TURRET PROJECT DEC.2 - So I'm at the point of building the fundamental code on my Arduino Uno board, and realized I should specify what my goals are here. This project will comprise two design milestones, as shown below.
Turret Mk I

• Two degrees of freedom
• Manually operated by pushbuttons or joystick
• Airsoft gun and laser site mounted on end effector

The second milestone will include everything from Mk I and the following:
Turret Mk II

• Camera attached to the end effector, and linked to a computer for target recognition
• Computer/software (most likely a custom program that uses the OpenCV computer vision library) to automatically manipulate the turret and "trigger" the airsoft gun, with no operator interaction

There is still quite a lot of work to be done, but Mk I is feasible in a very short time frame depending on how quickly the mechanical assembly of the turret can be completed. I'm probably going to buy an off-the-shelf pan/tilt servo assembly to expedite this, since it'll be simple and cost effective.

TURRET PROJECT DEC.1 - Allright - so its obviously been months since I've touched this blog. When I "started" this project, I was working 11 hour days and flying home every weekend, essentially making me work a 65 hour week. The last thing I wanted to do when I got home to my beach house (yes, I happened to be consulting on the Atlantic at the time) was nerd out with a soldering iron.

Now that I'm back home and have more free time, I purchased an Arduino Uno and plan to use that as my main microcontroller board for the turret. I have never used one, but I have experience with Parallax's BS2, and plenty of programming experience... so I don't expect much trouble coercing the Arduino to do what I want it to.

I'll be updating more frequently, starting with a breakdown of my plan, from here on out.

TURRET PROJECT JUNE.1 - The objective of this project is to create a stationary weapon that is capable of autonomously recognizing "targets". I plan to use a webcam to collect visual information, OpenCV to process the information, my laptop as the brain, servos controlled by a micro controller to position the weapon, and an electric airsoft gun as the weapon. I have had previous experience with OpenCV from a computational perception class, so making software for the turret to rudimentarily recognize what it is "seeing" shouldn't be difficult. Additionally, I have programmed micro controllers to make fairly complex digital circuits, like a digital tachometer and a line-following robot. Other than those bits of experience, this will be quite the ride for me!

Work Project #1 - Failure analysis of a new product

After finishing my freshman year of school at Iowa State, I came home for the summer of 2006 to find my services at the local Circuit City were no longer needed. Instead of settling for the first minimum wage job I could find, I found a job through an employment agency. They matched me with Joule Technologies, a small company which specializes in test solutions for printed circuit boards and other electronics. As it turned out, my low-level temp position at the company turned into a much greater employment opportunity down the road, and before that summer had ended, it was agreed that I would return as a full-on mechanical engineering intern.

Internship duties during the summer of 2007 included a lot of machining of piece parts, I became very proficient on mills, lathes, and operating 2D CNC rout/drill machines. Essentially, I was aiding the engineers in the fabrication and assembly of their test fixtures (side note: the test fixture industry has an extremely short time window from P.O. to delivery of the product, sometimes less than a week).

A final and important duty for me was to design a means and procedure to find the expected cycle life of a custom handle that they created in-house for their test fixtures. Up to this point, they had been ordering another company's handles for every job, at a significant markup in price and ultimately subpar quality. The goal was to improve product quality and value for the customers while reducing cost to Joule Tech.

The physical solution I came to was a PLC controlled robot which would pneumatically actuate the handle, open the fixture lid, close the lid, then release the handle. I designed and fabricated the "robot" using autocad, our 2D CNC machine for delrin components, and a mill for aluminum components. Ultimately, the fixture was cycled a million times and it was determined that the internal mechanisms of the handle would statistically be expected to begin to wear after about 50,000 actuations, far too soon for the product's intended lifetime. As an aid to the statistical analysis (which was done well before my machine design class, for the record), the wear was documented with a 50x magnification microscope/camera using both aluminum and steel internal components in two different trials.

The testing done was helpful to the company, since they had shipped several fixtures to customers without doing any failure analysis on the product. They knew which type of internal components to provide to customers who already had the product in case of premature failure, and they knew what improvements needed to be made to the design before selling any more.