Hello, my name is Robbie Valentine and I'm an Electrical Engineer. I spend most of my engineering time developing embedded hardware and software. I enjoy designing electronics and seeing development cycle start from scratch and end with a built usable proto-type.
*Note: in the video I mention this project having a 60 day funding duration. Since recording I've decided to change that down to 30 days. Therefore, this project actually has a 30 day funding duration.
I would love to eventually develop an electrical design company consisting of both products and design services. I've embarked on my journey of satisfying a need I've found in industry, but it requires a prototype/development platform so I can do my R&D. This development platform is the heart of this project and it is a USB development board. I'm designing my own because I couldn't find anything out there that met my expectations well. I need a board that is very expandable, very powerful, versatile and especially low cost. At least low cost for the amount of power you would be getting. I also wanted a lot of IO to provide more freedom to the designer and an expandable platform for any kind of embedded processor development, not just my particular project at hand.
This led me to where I am now and this project. I will be designing and developing an embedded USB capable development board based on Texas Instrument's MSP430Fx5xx and x6xx line of micro controllers. The hardware will support both processor families allowing for a versatile platform in both processing horsepower and price. All parts will be of the LQFP100 package. This is a 100 pin package which provides plenty of IO and functionality to the user. It will have two large expansion headers allowing for a plugin card. This is very important because it allows you to develop a cheap plugin module with your sensors etc... and you just plug in and go.
For example, I am working on developing a certain type of data recorder based on the MSP430Fx5xx line which will be USB based. I want to develop the firmware, communication links and USB software before I worry about adding all the sensors. Once I get the foundation (processor firmware and computer software/comm. link) developed I can then spin a simple plugin card with all of my target sensors, memory chips etc...plug it in, and away I go. Now, I've got a working foundation and can continue development targeted specifically at the sensors and other electronics I've chosen. If I want to try out some new/different chips than I can just spin another cheap PCB supporting these new ICs, plug it in, and I'm off.
So, what exactly "is" this project? This project is a development effort with hardware deliverables. I will be designing a development board described above as well as fabricating them. I put this up on Kickstarter because I think that there might be some interest in seeing an embedded design come to life and seeing how it comes to life. At the end, I will have usable circuit boards, but the journey through developing could be very interesting to some out there as well.
This project is not a back of the envelope project, but has been started a little bit. The parts selection and system design has been thought through as well as research into stencil production and quality of fabrication. I've also modeled the board and some of the parts as well so you can get an idea of dimensions, population look and feel etc...
Below is a simple block diagram of the development board. This is just something I put together as a visual aid so you can get an idea of the system we are working towards. As you can see, Most of the GPIO (general purpose IO) will be run out to the expansion headers with the core of the system being the micro, the USB block and some simple user IO such as LEDs and buttons. Keep in mind that the point of this is a development platform for firmware (the C code for the micro) so I really don't want a lot of extra parts taking up real estate that I may not need. I chose LEDs and buttons because they are very good indicators and event triggers which are prevalent in almost every embedded design.
Here is a 3d model I've put together. Of my current board design and technical strategy. These components are very similar to the ones I'm planning on using so far.
I also wanted to add a small logical design flow sequence so my plan forward is very clear. I'll lay the large pieces of this project out in the order I plan on executing them.
1. Design Ideas, functionality requirements. (these are pretty much finished. My main requirements or needs are mentioned in the video above. This includes expandibility, comm. links etc..)
2. Parts selection, BOM development. (This process is arguably always on going for designers, but I have determined most of the parts I plan on using with a large degree of confidence)
3. Real estate requirements (how big can this design be and why)
4. Build Library/Parts
5. Schematics capture
7. Tweaking and aesthetic mods to PCB (this would be lining up parts for neatness, silkscreen labeling etc...)
8. Rev.0 prototype board, parts and stencil order
9. Build up rev.0 prototypes
10. Debug, make mods, mod wire in fixes
11. Schematic/layout fixes and modifications. (apply mods above to the design)
12. Rev. 1 board/parts order (this is done when all current bugs are solved by the above white wire or hand modifications)
13 Build up units, exercise functionality
14. Ship products
Risks and challenges
Fabrication/Assembly - Since I will be using very small surface mount parts I will need a stencil to actually apply the solder paste and then reflow the boards. I have the reflow capability now, but I will have to find a place to provide me with proto-type stencils.
Current strategy - I've already been in contact with a group that can cut kapton stencils with a high precision laser that more than meets my mechanical requirements for layout. I think this is a solid path forward, but have not personally used them so there is still risk involved.
Timing - this development effort will be challenging because I'm using small parts and a larger complex type of PLD (micrcontroller). For stats on one of the chips I will use during development look at the MSP430F5628. These are somewhat complex devices and any requirement for a re-spin of the board will delay me roughly 2 weeks. I'm anticipating two rev's for this PCB which calculate out to about one month plus time for debugging.
Current strategy - I plan to spin two boards using the first one as my mod board. This board will be primarily for correcting any initial mistakes. Once I make the necessary mods (hardware changes) to the PCB in my electronics lab I will prove they work and modify the design for my rev. 2 board to incorporate these fixes.
Fabrication - Fabrication is another risk I would bring to light. This is heavily dependent on the quality of the stencil and accuracy of the laser etching. It is also dependant on the ability to properly place and re-flow the components once the solder paste is applied. This will be challenging because of the fine pitch (space between leads on the MSP's LQFP100 package) as well as the tight placement of components. Remember, this is supposed to be a fairly small/dense design/layout to help keep costs down by keeping the prototype PCBs small.
Current strategy - I have already done some experimenting with my re-flow system using small components and IC packages and saw very good results. I flag this as a risk because I haven't re-flowed a package with this many leads that are this tight.
This reward provides access to project updates, my design files and a blank PCB that will be mailed to you as sort of a souvenir. Remember though, you will have my BOM (parts list) so you could always build this up!