18 May Ivette – Build Log 2 – Case Build

In the interest of actually getting this done, I will try and keep this as brief as I can. I might, however, elaborate at a later stage.

Day 0

After much waiting the day had finally come, when the parcel collection card for my screen showed up in my mailbox, meaning I could finally start the build. I picked it up from the post office straight away, and un-boxed it on my lunch break at work.

I also tried out some ideas for possible internal component layouts, using the provided cables. In doing so, I found that I had to have the HID USB cable sticking out of the case in order to have the rest of the RPi’s ports be accessible from the outside. While this was inconvenient, there was not much I could do about this without doing some heavy modifications to the ports, which I was not ready to do. So I would have to live with that slight issue for the time being.

Day 1

The actual case build started the following day. I selected a few good thick pieces of cardboard, with as few dents and damage as possible, from my stash of old boxes. I then traced out the edge of the screen with a felt tip pen onto one of them, and cut that out as a template. That template was in turn used to trace and then cut out 6 case layers (which I will refer to as Layers 0 through 5).

With the case layers ready, I decided to test the screen to make sure it actually worked. But as described in the previous post, I could not get it working at first. But luckily after first sort of getting it to work on my house-mate’s Windows Laptop, I finally got it to work completely on the RPi. Huzzah!

Relieved the screen was working, I continued with the case. I cut out holes in Layer 0 to fit the back of the screen, as well as some holes to possibly fit camera, light sensor, and microphone in the future. Layer 1 was also marked up fit the controller board.

I then had to pack up shop, to get ready for a weekend full of events, followed in the next week by my friend’s wedding up in Brisbane. I did, however, made sure to pack everything I needed for the case in my luggage (except for any sharp objects of course) so I could continue building during any downtime while I was up there.

Day 2

I arrived in Brisbane for my friends’ wedding, and after some initial wedding commitments, I went and bought the tools I was unable to bring with me on the plane, and then set up shop in my friends’ garage/workshop. There I worked on Layers 0 to 2, cutting out holes to fit the RPi, which I had now decided to put on top of the controller board, as well as grooves and holes in Layers 1 and 2 to fit the inner cabling.

The next few days where full of wedding celebrations and other fun events, so I did not work much on the case before heading back home.

Day 3 + 4

Once back home again, I continued by working on Layers 3 and 4. Cutting out holes to fit the Battery and the RPi. I tested the cut layers with the RPi and battery inside which fit pretty nicely. I then tested with the cabling and found that the fit was not as good, so I cut out some more. Then with the RPi and the Battery connected to each other inside the case, the RPi turned on with all the pretty lights shining out from the holes in the back of the case.

Day 5

A few of days later I cut out the final layer, Layer 5, to fit the back part of the battery. And now, with all the layers cut, I needed a way to hold them all together. I was going to glue most of them together, but since I needed access to the inside, gluing them all together would not be a good option.

So one night, a couple of nights previous, when I was pondering the issue, I was going through what I had available in the house, which could be used to hold the layers in place. I did not have any suitable nuts and bolts so I tried to think of what was more readily available. I then thought, in keeping with the case’s somewhat recycled theme, maybe some plastic bottle caps could do the trick. But a quick check of the recycling bin proved fruitless. But then it dawned on me, something we did have was used toothpaste tubes (I lived in a sharehouse, so we often had multiples of things). After a bit of shameless bin-diving, I managed to find three tubes, which I deemed would be enough.

I cut the top parts of the tubes off and cleaned out any left-over toothpaste. Then using the base of one of the cut out lids as a template, I marked up and cut out a hole for it to sit in.

Day 6 + 7

I worked out that creating and using a template was the best way to ensure that the holes would line up across the layers. I made a smaller test template first, for just one of the caps, and then a couple of days later, a full-sized one which covered the whole case. I first traced the toothpaste caps onto the template and then used a compass and ruler to make the circles more precise.

Day 8

I had a bit more time on this day so I decided to work on a script to control the back lighting of the screen. This was done as a sort of proof of concept for developing on the tablet. I was unable to get Sublime running on it (since it’s an ARM system), so I had to resort to using nano (I know, I know, how terrible. I don’t know enough VIM to use it comfortably, okay?). Progress was a bit slower than I was used to, but I managed none the less.

I used the documentation provided by the screen’s manufacturer along with the cython-hidapi library to work out how to communicate with the screen correctly. The script ended up working pretty well, except running it would disable the screen’s touch capability for some reason, which was quite an obvious downside. The lack of good documentation and the manufacturers reluctance to answer my emails meant that this issue has not yet been fixed. But my assumption is that this happened because the documented commands where meant for the latest version of the firmware, and since I have not yet updated the firmware, the commands won’t work as expected. I didn’t have the time to debug the issue any further so I decided to leave it until a later date, and continue with the rest of the build.

The script can be found on my github page.

I continued marking up and cutting out holes for the toothpaste-case-screws using the template I had made.

Day 9

I finished cutting out the holes for the toothpaste-screws through all the layers, meaning that the layers where finally ready to be glued together. I started by gluing in the Audio cable with gun glue, and continued gluing the layers using PVA, ending up with two halves. I put the glued case halves under a bunch of heavy books (mostly the whole Song of Ice and Fire series) to serve as a sort of vice, to make sure all the layers would bond properly.

Day 10

The following day I got the two glued halves out from under their temporary book-vice and found that all the layers had bonded nicely. I was a little bit rushed to finish the build that evening, because I was determined to show the tablet at work the next day.

The final step involved the only bit of soldering I had to do for the whole build, which was to create the power cable for the screen. This cable would be created from the one provided with the screen as well as an old spare USB cable I had lying around. I cut the two cables and stripped back the outer sheathing to expose the wiring within. Through the magic of looking things up on the internet, I worked out which of the coloured wires needed to be connected together for the cable to work. Turned out that red was to be connected to red, and black to black (I did say that my electronics knowledge was a bit rusty). I stripped the insulation from the wires and twisted them together and tested the cable, which to my delight worked. I then popped some shrink tubing onto the cable, soldered the wires together, wrapped some electrical tape around the soldered wires, and finally heated the tubing to shrink around the join. A final test of the cable proved that everything had been done correctly (yay!).

With the screen’s power cable finished, the tablet was now ready for its final test. I hooked up all the components on the screen half and then attached the back half. When I turned the device around I saw that it was booting correctly and I was happy to know that it was finally ready for demonstration and general usage.

Day 11

Having completed everything the night before, I now brought the finished prototype tablet with me to work, to demonstrate it for my Show and Tell during the Monday morning meeting. The demonstration was met with impressed bemusement.

And with that, the case build was complete.

Hope you have enjoyed reading this article. It took me way too long to write but I am glad it’s finally done. The next iteration of the tablet (Ivette 2) is well under way, and there will hopefully be some updates about that one sometime soon.