More progress, less time…as things likely go.
Kerry and I got travel grants to go to SIGGRAPH, which is very helpful. So now we can go ahead and start booking our trip…A month and a few days to go!!! Work days are getting very long at this stage ie. 12+ hrs. I want to take breaks when a natural lull comes about. Last weekend went to Kerry’s Dad’s cottage because I was waiting on parts. Was the recoup I needed. I hope to go out to Toronto Island with Kerry this weekend, and maybe drinks with a friend at some point. Mostly though I’m holed up in my studio. It’s a gorgeous summer, but I miss many days. That’s okay, because it’s my choice.
I got the second draft of the shell back, as well as the new circuit, and have been testing those over the past few days. I made some minor adjustments to the shell, and yesterday sent off a third and final version off for printing. Why is it final? Because I got 10 printed at a cost of $1000+. I’m “all in” at this point. These better work or I’m screwed. One thing I’ve noticed this past week or so is that the quality of the work has improved. As you can read in other posts, twice I found myself on the wrong track and so tried to refocus on what I thought was critical, the object’s design. As a result I’ve made time to work and rework the model, and the improvements are tangible. The shells fit together well, the mechanics and electronics are more robust, and the circuit has more features which work, and will be used. I expect this version to be a significant improvement on the first.
The two most significant things I changed about the second draft of the shell is that I provided reliefs for a power switch, so that power can be totally cut while assembling, and for the programming header so that the module doesn’t have to be disassembled in order to reprogram it. The second change was inspired by the first. Realizing that having the innards totally inaccessible when assembled could cause issues. I might need to poke around inside, or observe what’s happening. So I added some holes on the bottom of the shell through which a few tools will fit.
This week, I also tested that the module would properly fit everything. It was tight at first, so I ripped out some inconsequential features of the shell that were getting in the way, and now it fits together better.
I also widened the latch so that the modules could more easily weave together.
As I explained before, the new circuits implement zigbees and ATMEGA644 which have two serial ports. Other than that there is not much on the circuit board. The MCP73862 li-poly recharger works well as was tested this week. There are things I have to change about this version as well, but they are minor. I am planning to add a zener/mosfet shutoff switch so that if one plugs in a higher voltage (say 12 volts) the snake is dead as a doornail. This will be a surefire way to make sure the snake can be turned off completely very quickly.
I also want to note that the original reason for adding the zigbee was because exchanging data through the slip ring was unreliable. I am happy to say that the second draft of the shell seems to have improved the continuity of connections in the slip ring by making the pressure between the roller levers and the disk even and continuous. While I don’t think it will be 100%, it might be good enough to use, and maybe even more reliable than wireless. This is all to say that now there are options! I can use zigbee, slip ring or both!
Here’s the list of items that are/will be changed about these second drafts:
- higher and wider roller reliefs
- remove servo wings
- add power switch relief
- remove excess material – access from bottom
- change pin to ring connections
- buy some good bolts to connect to motor hub.
- swap rx/tx on ftdi
- add power jumper
- separate xbee from ftdi dtr
- limit power flowing into 644 from stat pins.
- lower stat input voltage to 5v
- add zener cuttoff switch
- add manual switch
- add a jumper to select between serial port 1 and 2 connections to slip ring.