Annalise accuracy gyro from the datasheet Hello i am kind of new here. The thing is i want to understand how gyroscopes work. Especially which factors can make them malfunction. From my understanding the values can drift to a point where the gyro is not usable and as they are expensive so i want to understand them better before i buy one. I took this datasheet: www.analog.com/static/imported-files/data_sheets/ADXRS453.pdf and i did read a lot of ap. notes etc. One of them is http://www.alexandertrusov.com/uploads/pdf/avar-trusov-gyroscope-2011-03-31.pdf The questions are: (1a)If i take a look at figure 12 of the datasheet of AD. And try to understand it, the graph means something like: when you average the signal the accuracy first gets better. So when you average the signal for 0.1 hour you get a accuracy of about 0.00135 Deg/sec. Is this true? (1b)I understand that if you average the signal for a longer time the output signal gets better. But why does it get worse after aprox. 0.1 hour? Is this what they call red noise in the article from the second link? (1c)Graph 12 uses averaging time wile in fact it is a spi device so what sample rate should be used? Taking 10 samples for 0.1 seconds or 100 samples in 0.1 seconds? In other words what is used to make graph 12? (2)How can i make the graph of figure 12? (2a)Can i use a matlab program like:http://www.gnu.org/software/octave/ and the rate noise density (deg/sec/√Hz) ? I found this link that looks ok for white noise generation etc.: http://www.tty1.net/blog/2009-09-12-filters-with-gnu-octave_en.html (2b)Can i calculate it with a spreadsheet program? (2c)Is there a other way to make this graph from the noise data? I ask this question because most other brands gyros do not ad this graph in the datasheet. If i can somehow learn calculate this graph i would be able to compare the gyros. I am not used to work/calculate noise but i see this as a great opportunity to learn how "it works" even if i maybe wont be able to buy the expensive gyro. I hope you guys can help me Thank a lot in advance