Issue link: https://iconnect007.uberflip.com/i/1108006
22 FLEX007 MAGAZINE I APRIL 2019 shoulders and adapt them to our market needs. For instance, while we're using a standard BLE 4.2 chip, we tweak the protocol slightly to conform to HIPAA. Our devices won't show up on your smartphone and authentication is handled at the application layer. That makes it extremely difficult to hack, yet we can still leverage all of the development in the existing technology. In terms of integrated circuits, we don't have to do our own ASIC. We're not trying to do our own FCC certification; it's all about time to market for a startup, so you want to leverage as much as you can that's out there and adapt only as much as you must to meet the spe- cific demands of your market. We chose a very powerful, robust, low-power, fairly mature technology with Cypress PSoC, and we evalu- ated about four or five different other medical microcontrollers in the market. The second area of great consideration was power. In this iteration, we're going with a pri- mary cell, which has some advantages. The energy density is usually about twice what you can get out of a rechargeable cell, and self-discharge rates are very low; thus, we can build our device with our power management scheme and still achieve a shelf life of a year with no need to recharge the device or some- how insert a new battery into it. This helped us with our design objective of the forgettable SmartPatch. It goes on your arm, it's very light, and part of the reason we can do that is because of the energy density of our battery. It's a semi-custom battery that we had made by a supplier to meet our form fac- tor requirements, so the power management was optimized. Our quiescent current is about 50 nanoamps, so we're down in the region where we're comparable to the self-discharge of the battery, which enables us to get that super long shelf life. There's also the sensor suite itself in terms of the circuit design. Every one of those sen- sors is designed to support a specific aspect of the metrics we're trying to collect. There are optical sensors, which we use to measure hemoglobin/hematocrit, SPO-2, volumetric flow rate, heart rate, etc. We get SPO-2 and hemoglobin hematocrit through some very sophisticated proprietary algorithms. Every- body has a PPG chip on the back of their smartphone, and there are all of these apps for putting your finger up against that to get heart rate and SPO-2, and that's called photoplethys- mography (PPG). But we're doing what we call PPG+. We're pushing that technique to get hematocrit out of a major artery as well as volumetric flow. We do that by combining multiple different data streams and optical channels in a way that we've been developing over the last two years. It's a proprietary approach to generate these metrics, but we're still leveraging off- the-shelf sensors. We're trying to do as much as we can and take advantage of a lot of great work out there to build PPG chipsets and ana- log front ends. Right now, if you go to the Sen- sors Expo and Conference or the Embedded Systems Conference, many wonderful chipsets are available to support the wearable space. We took this circuit design to a couple of different vendors and said, "We want a super flexible region that will be bendable around somebody's arm or the curve of the bone in their arm." At the same time, we had to have regions in the design that were rigid because solder joints don't like to flex, so reliability was a concern. If we had gone with a rigid- flex approach, that probably would have been okay, but one of the things that we liked about Lenthor is they were willing to tackle this problem with what I call rigidized flex, which they had proposed through working together and looking at the cross-sectional structure. The second area of great consideration was power. In this iteration, we're going with a primary cell, which has some advantages.