Actually, the solution is pretty straightforward but it does have some considerations. For such a system, you would have to define a standard interconnect architecture and make the modules intelligent so that they know who they are talking to. Making each module ‘smart’ or intelligent will not be much of a challenge as we have many low-power and optimally-powerful processors available to do the job.
This way the base module or the main master or the controlling module can easily make decisions on what kind of module is attached to it; it can be a camera or a battery or any other module. Now everything will only depend on how much area we want to use and what kind of functionality we want to have in the phone.
Some people suggest that we can also have architecture of common control bus and have the modules complying to the control architecture, with a separate data bus which takes care of data. Ashwin Ramachandra, Sasken Communication Technologies, explains, “Looking it from signalling perspective, you need to start looking at signal flows and then identify control signals. If you are able to create a backplane where all your signals flow on one particular set of lines, you are beginning to establish something like that.”
We can actually borrow some mechanisms from existing technologies as well. Ashwin adds, “If you go to the automotive segment, there is something called a CAN bus. Let’s say, if you start loading your operating system and a signal goes on to your CAN bus, the CAN bus has a mechanism of arbitrating and then trying to figure out how to start communicating with the signal.”
However, such solutions will also carry with them complexity issues.
As already discussed, by making the phones modular one thing we are doing is, having an architecture where we are adding connectors (or interfaces), and in the process of adding these we are adding a lot of communication chipsets. This is going to make a significant impact on the power requirements of the device.
Also, if we are making the modules intelligent enough to be swappable, every module will have a computing element on it, and instead of one computing element now we will have multiple controllers, which means additional power requirement.
Increasing the amount of intelligence on the device will only add to power requirements and power needs will increase in proportion to the number of modules we have on the device.
If we just have a single physical PCB, the shortest route is much smaller. But if you are drawing wires across components, obviously we are taking a longer route. Pure physics tells us that we need extra assistance if we are taking up such a design, which directly affects the performance as a power consumption issue.
RF behaviour is one of the most significant aspects of regulatory approval for any device. While the industry is still waiting for a functional prototype to test RF behavior on a modular device, many experts predict that the modular architecture would show better RF response.
Nate Srinath, founder-director, Inxee, says, “RF behaviour should have no impact here, or in-fact in modular architecture it will be even better. In our experience we have seen that the more we isolate RF interfaces, the performance gets better. It happens so because when we have the integrated PCB, lot of noise issues from other components like the battery, LCD or the touch sensors have an impact on the RF signalling. But as we will move towards modular, we will actually isolate all of the RF components from the other parts of the system, so modular will actually make the performance better.”
The whole project of modular smartphone is a big challenge. Most mobile phone manufacturers are focused on reducing sizes of the phones for better user experience and looks. The current integrated design allows designers to innovate in areas such as curved displays or higher-resolution cameras.
But when we talk about a mobile phone, a major concern is the form factor. Making the device more modular will definitely impact on size of the device. We can easily predict that, with a modular architecture size of the phone is only going to increase. So, keeping the form factor same for the modular device will be a challenge for the designers and engineers.
The other issues that should be taken care of while designing a modular device would be liability and overall ruggedness of the modules. What will happen if the device is dropped? If it starts coming apart then we have an issue. Additionally, the modules and interconnects must be rugged enough to handle frequent plugging and unplugging.
Besides, today’s non-modular devices are made to resist mechanical shocks in terms of functionality as well. Ensuring the same thing with the modular device will be a bigger challenge. We need to engineer for the fact that mechanical shocks have least impact on the performance of the individual modules as well as the overall body.
A feasible technology standpoint
Modular architecture in smartphones is challenging but it opens up a new area of research which is probably going to be very exciting for the engineers. It is definitely a feasible technology standpoint but the associated considerations must not be sidelined.
When a device becomes mobile or mobility oriented, we take care of many issues. But if we are going modular, we are actually increasing those engineering challenges that are not going to be very easy to solve.
The key focus in the modular designs would be to ensure there are common communication and interface mechanisms at both hardware and software levels. As long as the modules adhere to these specifications, these issues would be addressed.
The author wrote this as a technical journalist at EFY. He recently opted and shifted to the hands-on training division of EFY