Let's Make Robots!


Vendor's Description: 

Accelerometers are very exciting little things. When I say "little", I mean it, as they usuallyfits 4 or more on a finger nail! And each typically has some 7-20 points so solder. This makes them quite hard to work with, and so I recomend that you also buy a board of some sort, so that you end up with something that you can hook up to your Microcontroller easily.

What you get from the thing is a signal. The signal can be analouge, I2C, or anything else, depending on the thing you get.

I prefer analouge, it is just SO easy to work with; hook it up to an analoge input, read, you are go :)

So what does it return to your microcontrolelr, then?

Well, inside the tiny box is a liiitle chamber with a gas or something. And around this gas cloud or whatever, there are sensors. The gas is lighter or heavier than air or something (this is very precise info, no? ;) - and  the resoult is that you get info about how the thing is tilted, or moved around in the air.

You can get them with 1, 2, or 3 axis, I have not seen them with 4 yet (joke).

You know them from the Ninendo Wii handles, and you can use them for such things as:

  • Balance
  • Registering drift on fast robots
  • Registering if the robot-pet-animal is lifted up, perhaps shaken by the stupid kid
  • Cool "In Air" joysticks
  • To build something that will tell you how fast your car is accelerating.. or your bike ;)
  • Registering if your Sumo-robot is just getting hit by another
  • Making auto-corecting flying things
  • and much more :)

I got mine here:


I just stumpled upon this place that aprantly sells some nice ones:


- and have a nice background info:


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Practical, readily obtainable accelerometers are the ST range of micro-machined range. Here's the 3-axis 6g version.

ST are more than happy to send you freebies "engineering samples" of these little twinkies. They're dead easy to use: you just hook 'em up to an analog input.

AND you don't have to worry about gasses and clouds and litte chambers. These work by having a little micro-machined springboard (like a diving board inthe swimming pool) in, which produces a voltage when it is deflected. This voltage is amplified and sent out to your microcontroller's ADC.

One pain about them: they are CMOS and as such prefer a 3.3V supply when we seem to mostly work off 5V.

These used to be my favourites and if anyone can find a source for them, please let me know. I'd love a couple more of teh 10g PWM versions. Rather than an analog signal, they output a  PWM rather like the one used to position a servo motor. AND they are happier to run off 5V!

PS3 controllers claim to have "6 axis" acceleration sense.

 I assume they mean 2x 3 axis accelerometers?

Don't know what they might mean. I'm not sure I see the point in 6 axes, though. I wonder if they get some resolution enhancement by offsetting them slightly, or something like that?
yeah, that's what I assumed
They sense rotation and translation on each axis (wikipedia).

Ah. I get it. 3 accelerometers for translation and 3 gyros for rotation.

I find it irritating the way the article seems to imply that one of the companies involved invented the technology and the other copied them. In reality, roboteers and hobbiests have been using such devices for years, but without the funding or marketing ability to make millions from it.

I don't really get your gas-stuff (what a surprise ;) ), I think you mean the type in which a gass is heated up, and it is moved inside the container. You can sense where it is by measuring it with a thermocoupler (I got this from wikipedia). A thermocoupler has 2 types of material which have another thermal strain ( don't know the english word ).

But if I'm not mistaken, the most common principle in accelerometers is far simpler and depend on the piezo electric effect.

 A mass on a layer of piezo electric material like so:

  OO   << MASS
==== << Piezo material

The inertial forces of the mass press the material.  You measure the voltage on the piezo material, with some additional circuitry to correct some of the properties of this material.


I can very well imagine this second type to be the cheaper version so you probably got that one.




Mine is aparently the gas-type: http://www.memsic.com/data/pdfs/an-00mx-001.pdf

But I don't care much, as long as I get the output :)

Thats because you are an arty!