Category: Design

Our initial supplies finally arrived a week ago, and I finally got to dig into the development for real. Below, you can see a photo of the supplies. They are organized by type: on the left you see the sensors, in the middle the micro-controllers, and on the right the communication platform (more on that at a later date).

I decided to start with the Hall Effect sensors since I had never used them before and because they felt very promising. Each small bag on the left contains 1 sensor, so I had 5 to work with in all. I began by simply plugging a single sensor into a breadboard and wrote code to get that working. At first, I struggled to get the readings from the sensor to make sense because they weren’t reading reliably. After a bit of research, however, I discovered that adding a 10K pull-up resistor was the trick, and then the sensor worked beautifully.

Next, I confirmed that 2 sensors would work in close proximity to each other. That went very smoothly and I discovered that the magnet has to be directly in front of the sensor to set it off, which can work well for our purposes, as long as we line up the sensors directly where the piece of metal that rest on top of the products sits.

Next, I wanted to get the sensors into the dispenser itself. This required soldering long wires to the end of each sensor pin to allow them to be farther away from the Arduino and to free them from the breadboard. I also added an LED to each connection which would turn off while the sensor was being triggered so that I could more easily see the state of each sensor.

I then mounted the sensors to a long piece of recycled cardboard, which allowed me to more easily move them around and to regulate the amount of space between each sensor. This version of the prototype has a lot of wires, as you can see below. One of my goals for the next week is to work on the wire management aspect of this design. The horizontal lines on the cardboard are roughly the size of one pad box and the sensors are placed every 5 lines, except for the bottom one which would signal an empty dispenser. If there is a case where the dispenser is ever filled but not completely loaded (thus triggering the top sensor), this could present a problem for measuring the amount of product and it will have to be studied further.

Finally, the sensor can be see in action. Notice that in the first photo, the green light on the right is on, while in the second photo it is being triggered by the sensor and is off. A video of this can be found here

Overall, things this week went very well. The biggest lesson I learned was that I need to plan more for issues around wire management, because it can quickly get out of hand. I think now that designing the casing for this project is going to be more important than getting the sensors working. In that vein, I have begun looking into some basics of 3D printing and boards that will allow me to more permanently mount the components in the hopes of designing a robust container.

Kiley, Allie, and I met with Lilian de Greef and Edward Wang, both PhD students in the Ubiquitous Computing lab at UW, about possible sensors for tracking product levels in feminine hygiene dispensers. We began by showing them pictures of the dispenser we’d recently disassembled and they immediately offered a variety of options, which were then vetted by the group:

• Magnetic Sensor
• IR sensor
  • Sonic – something from the top
• Circuit inside, resistance changes
  • More material you cover, the more resistant it is
  • don’t use dangerous amount of current! (what if the person servicing the dispenser is wearing a metal bracelet?)
• stretchable resistance, electrical resistance changes as it changes
• Sound sensor
  • how many times was the coin operator turned
• at dispenser – switch bend (momentarily)
  • momentary switches
• two magnets, one on the top, one at switch
  • they will interfere with each other, so probably not a good idea
• sensor mounted at top moving toward
• magnet – not work because of cross talk (proximity between tampons and pads)
• *absolute sensor
• TDR – know what part of wire you’re pressing on
• slider – resistive material – more material more resistive
• long linear potentiometer (like a dimmer)
• multiple switches as it goes down
  • could approximate # of tampons from that
• IR proximity sensors array – calibration/threshold for tampons
  • covered → something there
  • not covered → not there
  • uses “lasers”
  • whole array is 2 types of signals → make threshold between them
  • make sure to calibrate with door closed
  • but we might want to know what the signals are when the door is open and closed (so we know when it’s serviced)
• can tell if door is opened or closed (serviced)
  • complete circuit → the door is closed
  • not complete → door open
• contact sensor
• magnets – complete the circuit
• light sensor – note every time it’s exposed to light (door is opened)
  • not good because we don’t know dispenser would be filled all the way every time
• magnetic proximity sensor w/ coil inductors
  • change in magnetic field
  • draw wire to IO channel on Arduino

Power & communication concerns

• turn on and off by itself – idle and wake up after a certain amount of time
• trigger each time the dial is turned and reads tampon levels
• OR do a reading once when dispenser is refilled, count every time you turn dial
• wifi might be the right route

• less energy hungry
• problem: concrete bathrooms (cell signals are usually bad w/ concrete walls b/c there is metal inside them)
• cellular will drain power faster
• base station – wifi or cell, plug it in
• solar? phone could serve as communication & power harvesting
• stuck w/ cellular if can’t get access to power
• radio  – no towers to receive it
• use dumb phone
  • with big car battery?
• steal power off lights?

On May 4, 2015 from 3:05 – 3:59, Sarah and I opened up the feminine hygiene dispenser that we had bought from eBay.

  

Below are my (Kiley’s) listed notes as we unlocked the dispenser and examined the mechanisms inside:

– Sarah takes pics of outside

– Use key to unlock it

– The directions on the inside are in English AND Spanish. This seems interesting/pertinent but ultimately not surprising.

– The second set of directions on the inside of the dispenser aren’t very good because there aren’t any pictures.

– There is a crunched up yellow something at the bottom.

– We got it out (it was taped accidentally down there so it took a while), and it seems to be the same instructions as the set of directions on the inside but with pictures.

– We unlocked the coin compartment and there were 4 screws. We think that they are for hanging up the dispenser.

– We thought that there was only one set of the weights, but both of them were just on one side. The pictures on the door of the directions make it clear which went where.

– Nevermind, the directions did not make it clear which went where because I tried to put it in opposite. I feel stupid now because it’s obvious that one is wider than the other and one column is wider than the other!

– Figured out how to spin the red lever things all the way around, clockwise. I didn’t know why they would need to go down all the way, but Sarah pointed out that it’s because it’s how the tampons and pads are dispensed. That makes sense now.

– The pins are kind of difficult to spin and hurt our fingers while we were spinning them.

– We simulated a non-empty and empty dispenser to see what happens with the weight and when you try to turn the lever. As expected, the weight gets in the way of the lever so it’s unable to turn when the dispenser is empty.

– We are missing the levers for spinning (after you put the coins) from the outside and the red flags that say empty. This is the coin mechanism that the instructions on the inside and the crumpled up yellow sheet of paper are about.

– I think there are a couple ways of going about alerting when the dispenser is empty or close to empty. I assume we’d want to alert when it’s close to empty AND when it’s completely empty. Actually, we will want to say exactly how many are left in it.

– Since there is a standardized size for the machine and a standard set of tampons and pads that go into it, we can either:

(1) build a sensor that can sense how high up the products are and then does some math to say how many there are in it.

(2) build a machine that counts how many times the coin mechanism is turned with coins in it.

(3) build a machine that counts how much money has been put into it and senses which product was chosen.

– Besides the fact that we don’t have the coin mechanisms anyway, Sarah pointed out that (1) would probably be the best because it wouldn’t require any resetting by the person who refills the machines and so their jobs wouldn’t be changed.