me 102a spring 2010 april 21, 2010 jennifer lew stephen tu
TRANSCRIPT
FSD – Fan Speed Detector
ME 102A Spring 2010April 21, 2010
Jennifer LewStephen Tu
Motivation
Fans provide essential cooling to a computer The fan’s rotational speed is directly related to
its cooling capacity Without properly
calibrated fan, CPUwould almost instantly melt
CPU: http://hardforum.com/showthread.php?t=1423185Fan Grid: http://forums.legitreviews.com/about12716.html
How It Works
Idea: Use fan blades as an Optical Encoder By measuring the number of pulses in a fixed period of
time, can determine speed of blades
Rotary: http://hades.mech.northwestern.edu/index.php/Rotary_EncoderFan: http://computershopper.com/shoptalk/components/silverstone-fm83-cram-more-fan-into-the-same-space
Implementation Plan
1. Fan Voltage Supplied Current Supplied Impellers
2. Sensors Infrared LED Phototransistor Circuit: Resistors, Voltage Supplied
3. LabVIEW VI Input number of impellers Implements GenerateCalibrationCurve() Displays RPM, voltage supplied to fan
4. Error Analysis Sensor is noisy, so must take many samples and
consider confidence intervals
Procedure for calibrating fan
IR LED: http://www.solarbotics.com/assets/images/ir-led/ir-led_pl.jpgPhototransistor: http://farm3.static.flickr.com/2325/2105641666_1487ce594e.jpg
Issues and Scheduling
Issues Current limit on power supply – data at higher voltage values not
truly reflective of fan characteristics▪ Will resolve by using analog lab equipment that does not impose as stringent
current limits.▪ However, this makes the UI less robust, due to less automation
Scheduling Basic prototype implemented last week, but ran into the current issue
above 4/26/2010: Have prototype finished, current issued resolved 4/30/2010: Have prototype tested for robustness, finish error analysis 5/3/2010: Present prototype to customers 5/7/2010: Turn in final report and give final presentation