Slashdot Mirror
Harvard Professor Creates Paper Accelerometer
SuperSlacker64 writes "In an age where just about everything starts going digital, it's refreshing to see someone going back to our roots: paper. Well, sort of. Researchers at Harvard have created a cheap, dime-sized, paper-based accelerometer that they believe could be used in various ways, such as inexpensive medical testing. The device works because a carbon bridge stretches and changes resistivity as the device is accelerated."
When they say "cheap," they mean it; the cost per device is estimated to be about four cents.
I thought professors were people who couldn't hack it in the real world, but it turns out academia is behind most innovation in one way or another? Color me retarded.
Everything works on paper!
One of our competitors trademarked the term "hypothesis". From now on, we will call them "boneheaded ideas".
I hath sinned. I read the article. *hangs head in shame* The paper counterparts are far less sensitive. Silicon sensors give about 80 micronewtons while the paper give 120 un.
80 vs. 120 micronewtons isn't too bad. If by 'sensitivity' you mean the expected standard deviation for measurement noise, and assuming such noise is roughly gaussian, then you can almost achieve the precision of a silicon sensor by using two paper sensors and averaging the results (120/sqrt(2) = 84.9).
Throw together 25 of them (for a total cost of $1.00) and you can achieve 24 micronewton 1-sigma precision.
It says for use in medical testing equipment where I'm pretty sure one would want precision, no?
Not every test, especially when aiming for low-cost, needs to reach Starfleet standards.
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
many applications require high-precision measurements, not just good approximations
And your point is...?
For any application that needs high-precision measurements, there are many others where lower precision would do fine, if only it wouldn't cost so much.
Most accelerometers are not just a sense element.
The Analog devices ones from a few years back included an onboard reference (to allow single supply operation), and an on board buffer amplifier with externally settable gain (i.e. integrated op-amp). The 4 cent sensor still requires precision op-amps, nulling trim pots, etc to get a usable sensor. Don't forget packaging to protect the element from kinks, moisture, etc.
BUT, the most important thing about the accelerometers from Analog I used was they were laser trimmed for 0G, and used a very cool sense mechanism. The sensor was a micro-machined silicon mass on springs with a capacitive force/sense system that detects a perturbation of position by using a 1 MHz AC pump signal with a capacitive bridge and synchronous demodulation feedback. The mass is forced to be stationary by applying a DC electrical voltage on the capacitive sense plates, imparting a force 100% proportional to the applied DC voltage (i.e. purely LINEAR voltage to force relationship).
So what? Their system keeps the mass centered, thus making micro-machined springs that hold the mass irrelevant to the output voltage. In other words the NON-LINEARITY of crap springs is servoed out, and the output voltage is not an open loop measurement with gain, but a true measure of the force being applied to the tiny little mass by the acceleration trying to be measured.
Paper as a spring is not going to be repeatable from batch to batch, is subject to moisture, subject to fatigue, hysteresis, etc. Mad props for new uses of paper, but practical as a replacement for the devices it mocks it is not.
Commenting to undo unintentional mod.