This is a two-part series where we’re going to look at developing a hypothetical tracking device that incorporates Bluetooth Low Energy (BLE), GPS, an accelerometer, and a USB rechargeable lithium battery.
When you have a new product that you want to develop, the first thing you should do is look at the big picture. This is what established hardware companies already do.
You don’t want to jump right into developing a production prototype because you will get lost in all the fine details.
Instead, first focus on what I call a preliminary production design (or sometimes just called a preliminary product design). The key is that it focuses on a manufacturable version of your product, and not just on a Proof-of-Concept (POC) prototype.
This case study is based on a sample of my do-it-for-you Predictable Hardware Report service. You can download the sample report here . The goal of this article is to teach you about the preliminary design process and all of the information it provides, not to talk about my service.
A preliminary production design focuses on answering some of the important questions that you will need to know as soon as possible. This includes questions like:
– How difficult is it going to be to develop your product?
– Is it feasible to manufacture your product?
– What is the cost to develop your product?
– What is the cost to scale from a prototype to volume manufacturing?
– How much will it cost per unit to manufacture your product?
Ultimately, one of the most important questions you want to answer is how much will it cost to manufacture your product. The manufacturing cost is going to determine how much profit you will make and how much you can sell it for.
Manufacturing cost is something you need to know as early as possible. Unfortunately, many entrepreneurs make the mistake of first fully designing their product for production. After all that’s done, they go back and start figuring out how much it’s going to cost to manufacture.
That’s not the way experienced tech companies do it. They answer all the big questions before they start sinking resources into fully developing the product.
There is no point in spending all of the time to try and develop a new product if it isn’t feasible to manufacture, or costs too much to manufacture.
A preliminary production design gives you all of the information needed to accurately estimate the manufacturing cost for a new product before the product has been fully developed. It also gives you early insight into the manufacturing feasibility of your product.
By starting with a preliminary production design it allows you to more accurately estimate all of your costs, create a more accurate timeline of milestones, and be able to better determine the manufacturing feasibility of the product. Ultimately, it allows you to develop a better product that is more likely to be successful.
System Block Diagram
The first step in developing a preliminary production design is to design the system level block diagram. Our hypothetical product includes four primary functions: Bluetooth Low-Energy microcontroller, GPS, an accelerometer, and power management.
Figure 1 – System block diagram for a hypothetical BLE GPS tracking device.
Bluetooth Low-Energy (BLE) microcontroller
The core of the product will be a Bluetooth Low Energy microcontroller. This includes both the Bluetooth Low Energy radio and the microcontroller running your firmware code. Those two functions are embedded on a single chip called a System-on-a-Chip (SoC).
However, I always recommend that you start off using a pre-certified module solution, instead of a chip-based solution, for any wireless functionality, like Bluetooth Low Energy. This module will still be based on a BLE SoC but it will also include the antenna.
There are two reasons why you should start with a module. First, the use of a module drastically reduces your FCC certification cost. That can save you about $10,000 to certify your product.
The second reason is that modules will simplify the design process. Wireless functionality tends to be hard to design properly. In fact, very likely, it will require several iterations to get things just right.
Using a module reduces your time to market which is always important. Even big tech companies start off using a module so they can get their product to market faster.
If you scale up to producing 100,000 units or more, then it may be a good time to switch to a custom chip design to maximize your profit margin.
The other big piece involved with this product is GPS (Global Positioning System) which communicates with the BLE microcontroller via a UART serial interface.
As with BLE, we’re going to also use a module solution for GPS. The primary reason for using a GPS module is a bit different, however.
GPS is a receive-only technology. You’re not intentionally transmitting radio waves outside of your product. This is in contrast to Bluetooth Low Energy, where you are obviously receiving and transmitting data.
The FCC is primarily interested in the electromagnetic energy a product radiates (transmits), not what a product receives.
I still recommend using a module for GPS because it’s a very challenging technology to implement properly because you’re detecting extremely weak signals from space. In order to detect as many satellites as possible, it’s critical that a GPS design be fully optimized.
The accelerometer is an electronic sensor that measures both acceleration and orientation. For example, if you stick an accelerometer to a car, it will detect when the car starts moving, stops, suddenly hits a bump, or any other movements. You can also use an accelerometer to detect orientation.
This is how a smartphone can switch between landscape and portrait displays. An accelerometer can detect orientation because it detects the direction of gravity which defines the up-down direction.
The accelerometer interfaces to the BLE microcontroller via a simple two-wire I2C serial interface.
Now let’s look at the power section. We’re powering it off of a lithium polymer battery which outputs 3.7V. It will be recharged through a micro USB connector, which feeds into the battery charger chip. This chip takes care of all the complex stuff associated with charging a lithium battery.
Then, that lithium battery will power a voltage regulator that outputs 3.3V. This is what powers the Bluetooth Low Energy microcontroller, the accelerometer, and the GPS.
Also included in the power management section is battery monitor function to measure the charge level of the rechargeable lithium battery.
If this were an alkaline battery, you can measure its charge level pretty easily by measuring the voltage on it. For an alkaline battery this is quite simple because of the linear discharge curve. That’s not the case with lithium batteries.
A lithium battery has a very non-linear discharge curve. It remains fairly flat up to a point and then it quickly nosedives. You need to use a special chip to measure the charge level of a lithium battery.
This battery monitor chip is also commonly called a fuel gauge. Just think of the energy in your battery as the fuel for your product.
The battery monitor chip will then pass along the battery charge level through I2C to the BLE microcontroller. The battery charge state can then be transmitted to a mobile app for monitoring by the user.
Now that we have the system level block diagram, the next step is to select the production components that will perform each of the functions specified within the block diagram.
For each component selected, I recommend that you list the range of allowable supply voltages and the maximum current consumption. Doing so offers several benefits including:
– Quickly determine what supply voltages are required for your product.
– Easily determine the current levels each of those supplies must supply.
– Becomes more obvious if any level shifters are required.
– Allows you to make early estimates on battery life and battery size.
Figure 2 – List of production components selected for our hypothetical product.
Bluetooth Low-Energy (BLE) microcontroller
For the Bluetooth LE microcontroller I’ve selected the BT832 module from Fanstel. This module is based on a popular SoC from Nordic Semiconductor called the nRF52832 which embeds an ARM Cortex-M4F microcontroller running at 64 MHz.
Figure 3 – The BT832 Bluetooth Low-Energy microcontroller module from Fanstel.
The BT832 includes 512KB of flash memory for program storage and 64KB of RAM. It’s programmed via a Serial-Wire-Debug (SWD) interface but also supports Over-the-Air (OTA) updates. It has a line-of-sight range of about 100 meters.
The BT832 also includes a built-in antenna. Although many wireless modules may require an external antenna (external to the module, not necessarily external to your product), I’ve never seen a BLE module without an embedded antenna.
For the GPS function I’ve selected the SIM33ELA module from Chinese manufacturer SimCom. If you are a maker you already may know this company from their SIM800 cellular module that is popular with hobbyists.
Figure 4 – The SIM33ELA GNSS/GPS module from SimCom with integrated antenna.
Although I’m calling this a GPS module for simplification, it’s technically a GNSS module. GNSS (Global Navigation Satellite System) is the broader term used for satellite navigation since GPS really only refers to the U.S. system.
A GPS module will only work with the U.S. operated satellites, whereas a GNSS module will also be able to make use of other satellite navigation systems.
For the accelerometer I’ve selected the LIS2DE12TR from ST Microelectronics. This is a 3-axis accelerometer that can be used for measuring both acceleration and lateral orientation (using the direction of gravity as the reference).
The LIS2DE12TR consumes only 270uA of current and interfaces easily with a microcontroller via I2C.
The power management section for our hypothetical tracking device consists of four key components: a battery charger chip, a fuel gauge chip, the battery itself, and a single linear voltage regulator. It also requires a microUSB connector and a JST battery connector.
For the battery charger chip I’ve selected the BQ24092 from Texas Instruments. This chip will be powered from the USB connector which provides a 5V supply. The charger can then supply up to 1A of battery charge current.
For the fuel gauge chip I’ve selected the STC3100 from ST Microelectronics which measures both battery voltage and battery current to calculate the battery charge state.
For the battery I’ve selected the GM571322-PCB from PowerStream. This battery has a capacity of 120mAh.
I’ve selected a battery that includes the protection circuitry required to prevent the battery from possibly exploding or catching fire when overcharged or short circuited. This protection circuit is embedded on a small PCB taped at the top of the battery.
The final component is the TLV70233 linear voltage regulator from Texas Instruments. This chip steps down the 3.7V battery voltage to 3.3V for powering all of the components in this design.
Next we’ll take all these components and start pricing them out to understand the full production cost for the product. But before doing that, we’re going to first look at some of the technical specifications that we can now determine.
Once all of the critical components have been selected you can determine many of the key technical specifications for your product.
Figure 5 – Some of the key specifications that can be determined from a preliminary design.
I recommend listing them out in one place for future reference. Some of these specifications will also be needed for estimating your product’s manufacturing cost which I’ll discuss more in part two of this series.
There’s no set requirement on what you have to include in this section and it depends on your product.
For a battery operated product you will definitely want to list the power specifications such as average power consumption and estimated battery life.
When specifying the average power consumption for your product make sure that you take into account the duty cycle of any high current components.
For example, the GPS module I’ve selected for this case study consumes up to 25 mA, yet I only specify the total current consumption for the product as 20mA.
This is because I’m assuming the GPS is only on half of the time. Another way of saying this is that it’s duty cycle is 50%. The GPS might be setup to turn on for two minutes, then to turn off for two minutes, and then back on again for two minutes, and so on. The SIM33ELA GNSS module specifies a time to first fix at 28 seconds so a two minute on-time should be sufficient (in the video I mention a one minute on-time but that is probably a bit too tight).
With a duty cycle of 50% the average current consumption drops in half, so I’m using only 12.5mA of average current consumption for the GPS module.
When you add the 12.5mA to the current consumed by the BLE microcontroller and accelerometer then you get the 20mA of average current I’ve specified.
I’ve selected a 120mAh battery so that gives an estimated battery life of about six hours.
It’s also important to list some of the specs for the PCB that will be necessary in order to estimate your prototyping and manufacturing costs.
This includes the number of PCB layers required (2, 4, 6, 8, etc.), the number of soldered pins, and the number of leadless packages.
Leadless packages are nice because they take up less space, but they add considerable additional cost to the PCB assembly process, and they complicate any required debugging because the pins are not easily accessible.
Finally, you should include the critical specs for your product’s enclosure including: the type of material, the enclosure dimensions, and the total weight.
This information is necessary in order to estimate the manufacturing cost for your enclosure.
You should also list how many injection molds are required to manufacture your enclosure. The simplest case is a two-piece enclosure with a top side and a bottom side.
Always strive to keep the number of molds as low as possible because injection molds are very expensive.
In the first part of this multi-part series we’ve designed the system block diagram, selected all of the critical components, and listed out many of the key technical specifications for our hypothetical BLE tracking device.
In part two we’ll price out everything that goes into manufacturing the product so we can estimate the manufacturing cost for the product. We’ll also look at how to determine the suggested retail price based on the estimated manufacturing cost.