soil moisture sensor

Part of my ongoing garden project. Sort of inspired by this but manufactured in the "engineer" technique rather than the "maker" technique, using turned engineering plastics and epoxy rather than drinking straws and picture hooks.

This is basically two electrodes sealed inside a plaster-of-paris shell exposed to the ambient media. As water is absorbed the resistance drops. Unfortunately for some reason there is a dc bias, due to some chemistry anomaly or dissimilar metal issue -- a potential of around 145 mv is apparent in addition to the series resistance. My solution -- as most sources recommend for this type of sensor -- is to use an alternating current on it. My prototype so far uses a CD4052 to flip the polarity at a few kilohertz. The multimeter reads zero volts and the resistance is sensible.

Sensor bandwidth is worrisome. When the sensor is submerged in water, the resistance dropped from off-scale (3M+ ohms) to 850 ohms in less than a minute, but it seems to take a very long time to dry out. Possibly much longer than it takes strong lighting to dry out the soil it's buried in. I may need to re-engineer this with a smaller plaster casing with less lexan surrounding it.
[April 15, 2009 1:07 AM]

ads are dumb

This car is a Saab 900T. This is definitely one of the more awesome cars out there, partially because its clutch petal is connected to its clutch with a cable and almost all Americans are too dumb to drive it. Avis is a company that rents cars, and this is a frame from one of their television ads. It's got a voiceover that is basically a personified bitter 900T complaining about her owner having an affair with one of the shitty cars car that rental companies rent, like a ford shitmobile or whatever they call it this year. (I rented one a few years ago, it was called a "focus".) Imagine Heidi Klum whining about her husband having an affair with Monica Lewinsky.

Update: my dad tells me that the 900 clutch was indeed hydraulic. At some level I refuse to believe that the 900 and 95 share the same clutch linkage, though.


[March 17, 2009 12:41 AM]

delrin crash

neat, almost reaction-diffusion-like patterns created when delrin is hit too aggressively on a lathe.
[March 1, 2009 3:57 PM]

sonar positioning system

This is a project that's been sitting on the back burner for well over a year. It uses infra-red light and 25 khz ultrasound to measure distance between stations over a range of maybe 10m, or with at least 2-3 fixed stations, should be able to measure the position of a robot in my apartment down to a few cm.
[February 25, 2009 10:05 PM]

delrin is awesome

It machines like butter, so you can drill it smoothly and quietly right past where you meant to. I think I might start uploading photos of all my screwups and label them "art".
[February 18, 2009 1:36 AM]

brooklyn bridge in hdr



This is an HDR bracket I took of the brooklyn bridge a month or two ago. It's a pretty terrible photo but my mom really liked it, possibly because she hasn't seen enough HDR images for it to be cliche. bigger


[February 17, 2009 12:19 AM]

128x64 graphics lcd

A recent acquisition from SparkFun which came with datasheets written in authentic engrish. I got this for a forthcoming project but for now just spending some time writing AVR drivers for it on a '48. more info
[February 16, 2009 11:33 PM]

fun with airfoils

You know you're a loser when you spend half your long weekend writing code to plot NACA 4-series airfoils and building wing sections. Cutting the ribs with a razor is a huge pain. [reason #99 to buy a CNC laser.]
[February 16, 2009 11:27 PM]

new dro

The old digital readout on my mill, which required being attached to a computer to act as a display, finally crapped out, motivating me to actually build a good one, seen here. Between that and (1) the super-quiet aftermarket belt drive (2) the cool alien sticker replacing the warning label (3) all the safety guards removed, this is probably one of the most awesome cheap-ass chinese mini-mills out there.
[February 16, 2009 11:20 PM]

an obvious idea whose time had come

Low-end AVRs are like $2 these days when you buy in bulk. Unfortunately, the hassle of wiring up the ICSP and RS232 connections required to program and debug them often exceeds the actual value of the part. I've come up with a neat solution to this. I took my latest board -- which takes a mega (48|88|168) -- stuffed only the icsp header, rs-232 level converter, and reset button, and wired the power/programming/serial pins on the AVR socket to a DIP28 test clip. This way it's possible to program and talk to any DIP28 AVR -- on a breadboard, ratshack board, or whatever -- without any additional effort.
[November 15, 2008 3:57 AM]

projects
photos
email

archive:
April 2009
March 2009
February 2009
November 2008
October 2008
January 2008
December 2007
November 2007
October 2007

mt login