Recent Changes

Thursday, August 25

  1. page Hybrid Proof of Concept Device (deleted) edited
    5:20 am
  2. page Phototransistors and Testing (deleted) edited
    5:19 am
  3. page Hybrid Proof of Concept Device (deleted) edited
    5:19 am
  4. page Phototransistors and Testing (deleted) edited
    5:19 am

Tuesday, August 16

  1. page Other Sensors edited ... Why Don't LEDs Work? That is a great question, perhaps one that I can't entirely answer. Base…
    Why Don't LEDs Work?
    That is a great question, perhaps one that I can't entirely answer. Based on two recent publications - "Very Low-Cost Sensing and Communication Using Bidirectional LEDs" (Dietz et al, 2003) and "An LED-only BRDF Measurement Device" (Ben-Ezra et al, 2008) - using LEDs as sensors has been confirmed to work. So why have I not gotten better results from my LEDs?
    RadioShack LEDs.

    back to
    production costs.

    I'm saying
    old stock.
    The first instance of LED sensing occurred in the 1971 by Forrest Mims III, designing a sunlight photometer using LEDs. I'd like to get some vintage diodes in the lab and test them before I end my research, just to see how they different from new diodes (both physically and in terms of sensing abilities). I picked up a set of two Western Electric LEDs (seen here) that probably date back to 1975 or so. They were used in telephones, replacing incandescent bulbs and allowing the whole phone to be powered from only the phone jack. They were among the first LEDs that were mass-produced in the US (a few years behind the Monsanto MV1 and MV2). When these LEDs arrive (likely next week), I'll be sure to post test samples for a set of LEDs that have likely not been used as sensors by anybody. 
    Looking Ahead
    With the initial scope of the project changed, it's time to rethink things a little. The original concept for this project was to create an imaging device using only LEDs - both as sensors and as emitters. However, now that we've found that modern LEDs generally make very poor sensors, it is time to explore other sensing options. It is important to keep in mind some of the other project objectives - create an materials-classification device that is portable, fast, accurate, and inexpensive.
    Photodiodes were initially an option, but I've ruled them out because they are not "inexpensive" - many photodiodes cost at least $5 per diode. While this is "cheap" in scientific terms, I think a better solution can be reached. Some searching today revealed the use of photoresistors as light sensors. Photoresistors vary their resistance based on ambient light conditions, and we can get a sense of ambient lighting in an area by measuring the voltage drop across the photoresistors. Unfortunately, though, these photoresistor-based systems do not seem to be highly accurate; this does not comply with the requirement of an "accurate" device.
    Eventually we came across another "photo"-prefixed device that seemed very appealing - a phototransistor. Phototransistors are very similar to LEDs, except they're actually designed for sensing. When we were sensing with LEDs, we reverse-biased the diode to use it as a (poor) sensor. Phototransistors are designed to always run in a reverse-bias mode. Even better for this application is the fact that phototransistors are available in common LED sizes - such as the T1 3/4 (5mm) round size that we've been testing on. So unlike some of the surface-mount diodes or hexagonal-shaped photoresistors, phototransistors would be easy to mount in a hemisphere using only a simple drill. I've put in an order for a set of Vishay Semiconductors TEPT5600 phototransistors.

    (view changes)
    7:21 am
  2. page Coding the Dome edited ... Revisions (II) With the help of some wonderful people in the Arduino forums (see http://ardui…
    Revisions (II)
    With the help of some wonderful people in the Arduino forums (see, I have solved the issues I was having with the averaging system yesterday. Now, the Arduino serial monitor prints a single output - an average of ten readings - for each sensor with every given emitter. This simplifies the output of the system greatly. I hope to implement standard deviation here too, as that will provide some useful data as to how large the variations in the system really are. Things will be looking even better when I get enough LEDs that are actually useful sensors - then I can fill my new system with good LEDs and begin testing its outputs.
    Revised Setup for Phototransistors
    With LED sensing determined unviable in this scope, I will have to redo some of the code to use phototransistors and use LEDs as sensors. This may be some work to redo, but it will greatly simplify the code! Instead of using 30 or 40 lines of code for the sensing procedure with LEDs, I can perform a sensing procedure with phototransistors in less than three lines of code. This will greatly simplify the sensing process and (hopefully) speed things up. Phototransistors are much faster sensors than LEDs (because unlike LEDs, they were actually designed as sensors), so the sensing/emitting process for the whole Mega board will likely be significantly faster. Emitter LEDs will be used to emit light while the sensors determine how much light is in the hemisphere at that point, and the emitters will be driven from the Arduino's digital pins.
    With this setup, I can still run 16 sensors as I had before - 16 phototransistors for 16 analog pins, and a variety of emitters from a huge number of digital pins. The Mega has an enormous number of digital pins, and it only requires 1 pin per LED now, since the LEDs aren't sensing. Cathodes will be connected to a given digital pin, and the anodes will be connected to a common ground. Thus for n LEDs, I will only have to account for n + 1 pins on the controller. 
    Phototransistor Code Complete.
    After coming back from lunch and doing some other work, I decided to take a shot at reworking my first set of code into something that could be used for phototransistors. It is far simpler than the LED sensing code, requiring only 115 lines of code as opposed to 187 for the original LED sensing/emitting code. Part of this is because I was able to eliminate two enormous arrays (15 by 16, for a total of 240 integers per array). This simplified code should be much faster as well as more memory-efficient on the Arduino. The code can be downloaded below:
    Now it's just a waiting game for those phototransistors to see if they work and if this code works!

    (view changes)
    7:20 am
  3. page Phototransistors and Testing (deleted) edited
    7:14 am
  4. page Hybrid Proof of Concept Device (deleted) edited
    7:13 am

Tuesday, August 9

  1. page Test (deleted) edited
    1:15 pm
  2. page Test (deleted) edited
    1:15 pm