Shortly after writing my last post I read an interesting article from Tom Maiorana on 5 Prototyping lessons from a BMX backflip this is an extension of the Fail Early, Fail Fast, Fail Often principle.
This consists of 5 stages.
- Deconstruct the challenge
- Minimize risk
- Use ignored resources
- Remove as much as possible
- Maximize cycles of learning
It seemed a good principle to apply to my new board prototype.
Deconstruct the challenge
When you are looking at your product or service, break the whole into all of its different elements. Then focus on the one that is most critical. For those of you familiar with the Lean Start-up, this would be what Eric Reis calls your “leap-of-faith assumption”.
The main steps of my project are:
- Design the schematic
- Design the PCB layout
- Assemble the boards
The schematic is effectively straight off the data sheets.
Assembly of the boards is something I'll likely need a bit of practice with but the processes for doing this are all fairly well defined, I'll likely hand solder the prototype boards.
This means that the critical step is the design of the PCB layout.
However following on from this is a shift from prototyping to production so it might be worth re-running this process at that point to ensure success.
Minimize risk
As you think about prototyping your offering, brainstorm ways in which you can learn with minimal risk.
For me the key risk is that I'll spend lots lots of time and money and have nothing to show for it. This breaks down into two parts, firstly not purchasing lots of tools up front and making do with what I have. The other risk is having PCBs made and then finding out that they don't actually work. So what can I do to get to a working design quickly without having lots of PCBs made that don't work?
Use ignored resources
Think about the untapped resources that you might have at your disposal? Is there a bored developer who’s looking for an interesting project? A space no one cares about? Find them and put them to use
Looking at the resources I have available I don't have any PCB creation capabilities but I might be able to talk to a local hackerspace and get my boards made that way.
I did think about what I could do before making the PCBs. There's lots of useful information in the CapSense
Design Guide as well as a spreadsheet that allows you to check your parameters. I did wonder if I could use the design rule checks or a Write an Eagle PCB ULP to validate my design such as checking the trace lengths, separation and routing. However given the size of my board and the max trace lengths (e.g. 50cm) this does not seem like good time spent.
I also found a useful article from Sparkfun - Better PCBs in Eagle which suggests tips for ensuring that boards have good tolerances so that shorted connections are less likely to happen.
Another idea I found from Element14 is Using Paper Protos to Check Footprints this allows me to check that I'll be able to place and solder the components before getting the board made and also that the button layout is sensible.
I could also Render 3D images of PCBs to ensure the board looks ok and use a Gerber Viewer to ensure that the Gerber files are generated as expected.
Several of the PCB houses can run their own Design Rule Checks against your board or provide a rules check settings file to download. I may take advantage of the DRC file provided by Sparkfun for this.
Other resources I can utilise are some of the Element14 members, I've had a offer for someone who wants to take a look at my project.
There are also some free electronic circuit simulators List of free electronics circuit simulators - Wikipedia, the free encyclopedia although I'm not sure how I could utilise these in my project unless I can also simulate the MBR3 chip.
As mentioned in my last post it should be possible to re-use the CapSense evaluation board along with EZ-Click software to test and configure the boards. The MBR3 chip has some onboard diagnostics so it might make sense to write a bit of Arduino code to read these or use the testing tools that come with EZ-Click.
Remove as much as possible
For organizations, this often means your prototypes will have a lot of throw-away work. But you’ll be moving so fast that it’s actually more efficient than a traditional way of working.
For my prototype my first circuit is simply one from the data sheet with the 5 inputs and host interupt. I was wondering if I could make this as a breakout board and have the sensors as a separate board. This might give me a little advantage in that I'd not need to solder the components for each iteration although that might be something I'd want to practice any way. I also wondered if I could create a board with a blank area on it and use conductive paint to draw different pads. However I think these might actually add complexity to the project and doing my iterations on screen electronically might be preferable.
Maximize cycles of learning
As you prototype, think about ways to create an exponential jump in the number of learning cycles.
To maximise my cycles it makes sense to do as many as possible electronically using the above resources before sending the board out for production.
Once I'm happy with the design I can do another set of cycles to get the boards working.
Conclusion
As a process I'm not totally convinced that the BMX Backflip principles can be applied here but I think the process of deconstructing the challenge and reviewing resources has been useful.
Top Comments