Urban Sensing: Using Wifi Signals to Measure Crowdedness

(More documentation to come!)

#urban_sensing 


Energy: Growver

Last week, James Hosken and I wrapped up “Growver”, a roaming solar-powered bot that seeks sunlight to power itself. For our Energy midterm, we had presented an early version of Growver that functioned that was missing an intregration with a key power-saving mechanism I had been working on: a 555 timing circuit.

Here is it in action back then:


It should be noted that Brent Bailey was also working on the project with us at the time, and made valuable contributions in both the fabrication of Growver and its direction-finding code.

For the final, James and I decided to split up the work so that I focused on the 555 and he worked on robot behavior.

I had based my timing circuit on an oscillator tutorial, and need to make a key change to include a relay:


(For those interested, I wrote up some additional information the IC here: About the 555).

For the midterm, we had attempted to make the 555 trigger a transistor, but had into some trouble due to its placement within the rest of the circuit. We consulted Eric Rosenthal, with whom we had taken Basic Analog Circuits, and he had recommended simplifying matters by switching to a relay.

Unfortunately, we still ran into some trouble with the circuit – though the timing compontent worked over loads such as an LED, its behavior became quite erratic when triggering the relay. For my Homemade Hardware final, I had designed an PCB with the timing circuit, and for a great deal of time, believed that the erratic behavior was either due to my circuit design or soldering. I sunk quite a bit of time reconfiguring my circuit in Eagle and reprinting my PCB.

Fortunately or unfortunately, the errors were not due to my actions in designing the PCB or soldering it, but rather, a more fundamental error in my understanding of the relay in my timing circuit.

Because Eric had suggested the relay in the first place, I sent him a desperate email to update him on the behavior of my timing circuit. After a few messages back and forth, he asked which 555 chip I was using: I had poppped in the 555CN because that is what is stocked in the ITP shop, but unfortunately, this is a low-power chip was not providing enough power to trigger the relay consistently.

Thankfully, I was able to find a 555P in Tinkersphere.

I had in passing complained in David Rios that nothing was working in my circuit, and he also took a quick look at what I had done so far. He suspected the magnetic coil in the relay may have been holding a inductive charge longered than it seemed on paper. He suggested that I add a pull-down resistor (in this case, called a “bleeder resistor”) to ground so that the coil could discharge when it was not powered.

The bleeder resistor, along with the new 555 chip, all within the span of a hour, solved the problem that had plagued me for days. The new circuit looked like this:



At this point, we connected the timing circuit to James’s motorshield + Arduino setup, and hoorah, it finally all worked well altogether.

At this point, the timing circuit had a 47k ohm and a 680k ohm resistors, yielding a duty cycle of approximately 1 minute on, 9 minutes off. The timing circuit was independently powered from 4 AA batteries. Based on that, my understanding of the current draw of the timing circuit and the life span of the batteries is calculated below:


The current configuration is less than ideal, but it should be noted that in earlier iterations of the project, we had much larger resistors that allowed us to power on for 3 minutes and then off for 90. Now that the timing circuit and motors can function together, it is a minor change to adjust to a new, more energy-efficient duty cycle.

In the video below, you can see Growver in action. Its light-seeking behavior leaves something to be desired, but the timelapse allows you to see the timing circuit at work.

#energy 


Energy: About the 555

For my final presentation in Energy, I had created some slides to explain how the 555 timing chip worked, but based on the advice of smart people, I decided not to bore my audience with the details during the actual presentation. However, I thought I’d share the slides here (perhaps, mostly for my own sake, for my later reference.).

The hand-drawn portions of the image represent the actual circuit design we used for the Growver project.

The slides are based on my notes from Practical Electronics for Inventors.

#energy 


Soft Robotics: Final Project Idea

A very preliminary quad chart of a final project idea:



And a sketch:

#soft_robotics 


Soft Robotics: Custom Actuator

This week, I decided to make a variation of the mold we used last week. My silicone cast had a lot of room for improvement, and I wanted to iterate before I moved along completely from the base design.

I researched some of the possibilites of how I might adjust that design to yield a different kind of actuation. I was particularly interested in the internal restraints I could introduce to the design. In my research, I came across fiber-reinforced actuators in the Soft Robotics Toolkit. These actuators consisted of a flexible bladder that were wrapped with inextensible fibers, or thread, to constrain radial expansion.

The site suggested that the the following arrangement would result in an axial expansion (rather than the typical radial expansion we see in other balloon-like actuators):



In order to implement this design, I decided to use a two part casting process. Initially, I covered the inner mold from last week in silicone (using less than 5g of mixtures A and B each). I allowed this to cure until the surface was squishy to the touch.



Next, I wrapped thread in a double-helical pattern around the length of the inner mold.



Then, in a fill and degas process similar to the one utilized last week, I placed the inner mold in the outer mold, this time working quickly enough that the mixture was significantly less viscous than last week’s iteration at this stage of fabrication.



I then allowed the whole mold to cure overnight.

[I’ll update with a video, once I’ve tested its behavior with an air pump.]

#soft_robotics