Wesley's Tool-Box

SK Telecom's Smart Home device line-up was conspicuously missing a motion sensor. It would have nicely complemented the Jikimi home security devices. Considering that the PIR (passive infrared) sensor itself doesn't cost all that much (some as low as US$2) or complicated to work with, it wouldn't have been so hard to create a product out of it.

It turned out that SKT wanted that as a part of a much more beefed up home security solution called "T View Sense" that came out last month. It's a cloud-connected IP camera with the optional sensor package (motion, door, temperature/humidity, smoke, and carbon monoxide). The sensors requires the IP camera to function because they communicate to the server via the camera's integrated gateway. I appreciate the effort SKT is finally making to counter the IP camera efforts from the other mobile carriers (LGU+ had theirs since 2013), but the pricy nature of the camera (official price of KRW159,000 or about US$140) is off-putting to someone who's just interested the sensor alone.

So I decided to be a bit creative and repurpose one of the existing Jikimi devices into a motion sensor instead. Initially, I wanted to modify the SOS button because it was cheaper. But the fact that the alarm it makes doesn't get differentiated between the buttons made it a deal-breaker. With the door sensor, I needed to simulate the door closing/opening with a magnet, meaning that I could either attach the included magnet to a motor or use an electromagnet to fake such action.


Obviously, I chose the electromagnet because it would be far more simpler if it worked. If it generated enough magnetic field, it would act like the original magnet and cause the sensor to send a "door closed" signal. The problem was that I wasn't sure what kind or how big of an electromagnet would be necessary. After looking at various offerings on the internet, I decided to take a stab in the dark and buy a small "lifting electromagnet" used in machines for picking up metallic items. The particular model I bought for about US$5 was rated for 2.5kg at 12V DC, consuming 3W. The model number KK-P20/15 apparently indicates a 20mm outer diameter and 15mm thickness.

12V was tad higher than what I wanted, but the electromagnet works at a lower voltage with reduced pulling power. All I needed to know was whether that pull is just enough to trick the sensor, so I created a quick testing platform with my kids' electric circuit kit. As you can see here, 3V was just barely good enough for the sensor to activate. At 6V, it worked more reliably. This confirmed two things - the electromagnet I had was fit for the job and that I could make it work with a battery pack, e.g. single-cell Li-Ion battery (3.7V nominal).

On December 12, 2016, SK Telecom expanded its Smart Home suite by releasing a device set called Jikimi ("Protector") for basic home security. It consists of an SOS button that can notify either the police or someone you know, and a door sensor that can detect intrusion. They can be bought separately, and while the list price is KRW35,000 (US$30.70) for the button and KRW45,000 (US$39.50) for the sensor, it's currently being sold at a significant discount - KRW22,500 (US$19.70) and KRW24,000 (US$21), respectively. As far as networked sensors go, the discounted price is reasonable. I managed to buy a few of these shortly after they became available to gauge their usefulness.


Both devices come in an identically sized box roughly the size of a hockey puck. An instruction manual and two security stickers are included with each device. The stickers are meant to warn off would-be intruders. While I doubt they would be much of a deterrent, the large one refers to an actual NSOK security dispatch service for the SOS button which you can optionally enroll for an extra KRW40,000(US$35) per year. It's not available for door sensor users, but this sticker is included with that device as well.

Last year, iPhone 5S came with a motion co-processor called M7 that can collect various sensor data relating to iPhone's motion. The most notable use for this was recording how many steps a user took while carrying the device. In other words, the device acted as a pedometer. This year, iPhone 6 series has updated the co-processor to M8 and added another sensor - a barometer. The Health app that comes with iOS 8 makes an immediate use of this. It records how many floors the user has climbed, adding another dimension of knowledge to the user's movement.


Apps that are updated to use the Healthkit in iOS 8, such as Argus, can read this stair climbing data. The notification widget from Argus is able to display the number of floors I moved up during the day if I'm using an iPhone 6 Plus. If I'm using an iPhone 5S, there's no such data available, so the widget simply doesn't mention it at all.

While I'm comparing the data from both devices, I should note is that the pedometer function acted very similarly between each other. However, as you can see in the pictures, iPhone 5S consistently recorded a bit more steps than iPhone 6 Plus, for some reason. Perhaps the bigger form factor and weight of the 6 Plus dampened the recorded movement a bit.

Apple had a big product release day just a few hours ago, with new OS X, iPads, and Macs coming out. While it was left out of the announcement, iOS 7 also had a new point release just after the presentation at Cupertino ended. iOS 7.0.3 lists a lot of changes and fixes, and among them is a mention that the accelerometer calibration was fixed. I personally checked on this issue earlier. What better way to verify the fix than actually seeing it for myself?

As you can see, the problematic horizontal tilt measurement of iPhone 5S is now in agreement with the other iPhones. I can also hold the device up and see that it's reporting fairly accurately. Same can be said for the other measurements, which you can see if you decide to keep reading below.

It seems Apple didn't make an independent patch for this problem and instead lumped it together with other fixes since it wasn't a critical one. At least it came out just a month after the first release and before the release in Korea, so I'm glad the Korean users wouldn't have to be concerned from the get-go.