So, if I got right mpu6050 has 100Hz bandwidth and it means that it can accurately measure changes in acceleration up to 100 times per second. But if I want to use in-build mpu low pass filter, I should specify a certain corner frequency. And here I got stuck cos I don't understand relation between frequency and acceleration.
The bandwidth is usually specified as a tolerance band, relative to the reference frequency sensitivity (usually 100 Hz). The tolerance band can be specified in percentage and/or dBs with typical bands being ±5%, ±10%, ±1 dB, and even ±3 dB. This is quote from web-resource that I also don't understand how from Hz they got dB?
I understand that vibration and movement have different frequency, but what about their acceleration ? If acceleration of vibration will very high, will I cautch it or not ?
Could u advice literature maybe?
Hi, if you mention something, can you give a link to it ? You could be looking at text that is generated by ChatGPT.
Theory
Are you familiar with audio curves ? For example for a headphone.
Suppose that it is 16Hz to 22kHz and those limits are when the sound is at -3dB level.
When a tolerance band of 3dB is drawn in the picture as horizontal lines, then it drops below the bottom line at 16Hz and at 22kHz.
If the tolerance band would be -10dB, then those headphones could be from 2Hz to 30kHz.
The frequency does not go from Hz to dB. The tolerance band is just how accurate the output should be, to measure the lowest and highest frequency which are still acceptable.
In my opinion, it is all a lot of words, a nice graph gives more information.
Practical
The MPU-6050 is noisy, outdated, and no longer manufactured. You should not use it. If you buy it on Ebay/AliExpress/Amazon, then it is a counterfeit.
A Arduino Uno with a MPU-6050 and using the 'dmp' inside the MPU-6050 can be set to a sample rate of 100Hz. That works reasonably well, but it is not enough for a small or fast moving drone. The MPU-6050 is not good to measure an impact and also not very good to measure vibrations, although it will work to measure the vibrations of the 3D printing head.
A "filter", such as the AHRS or Kalman is not just a low pass filter. It is more "sensor fusion", it combines the accelerometer and gyro (and magnetometer) to get a stable output.
You see how the practical use of the MPU-6050 is very far away from the theory that you mentioned.
Was my explanation good enough ?
I have some doubts with paragraph 2 on that page:
- The graph is for a professional sensor. Cheaper MEMS sensors can have peaks in the sensitivity.
- The gray curve is bending down near 0Hz. I have never seen that for MEMS accelerometers used with Arduino boards and smartphones. Every normal accelerometer can measure the static g-force in 3D with all axis.
- It says that 100Hz is the "reference frequency sensitivity". Those three words make it complex. In a datasheet the manufacturer just picks a frequency for reference. From that reference, when the curve below a certain dB then that is limit for the frequency of the vibration it can measure.
In short: Let's say the sensor can measure a vibration of 100Hz well. If, for example, the output signal has dropped 3dB with a 500Hz vibration, then the tolerance band is 3dB and the bandwidth is 500Hz.
In general, there is none, they are completely independent concepts.
If you specifically want to measure vibration, there is an associated acceleration and you need to specify frequency.
Explain what you want to measure.
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.