yellowbkpk

Hi everyone from Makezine Blog! Very happy to see you and hear your comments.

Here’s an update on where I’m going:

• We went to University of Iowa’s surplus store and picked up several small computer microphones. The guy there suggested piezos (similar to several comments that have rolled in) as sensors, but cautioned that the different sized piezos are sensitive to different frequencies.
• I’m looking harder at doing a current clamp to detect current flow in to the washer and dryer. This would eliminate a lot of the electrical noise picked up by the microphone (but would probably pick up other kinds of noise).
• Other options for data include temperature and vibration (as suggested by comments).

After speaking with Andy about reliability of the data using sound, I decided that we need to determine another option to collect data. One option Andy suggested was to use a current clamp to determine how much juice is going through the power cable to the washer or dryer. This should vary depending on the cycle of the washing machine and potentially the time in the dryer (as the dryer heats up, it would use less energy).

Here’s a website that describes how to create a current clamp: http://www.siliconchip.com.au/cms/A_30669/article.html

Once that is created, it would be a simple circuit to measure the voltage that is created from the circuit.

Tonight I recorded my first small tid-bit of data. There were only two dryers running here in Merner, so I recorded one of them. I started recording when it was at 41 minutes and ended when it was 36 minutes remaining. I tested using my Archos MP3 recorder in its highest-quality mode. I made two recordings:

• The first was made with the unit sitting with the screen (and thus the microphone) facing up. I recorded for about 30 seconds.
• The second recording was made with the microphone facing down on the surface of the top of the dryer. I recorded approximately 40 seconds worth of sound on this side.

Once I made these recordings, I transfered them to the computer and loaded them up in to an open source sound editing application called Audacity. Audacity has a wonderful little feature called “Plot Spectrum” that I used to perform a quick-n-dirty spectrum analysis on a small “average” portion of the sound. By average, I mean a portion of the sound that did not include any very loud peaks (in this case, the loud peaks were the sound of some hard object inside the dryers banging onto the metal drum). I was hoping that the “average” portion of sound would only include the rumbling sound from the motor, which is really what I wanted to record in the first place.

Below is a spectrum analysis of a portion of the first recording (note the logarithmic scale on the X-axis):

If you click on this image, you can get to a larger version and see that there is a very pronounced, short peak at 581Hz, which corresponds roughly to the sounds I was hearing with my ear.

Below is a spectrum analysis of a portion of the second recording. I had Audacity show me that the peak is at 360Hz in this recording. 360Hz is a multiple of 60Hz (which is the frequency our lovely American alternating current runs at).

Also worth noting is the relative quiet space above 3KHz, and the dead space above approximately 20KHz. This might be the edge of the frequency response for the microphone that is built in to the MP3 recorder.

Hopefully I’ll be able to find someone doing washing later and gather some more interesting data.

This coming block will be an independent study for me. I’ll be working with Andy on a project (or two), and I’ll be using this blog to post about it.

My first project is to use sound to determine the current cycle one of the school’s washers or dryers. The idea being that I should be able to determine how much time is remaining for each device and publish it to a web page. A student can then use this information to determine what the best time to wash their clothes is.

I have to start by using my MP3 recorder to record the sound of the washing machines and dryers in action and determine if I can determine any difference by ear and by spectrum analysis. Should be interesting!

Stay tuned for more.