This is exciting. I've lined up my Christmas presents for my wife to give me. First off is the latest Asus Tinkerboard S, which is many times more powerful than the Raspi3. I found I just couldn't get my old r3 to work with the new 'Stretch' operating system. Many people have reported the same thing. The r3 was full of many proprietary things and there are no new drivers. The tinkerboard is way more expensive but everything has open source drivers. It has the same sort of pins and the same Python libraries.
The new accelerometer, adxl355, is 20 bits, 4 more bits than the 345, which has been around for a long time. You can get a 345 breakout board for 2 bucks, but the 355 is $60. I wrote the drivers for the 345, and it will fun to read 20 bits on 3 registers.
So, theoretically, I'll use two of these babies, and it will give me the equivalent array of 32 adxl345's. I'll want to see footsteps on the concrete basement floor, and if I can do that, I'll call it seismometer grade.
For years now I've been trying to build something that is an order of magnitude (10x) cheaper than the existing equipment. I want every building in Alaska, and such, to be instrumented, and to have a noise test. However, I have found out that in order to calibrate, as installed, you need something that can record distant earthquakes. Thus, you need a huge dynamic range. 24 bits is ideal, but maybe 20 bits will do. You don't need anything more sensitive than standard noise.
So, when the Oklahoma earthquakes start up again, you instrument the old concrete office towers. A noise test might show the base frequency at 7 Hz, which is good. An earthquake might soften it to 3 Hz which is bad, and you raze the sucker. Standard inspection just doesn't do it.
No comments:
Post a Comment