Controls Peripheral SOM Unified v. Multiboard Design
Controls has the ability to have a different board to control lights, pedal potentiometer sensors, and cameras. The peripheral SOMs take care of these tasks but we can approach the board design in two different ways:
Unified Board Design:
Intended to be a singular daughter board design that is to be used with the standardized peripheral board, it will basically contain all the components that both of our peripheral SOM packages need, lights and pedals. Distinct boards will be used at each peripheral SOM package instance, but they will be the same design.
Multiboard Design:
Intended to be 2 different daughter board designs that are to be used with the standardized peripheral board, each design would be tailored to their respective package and usage. Basically, 2 daughter board designs 1 for any lights peripheral SOM packages, and the other for the pedals peripheral SOM package.
There are advantages and disadvantages of both! Let’s consider the constraints of each approach as well as the benefits.
Unified Board Design
1 Design → Used as daughter board for ALL peripheral SOM packages
Pros | Cons |
---|---|
Standardization Using a single board design will simplify the production and manufacturing of the boards, much like how we standardized the leader and peripheral boards and created the SOM system of PCB design | Overdesign By creating a single board with 2 different functions, we could end up increasing the complexity and include components that are only relevant for one purpose. This could lead to inefficient use of space and we could increase power consumption for no reason. |
Manufacturing Since the same board will be manufactured and produced for both of the peripheral packages, we can approach this design like how we approached the SOM, order higher quantities and always have backups. | Increased room for error With increased complexity and less modularity, if a board comes up to be problematic it causes every use of the board to need revising, in other words, if we find a problem in the design while debugging or revising, we would basically disable every problematic component across every board and nailing down where the problem is in a complex board would be harder as well |
Testing We would only have to revise and test one board design as opposed to many board designs | Software Logic would need to be used in order to differentiate between the different uses of the board, so when its used with the pedals and when its used with the lights. Again, this increases complexity. |
Easier updates/Iterations If we do need to update the board architecture after a revision or different design have been made, it would be easier to update 1 board rather than 2 | Wasted Resources If the pedals and lights have different functional requirements (more on that later) we could end up adding unnecessary components on the board that are only relevant to one use of it, this adds value and again complexity for the sake of easier manufacturing. |
Cheap Again, ordering bigger batches of would turn out to be cheaper for us rather than ordering multiple boards with different designs, we just order a bunch of boards with this 1 design. | Size The daughter board has now been updated to also include the power and CAN which would mean we would design a board with power, CAN, and relevant components for 2 different functions. This will more than likely cause for the board to be a bigger size. |
| Design Work/Ownership Multiple people working on this board would force them to coordinate working together during the week as well as during workdays. Communication can vary from person to person, issues could arise |
Multiboard Design
2 Designs → Used as daughter boards for respective peripheral SOM packages
Pros | Cons |
---|---|
Tailored Design Each board would be tailored to include the components and perform the functions needed by the lights and pedals. | Testing We would end up testing and reviewing 2 different board designs, instead of just 1. |
Reduced Complexity Since each board would be focused on including the circuitry that is needed for that specific function then the board themselves would be less complex than a board with circuitry for 2 distinct functions. | Iteration cycles while it would beneficial to review less board designs (unified board design) we have to consider reviewing/testing 1 big, complex board vs 2 distinct, smaller board. Which do we think is more beneficial to us in the long run? Iteration cycles are more of a neutral point. |
Modularity This allow us to separate the different functions, once we have a preliminary design done, it is easier to build upon it, if a problem were to arise, we would be able to localize the issue quickly as we can debug/test each module individually. A problem in 1 board wouldn’t disable the other. |
|
Smaller size Tailored design would allow us to manufacture boards that are more compact in components and therefore more compact in size. This would save space and money with creating boards where some components would not be used. |
|
Revisions Separating each board would allow us to pinpoint revisions when we have to make changes. The inputs on the pedals board don’t work? We simply revise and replace that board instead of having to replace every single daughter board (unified design). |
|
Design Work/Ownership Separate, simpler boards allow us to divide the work better between different team members, allowing a member to have “ownership” of a board and become an expert in it. We can still work together, but being responsible for your own board forces you to actively be involved and do work yourself. |
|
Summary
While the Unified Board Design offers simplicity w/ production and manufacturing once the design has been finalized, the Modified Board Design approach offers more modularity and efficiency when it comes to the specific functions and roles within our system architecture.
A good starting plan for the peripheral daughter board design cycle would be to pair people up, assign 1 board design to 1 person, and while they both “own” their own board, they must be able to become experts in that board and be able to communicate and relay that knowledge to the other. The pair can review the other persons board and question design decisions before boards are submitted for review.
Questions that might arise!
Isn’t more work to make more boards?
One might think that we are creating “more work for ourselves” with the Multiboard Design because we would be designing, reviewing, testing, and producing MORE boards, but overall it is the same amount of work and effort to design, review, test, and produce a singular complex board and doing the same for simpler boards
If standardization works well for the SOMs, why are we ditching the idea here?
The standardization of the SOM systems works because of the sheer amount of common components and functions most electrical systems need to perform their work, but this standardization does not benefit us within the system specific daughter boards because we want to be efficient with size and resources as well as have modularity.
Welcome to the University Wiki Service! Please use your IID (yourEID@eid.utexas.edu) when prompted for your email address during login or click here to enter your EID. If you are experiencing any issues loading content on pages, please try these steps to clear your browser cache.