For my project i need to measure accelerations up to 16g. My options are either digital accelerometers such as the MPU6050 or MPU9250 or an analog accelerometer such as the ADXL326.

What are any major pro's and con's of each. number of wires would be almost equal. the digital accelerometer is less pins on the board but i could easily use analog to digital convertors to bypass that. sampling rate is lower at 550Hz for the analog vs 4000Hz for the MPU9250. However i believe 550Hz would be more than adequate. And the biggest difference for me. £2 for an MPU9250 and £12 for the ADXL326!

Does anyone have any general advice on the main differences when choosing between analog and digital? I am measuring suspension motion on a motorcycle and need 5 of these accelerometers.

  • Is the motorcycle supposed to be driving while you take your measurements? I suspect you would only capture vibrations rather than the (relatively slow) movement of the suspension components relative to each other. Hall sensors with a mechanical linkage are traditionally used to measure suspension movements on cars. IR distance sensor may also be an option.
    – towe
    Jan 7 '19 at 12:37
  • Very interesting point. I successfully recorded suspension motion of a motorcycle moving over a specially designed obstacle and compared to a numerical simulation and the answers were relatively close. I've never heard of hall sensors being used. Would you mind sharing a link to an example of such a system or device? I've only seen linear potentiometers used.
    – Ross Hanna
    Jan 7 '19 at 14:20
  • It's mostly used in finished products, but both my BMW E39 and MB W211 use hall based sensors for regulation the air suspension (and automatic headlight adjustment) functions. Have a look for "W211 Höhenstandgeber" to see an example, I'm not sure about the licences on those pictures.
    – towe
    Jan 7 '19 at 15:25
  • Thanks, i had a look at the sensor, and it seems that it uses a linkage system with a hall sensor on a rotary shaft or similar? The type of system i was looking at was more similar to a suspension potentiometer. However i might start a new thread for people to comment on alternatives, but i will still test the accelerometer method as it is an interesting proposal in my mind if the small accelerations from vibration could be filtered out.
    – Ross Hanna
    Jan 8 '19 at 13:12

The biggest difference is where the measurement (the conversion to digital) is done. On the digital accelerometer, it is done in the accelerometer itself. On the analog one it is done at the MCU.

For simple arrangements where the accelerometer is near the MCU that makes no difference. However the analog one is more susceptible to noise. With longer wires / traces between the accelerometer and the MCU you can get less accurate results with an analog one since it picks up noise which you are then sampling and including in your results.

The digital one is more immune to that kind of interference - instead it is susceptible to digital noise - corruption of the data as it traverses long wires. But that is less of a problem since it takes far more induced voltage to corrupt a digital signal than an analog one.

One other main difference is that you are limited to the ADC that is in the digital accelerometer. If you want a different resolution you have to use a different accelerometer. With analog, since it has no inherent resolution, you sample at the resolution of your MCU's ADC.

Digital is also easier to interface in modern systems. Many SoCs don't have analog inputs. So adding an analog accelerometer would mean adding an ADC chip as well - and you don't really want to do that in something like a mobile phone.

So in summary:

  • Digital is the most convenient, reliable and suitable for current basic uses
  • Analog is more of a niche product that is rarely used in consumer electronics

The ADXL326 is only an accelerometer.
When you have one, you can measure the output of the sensor with a oscilloscope. For the final project a digital sensor is preferred as @Majenko has explained in his answer.

The MPU-9250 is an accelerometer, gyro and magnetometer. That makes it possible to use sensor fusion (combining the sensors with software calculations for a much improved result).

Are there vibrations involved? An accelerometer is sensitive for vibrations. I doubt if an accelerometer can be used to measure the suspension motion.

By the way, I'm ignoring the MPU-6050 and so should you. The MPU-9250 is less noisy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.