brahm neufeld

EE495 - Capstone Design Project (Work in Progress)

EE495 is the fourth-year design course for the department of Electrical and Computer Engineering at the University of Saskatchewan. It is a two-term, six-credit class which will showcase some of the skills we've developed within our discipline, and us the change to work in small groups to meet the needs of a customer, be it a professor or someone from industry.

My partners for EE495 are Mike Kachor and Brett Baerg. Our supervisor is Dr. Safa Kasap, a professor and researcher specializing in electronic materials in the field of X-ray imaging.

Our project is the design and implementation of a digital photometer device:

Digital Photometer: A Digital Light Meter
To build a digital light meter that would measure the total light falling on a sensor. The meter would measure the illuminance in lux. Such a meter would be used to measure the amount of lighting at home and in factories to find out if there is sufficient light to carry out "comfortable living" or "comfortable work". Is there enough light to read without stressing your eyes?

In addition to home and office use, some industries or professions also require a certain amount of light - a doctor cannot perform surgery in the dark; an cashier cannot accept money in low light levels. The Government of Canada has several published standards and recommendations regarding light levels in certain environments, but they are not widely known about or adhered to.

This device meets a need of providing a low-cost, user-friendly, portable way of quickly measuring the lighting conditions in different areas. Additionally, the device will also take a (very modified) "snapshot" of the colour spectrum in the surroundings. Colour is important in lighting - more specifically, a balance of colour in lighting is important.

For example, a doctor needs at least 3000 lux to perform surgery (in contrast, a cashier should have 500 lux to work accurately). Lux is a measure of apparent (read: visible) light hitting or passing through a surface. It is a unique measurement because it is wholly dependant on a (slightly) subjective measurement, the human eye curve.

In speaking with Dr. Kasap, we identified the following requirements for the end-user:

A display; the user has to see the measurements.
The ability to save and recall values for later analysis.
Measure lux in the 0-10,000 lux range at at least 10% accuracy.
Low power consumption.
<$300 production cost.

Our current solution is as follows:

Microcontroller: We selected the Atmel ATmega644, for its wealth of built-in memory for saving light measurement values, four 8-bit input ports, low-power features, and familiarity.
User Interface: A Newhaven 4-line by 20-character display provides room for a lot of information to be delivered to the user. Four contextual menu buttons are placed beside the screen for variable functions, depending on what the user is doing, and a dedicated "back" button is placed just below.
Lux and Colour Detection: Originally, light and colour detection was going to be the most difficult part of this project. Some careful research and phone calls lead us to Texas Advanced Optical, a company that specializes in light and colour sensors. We selected the TSL2550 for a sensor known to be accurate in industry for lux measurements, and the brand-new TCS3414 for digital color detection. Unfortunately, the TCS3414 was too small to use - at 2.1mm long and 1.56mm wide, we were far too inexperienced to mount it, and the attempts made by the technicians and other staff failed (the department did not have the right equipment for the job). We selected the TSC230 instead, a predecessor of the 3414.
Power Consumption: Our photometer will be powered by 4 "AA" batteries. All of the components we selected were designed to be extremely low on power consumption and run at 3.3V. In a test on March 7th, we measured a current draw of just 10.8mA - this is less current than a single blue LED would draw under regular conditions. with the backlight enabled, we draw approximately 200mA.
Cost: The cost of materials for this device are approximately $50, PCB fabrication and some sort of enclosure would be extra.

Testing & Verification

We verified the accuracy of the lux sensor by comparing it to several other lux sensors available, including a commercial lux sensor. To test the colour sensor, we used a spectrometer to measure the spectrum of various light sources (seven LEDs and two overhead lights) and calibrated the lux calculations to best approximate the spectrum of various light sources.

Demonstration

Updates to come

I would like to update this page (eventually) with the following:

A proper circuit schematic, block diagram, and parts list.
Results of our project and presentation, acknowledgements, and some neat tricks we learned during this project.

Stay tuned!

Last Modified March 15, 2009