even rougher time up there at the 5GHz band. In the 24GHz band, you'd be lucky if you got from one side of a wall in my house to the other.

But what if you didn't *need* penetration, but were only interested in throughput? If, for example, your access point was in a highly visible point in the room, and any device it would need to connect to would have line-of-site? 24GHz sounds mighty nice at that point. Except that it's extremely hard to create circuits that deal with those frequencies; Capacitance between wires in your circuits will be enough to give you headaches; There's a reason PCs don't have internal signal paths that operate at more than a few GHz.

What would be nice would be to be able to operate at a multi-GHz band and consume a huge amount of bandwidth to increase throughput. But that's going to screw with anything else using that area of EM spectrum; That's why we have parts 15 and 18 of the FCC regulations.

But wait...There is an area of the spectrum that isn't really regulated that much by the FCC. There's a nice little spot around 344,000 GHz (344THz, but that's not a number often bandied about.) Yeah, it's going to have severe penetration issues; A piece of paper will stop it, for the most part. But that's because we're talking about infrared radiation now.

Rather than using frequency modulation, use amplitude modulation. That lets you use a source like, I don't know, a standard 870nm LED.

The trick is finding out what the signal frequency response would be of the IR LED and an IR-sensitive photodiode or phototransistor. (Photodiodes will have a faster response time, but then you have to do your own amplification of the signal.)