Thomas wrote:jimlongley wrote:Thomas wrote: . . . I'm not sure if this is applicable to laptops. I think they said USB keyboards weren't susceptible, but if you have a desktop and a keyboard that uses a PS/2 connector (the round one), all anyone has to do to figure out what you're typing is plug into the power lines going into your home. Basically the motherboard supplies power to the keyboard, it allowed unique signals for each key to go through.
Also, every electronic device emits electromagnetic (EM) fields. So, if you have an electronic safe/computer/whatever, every time you press a key, it emits an EM wave. Each key produces a unique EM wave. So you don't have to be in the same room to detect what keys are being pressed.
Don't believe me, see
http://en.wikipedia.org/wiki/Computer_s ... re_failure
The problem is though, if you wrap your computer in tin foil, it will probably overheat
(If I recall correctly, it's pretty easy/cheap to build devices that will detect these signals. Basically it's security through EXTREME-obscurity.)
Back in the early days of computers we used to play music on them by tuning an AM radio off station near the cpu and then running specific timing loops. Some folks got pretty good at it.
As far as the PS/2 connector thing, this old telephone company noise engineer can't imagine how that would work. The power supply supplying the DC that runs the computer would adequately isolate the keyboard from the power mains to keep keyboard signals from "leaking" that far away. And in my case, since I protect my computers with UPS, which uses AC to charge a battery which continuously provides power to an inverter that provides AC to the PC, I have an additional layer of isolation.
Key signals are very weak to begin with, they only have to make it from the keyboard to the computer, and pretty short duration without a header or anything, so although I know of and even have used hardwire keystroke recorders, I find it hard to believe that they would make it very far. Even signaling systems that are designed for wireless have some pretty severe distance limitations.
You might be right, after three and a half years, I might have gotten the details mixed up. They said the grounding wire acted as an antenna, not the transmission line.
http://www.pcworld.com/article/161166/article.html
Maybe, but not likely, there won't be a direct path for ground all the way from the keyboard to the power line, and even if there was, it would be EXTREMELY noisy.
The wikipedia article is kind of poorly supported, at least in terms of detecting signals from the keyboard. First of all, they keep talking about "harmonics" which means that they could not detect the fundamental frequency involved. Since the signals involved are digital, they are therefore made up of the fundamental frequency plus all odd harmonics, creating a square wave. Harmonics are much lower in power than the fundamental frequency, rarely more than 1/3 the level of the fundamental. A couple of harmonics are additive in nature, how else would you get a square wave, but the subtractive harmonics would not be filtered out in this environment, leaving very poor signal strength to be detected, which would require sophisticated equipment just to start.
Secondly, they are talking in the PS/2 and USB cables about paired wires, the "ground" wire being the return path for the voltages driving the currents making up the signal. Since what is going out equals what is going int, the net current in the path is essentially zero. I know this is counter intuitive, but here's a simple experiment: Take a clamp on ammeter and clamp it around the power cord for a lamp, and you will see a very low to zero current flow (in most cases, there are exceptions, but beyond the scope of this little treatise) now, split the cord in half and clamp your meter around just one of the two conductors, and you will see all the current flowing. A simplistic explanation, to be sure, but it suffices, except we are talking DC and square waves (I know, DC doesn't have waves, but how else are we going to get the signal through than modulate the DC?) so there is a little variation from what I am stating, again, not enough space or time. This, BTW, is how GFI outlets protect you, they compare the current going to the device with the current coming back, and if there is a difference, they trip.
Paired wires are designed to cut down on leakage of signals, either out of the cable or in.
Without really getting too deep into the nitty gritty, the PC Magazine article is one I would have love to have seen and answered. Once again, the connection from the keyboard does NOT go directly, as the article states, to building ground, it goes to the PC, where it joins with the other grounds and eventually may trace to the PC's ground connection. The problem here is that the "grounds" referred to may not be "ground" in the term that I think of ground as a telecom engineer and ham radio operator.
The system ground in the PC is ideally separate and floating from ground, but may be, but hopefully not, at the same potential as the electrical neutral. I could go into great detail and explain why the neutral of the electrical system is not necessarily at ground potential, but the best would be for you to take my word for it. Yes, it should be at or near ground potential, but it is not always,and that leads to all kinds of issues for telephone and power companies. Remember that the PC has a power supply that is breaking down house AC current into Several levels of DC and just about everything after that point runs on DC, and it is unusual for the DC "ground" to be mixed with the AC neutral, which is just asking for noise.
In electricity and electronics, we have three different symbols for "ground."
The purpose of the neutral, which is often referred to as ground, is as a return path for currents driven by voltages providing power to devices. The purpose of true ground in building and home systems is to carry away fault currents. In a typical house (not mine and some other hams I know) true ground really only makes it as far as the breaker or fuse box, and connects to the power company neutral there, and the neutral is what "penetrates" the house. Most buildings incorporate a ground/neutral system which employs a "Principle Ground Point" which is where the power company neutral and the ground join, and then separate to travel further in the building, and they should never join again, that would lead to all kinds of issues, remember ground is for protection and neutral is for current.
The National Electric Code was revised in 1999 to recognize the difference between ground and neutral and made four wire connections for 240 volt appliances the law, with ground providing protection and the neutral carrying current.
And since the neutral is carrying ALL of the return currents, including lots of different keyboards, and motors, and fluorescent lights, and lots of other devices putting signals out there. How do you tell which keyboard is which?
I would be willing to bet that they conducted their tests with discrete devices and nothing else on the same circuits.
Boy I wish I had seen the article.
Sorry for the long post, when this old telephone engineer/technical trainer gets going, it's hard to stop me and this sort of signal troubleshooting just happened to be a specialty of mine.