Sleepace (Chinese name "Hide Sleep") is the brand of Shenzhen Maidiga Technology Development Co., Ltd. Its product RestOn Smart Sleep Monitor is the world's first portable non-wearing sleep monitor, launched in 2014. This sleep monitor can detect heart rate, breathing, body movement and other physical signs in real time and analyze sleep conditions. It does not have electrodes that come into contact with human skin, nor does it need to be worn anywhere in the body. Instead, it collects data through a sensor belt that is 80 cm long, 6.4 cm wide, and only about 2 mm thick. Simply use this sensor strip between the sheets and the mattress to monitor it. This sensor belt is integrated with the main body of the monitor. When not in use, the strap can be rolled up for storage and easy to carry when traveling. The friction between the sensor strip and the bed sheet mattress allows the main body of the monitor to hang from the bed and does not slip. The back of the monitor is a curved back cover, but there are no screw holes on the back cover. The front of the monitor has a cover that can be removed at any time and is magnetically attracted to the main body. Separating the cover plate, you can see the illustration of the operation method, and the cover plate faces the function of the power on and off control. Reposition the cover in the direction of the power-on prompt, and the green indicator light on it will illuminate for a few seconds to indicate normal startup. If the cover is removed at any time, the monitor will enter the shutdown state. Below the movable cover is the front cover of the monitor, which is also designed without screw holes. In order to disassemble, the author tried to open the back cover from the gap in the back cover, but finally broke the machine by opening the front cover. In fact, the front cover and the back cover are fastened to each other by a plurality of buckles, and are not respectively interlocked with the frame. Four magnets are embedded in the front cover, so that they can be attracted to the magnets in the movable cover for positioning. The interior of the machine is flat, with lithium batteries and boards each occupying half. In addition to the wiring of the circuit board, there are two strands of enameled wire connected to the sensor strip. The author guessed that the monitor uses the ARM Cortex-m series MCU as the master from the test mark marking on the board. The main components are on the other side of the board. In addition to the two-color LED, there is also a SOT-23 component on this side, which I speculate is a Hall (magnetic) sensor for detecting the orientation of the movable cover. Correspondingly, there should be five magnets in the movable cover, one more than the front cover. I guess that four of the magnets are aligned with the front cover. The fifth magnet is located in the Hall sensor when the cover is placed in the forward direction. Above. It is easy to remove the two screws and remove the board. On the right side of the other side of the board, there is a small, easy-to-obtain PCB module because it uses different colors. I guess this is TI's CC2540 low-power Bluetooth solution. At the center of the board is Freescale (which has been incorporated into NXP). It is checked that this is the KL16Z128 model, the ARM Cortex-m0+ core MCU, with 128kB Flash, 16kB SRAM. This family of MCUs has a 16-bit SAR ADC that has a higher number of ADC conversions than a typical MCU. There are several larger chips on the board: TI's TLV2764: low-power, low-voltage four-op-amp; Microchip's MCP41100: 100kΩ digital potentiometer; Holtek's HT1381: real-time clock, next to a crystal; Winbond W25Q64: Serial Flash, 64Mbit capacity. The author's analysis of the use of these chips is: TLV2764 combined with MCP41100 for sensor signal amplification, gain controllable, amplified analog signal is supplied to the ADC input of KL16Z MCU. The HT1381 is used to record time and keep running in the off state (as to why the RTC is not used inside the MCU, perhaps for lower power consumption). W25Q64 is used to store sleep data. Because the monitor can't rely on mobile phone or network when collecting, it must have its own storage. There are also many small-sized packages or transistors on the board. I have not checked their models. I only guess that they are mainly used for power management, including lithium battery protection and charging, digital systems and analog front-end regulated power supplies. To get an idea of ​​what the signal collected by this sleep monitor is, I connected the digital oscilloscope to an output pin of the TLV2764 for observation. When a person is lying on the sensor belt, a waveform that fluctuates with the breath can be observed. However, the human eye cannot intuitively determine whether there is a heartbeat signal. The reason may be that the signal tested by the author filters it out, or the heartbeat signal may be relatively weak, and the algorithm needs to be extracted. In addition, the apparent motion of the human body, such as the amplitude of the signal caused by raising the arm, is large, and the output of the op amp is directly saturated. In order to understand how the sensor strip generates signals, the author continues to disassemble the sensor strip. The strip has tight stitching on both sides, but after removing a stitch, the strap cannot be stretched from the side. It seems that non-destructive dismantling is unlikely. Then, the author tried to disassemble from the head. Solder the lead wire of the sensor strip and remove the back cover to separate the sensor strip from the monitor body. The author speculates that the sensor strip is glued and, like after bonding, is cut and formed, making the disassembly difficult. The outermost part of the sensor belt looks like a synthetic leather, and the difficulty of using the glue can be felt from the difficulty of dissolving the upper and lower layers. It is only convenient to disassemble from the end of the lead, where a layer of foam is sandwiched between the two layers. After gradually uncovering the surface layer, the author found that there was at least three layers of structure in the middle, and the bubbles were arranged one by one. This disassembly is completely unrecoverable. The leads from the monitor are still in the sensor strip, and the solder joints on the other end are sandwiched between another piece of foam. But from the strip of foam, you can already see the exposed sensor. By removing the local covering material, it can be concluded that the sensor is in the form of a long film. In order to separate the connector of the sensor from the foam material sandwiching it, the author used a hair dryer for local heating, and it was easier to remove the heat and peel it off. The foam is like a double-sided tape. The thickness of the single piece is about 1 mm. Two pairs of stickers are used at the head of the sensor tape and at the end of the sensor lead. The strip of the sensor can be seen after peeling off the upper layer of foam, but there is a layer of translucent film on top of it. This film seems to cover the entire sensor belt, but because it is too close to the gray surface material, it has been torn many times during the dismantling process. By removing the film here, the sensor can be clearly viewed. Disassemble to this extent, and then continue to strip the sensor from the sensor strip other materials. In fact, stripping a section is enough to predict its full picture. This is a film strip with a width of 13mm and a very thin film. The opaque part in the middle should be a conductive layer with a width of 9mm. From the point of view of the joint, the conductive layer is divided into upper and lower layers. Because it is too thin, the author has no condition to measure the middle. Media thickness. This sensor is said to be a product of TE ConnecTIvity (TE). Information on sensor products for sleep monitoring can be found on the TE website and can be downloaded to the drawings of the sensor product 10184000-01. The structure in the drawing is as follows: Although the sensor size of the RestOn sleep monitor is different, it should be consistent in principle. For the TE sensor used in this disassembly, the thickness of the PVDF (polyvinyl fluoride) layer in the middle is estimated to be only 28 microns. Adding the conductive layer and the protective layer of the positive and negative electrodes, the author uses a digital vernier caliper to measure the thickness reading to 0.05mm. The film thickness of this 5-layer structure is even thinner than ordinary A4 printing paper, and it is very soft. According to the material on the TE website, the piezoelectric film is a thin and transparent film of PVDF (polyvinyl fluoride); it is a material with good flexibility, low density, light weight and good mechanical toughness. The film can be made in a variety of thicknesses and areas. It can be attached directly to the surface of the machine without affecting the mechanical movement of the machine. The piezoelectric film has the advantages of wide frequency band, wide dynamic range and high voltage sensitivity. Another major advantage is that it has low acoustic impedance and its acoustic impedance is closer to the acoustic impedance of water, human tissue and other organic materials. A close impedance match facilitates more efficient transduction of sound signals (>>data sources) in water and body tissue. The author then connects the oscilloscope probe and the ground clamp directly to the two ends of the sensor to see if the signal output can be observed. At this point, the sensor has been disconnected from the monitor circuit and no more circuit processing. When the sensor is subjected to vibration and the moment of deformation, an obvious voltage signal can be observed from the oscilloscope. This means that the sensor does not need to be powered, but based on the piezoelectric effect, it can itself output a varying voltage signal. In addition, due to the high output impedance, the oscilloscope also received significant power-frequency interference under this simple experimental condition, so there is a 50 Hz disturbance on the graph line. So why does the RestOn sleep monitor make the sensor belt so complicated? The author speculates that the reason may be to better transmit the pressure and vibration of the human body to the sensor on the one hand, and to protect the sensor from the tension on the other hand, and to avoid extreme bending, distortion and other deformations locally. In summary, through the piezoelectric film sensor, the weak pressure changes caused by the body's heartbeat, breathing and other physiological signs are converted into electrical signals, and then processed by the internal analog circuit of the sleep monitor, and then sent to the ADC of the main control MCU for conversion into a digital signal. After that, signal filtering, data recording, data analysis, and the like can be performed. Through the data interaction between the low-power Bluetooth and the mobile APP, a graphical sleep report can be generated by the mobile phone software. With the more powerful data processing capabilities of mobile phones or remote servers, it is possible to implement richer health management applications. The content described in this article is my personal opinion and has nothing to do with Shenzhen Maidiga Technology Development Co., Ltd. or TE ConnecTIvity. The logos, products and/or company names mentioned herein may be the trademarks of their respective owners. 3 Mm /8 Mm Nano Tip,Electronic Board Marker Pen,Touch Board Marker Pens,Infared Smart Board Marker Shenzhen Ruidian Technology CO., Ltd , https://www.wisonen.com
