First draft on the left, second draft on the right
I’d like to address some comments to the prior two posts. The first post was essentially about identifying a problem that could be solved with a 3D printer. In this case, it was the cracked plastic case for my mom’s car key fob. The second post was about one method for measuring and documenting measurements in a meaningful way. Basically I trace out the parts on a sheet of paper and draw in arrows indicating the dimensions.
@Mark wanted to know why I was breaking up the posts.
Well, originally I had intended to live-blog the process of identifying, measuring, designing, and printing the solution. This was not to be for a variety of reasons. Instead I opted to break what would be a very very long post into discrete sections.
@Alex Guichet wanted to know when you’d get to see the results and if this was just going to be a teaser.
I’ve broken the entire saga into four posts. You’ll get to see the finished product at the end. Promise.
Draft, the First
From the photo above you can see the first attempt on the left. There were a number of deficiencies in this draft that I didn’t really think about until I was actually holding the part in hand. In my zeal to measure the electronic-beeper-thingie and the metal key, I had completely forgotten to measure or design the less critical part that allows you to put the darn key on a keyring. Next, the metal part that is the actual key is quite a bit thinner than the electronic-beeper-thingie. As such, there needs to be a raised section that is only “key”-width rather than “beeper”-width. In this first version you can see that there is no such raised section. I also failed to include a hole for the small screw that holds the key in place and both sections together. In this draft the two parts do not actually fit together, they are both just flat pieces that do not interconnect in any way. I also noticed that I made the entire enclosure slightly too big vertically. Just one more thing to tweak in the next version.
Once it was in my hand, I decided that the next version would allow a portion of one part to fit into a section of the other. Just looking at the finished product was enough to help me realize all of these deficiencies – that’s why I didn’t even bother to clean the part up. 
Onwards and upwards!
Draft, the Second
From the photo above you can see the second attempt as the two parts on the right. While these parts were much more suitable than the prior, there were still some issues to work out. I included a section for the keyring, a hole for the screw to hold the key in place, a lip on one side and a bevel on the other so they would fit together, and I reduced the interior height by 2mm to reduce the extra room. The parts fit together well – but came apart relatively easily. I decided the next version would have an even larger lip. It turns out that a reduction in 2mm was too much, since the electronic beeper just barely kept the parts from fitting properly. Increase by 1mm. I decided all of the buttons needed just a little more clearance on each side, so I gave them all 0.4mm more space. I made created a recessed section for the screw head to disappear flush into the top section once installed. Lastly, I also widened very slightly the two walls on either side of they key, since it was a very tight fit.
The benefits of drafting
Imagine you had to program by punch card. Get one punch wrong and that card was toast. Get one punch card wrong and your whole program was toast. Sure, you could double, triple, and quadruple check your program by inspecting every punch and the order of every card… But it’s not a very efficient use of time. For modern programmers there is little penalty for just taking a chunk of code and firing it up. The most common worst case scenario is an error code directing you to where the problem might exist.
Now, if software guys can try out a draft of code without any significant penalty, why not hardware guys? I don’t need to get everything exactly right the first time – and I might even find a few improvements along the way. With an investment of about 20 minutes of unattended printer time and $0.20 of plastic each I got to try out two interim designs – and improve my final design.
Up next: The final results!
Posts in this series:
