TL;DR - he found it, by sending a giant PDF thereto and recording network traffic info via 3 computers, extracted the signal strength of the desired device (Kindle), and mapped the convergence of 3 spheres of that signal strength, locating it within a couple feet.
I found the break in a 10m extension cable, by measuring the capacitance at each end and calculating the ratio of the two values. This predicted a break at roughly 1m from one end. Then started probing the cable from 10cm towards the ”short” side, moving towards the long side. Found the break within 15cm, saving roughly 9m of cable.
That's pretty clever. Also if you are interested (not 100% applicable because it's not something most people can do at home) there is a practice for finding breaks/partial breaks in cables using time domain reflectometry (TDM) [0]. You send a pulse in to one end of the cable and measure how long it takes for some reflected energy at the break to return. It's pretty cool stuff and you only need access to one side of a cable. I have used this in my job for finding discontinuities in large RF cables.
If you do have some electrics equipment at home (o-scope, sig-gen) you can make do it yourself [1]
We did this in the Navy to check circular wave guides to the ECM antennas in the leading edge of our bombers wings. We had to do this once every 6 months (I think) with a sweep generator to see if the cables needed to be replaced. Also, installing these were a complete PITA! Nothing was a straight run to where it needed to go, so you had to bend them as gently as possible as you fed the new line into the wing. Then you had to test it again.
These were very expensive and I've seen more than a few screwup's while installing them.
A coworker's cat-5 tester also has a TDR function that shows you where an issue is on the line.
Many _many_ years ago we used to have to do that on occasion in our office to find where the break in the coax network was. That brought some amusing memories back for me, thanks :)
I got the capacitance idea from having worked with TDR in the late 80's. As students we used them to tune FM antenna combiners, but I never quite got the hang of it at the time.
Unless there was an obvious reason for the break then there could easily be another "nearly broken" fault in the cable e.g. if break was due to stretching.
A partial break has the risk of local overheating.
So perhaps you should biff the cable anyway if it is likely to be unsafe?
This is a good point - the cable is probably 20+ years old and has been used for weed eaters and lawn mowers. It's been regularly uncoiled and re-coiled, not always correctly.
My solution would be to hire an estate sale company, who will bring in a team to inventory every item in the house while selling it off to the general public. Give them instructions to sell off everything except Kindle readers. By the end of the week you will have a clean, empty house except for one pile of Kindle readers. Problem solved!
Who said the house was not organised/clean? I don't think it's a fair assumption at all. Even ignoring the fact that this wasn't his house (so organising it was probably not an option), he found the kindle in a closet... meaning his mom (or someone else living in the same house) cleaned/organised it and then just forgot about it.
The simpler solution is to just remotely set off an alarm. I do this frequently with the "Where is my Android" thing. Internet-enabled devices should include this capability by default, anything more is just over-engineered.
Not supported by all devices. It's unlikely it's easier to get everyone to agree to support this feature than a firmware update to wireless access points (which might even be open source depending on the hardware specd) to support time of flight ranging. If you can ship it, ship it. Don't wait on someone else.
Unless I'm missing something your assuming multiple (3 minimum access points, having worked on a similar project our accuracy required more to make up for inconsistencies and multi-level facilities) at home (where I get away with just a single one) and that they have been calibrated to know where they are in relation to each other.
Yes, I am assuming that is more likely than convincing manufacturers to support alert functionality natively. Considering the rise of wifi repeaters (Google, Eero, Netgear, ec) and other IoT devices, I don't believe my assumption is far fetched. Consumers have smoke detectors, cameras, garage door controllers, and water sprinklers on their wifi now.
You could make a simple 3-part product that does just that. You place them in various parts of your house, they calibrate together (play an inaudible sound and measure delay), and then have them manufacture traffic to your target device to figure out where it is.
I imagine they could be somewhat close together (e.g. same room), which makes setup way easy.
I can't imagine it would be very difficult to make, provided small Wi-Fi capable devices are sufficient, it just needs some work on the details.
I wonder if RFID would be an effective solution here. They're cheap, small, and don't require batteries, so all you'd need to do is make a device that can generate a powerful, directional signal and you could quickly locate whatever you're looking for. I'm not an expert on RFID, but it should be a workable solution, and would work even if the device is off (e.g. ran out of battery).
