Can Gyroscopes/Accelerometers still be useful under constant vibration?

By all means give it a try but you are right to be worried about the vibration.

If you try to measure small low frequency signals from a sensor which is also subject to high levels of higher frequency stimulation (any sensor not just accelerometers and gyros) then you must at least make sure that the sensor stays within it's linear range.

Sometimes HF noise (random or periodic) is actually beneficial so long as you can filter it out again - example are dithering (used to help mitigate quantising effects) and HF bias used with magnetic tape recorders.

I suspect that accelerometers will be more tolerant of high levels of out of band vibration than gyros (because the gyros rely on vibrating structures inside to work) but I haven't tested this.

You would be able to use an accelerometer with perhaps a +/- 2g range if there were no vibraion but one with a bigger range might give better results overall in your application.

MK

I agree with Mike,

You can always filter out the vibration.  If you use another sensor that picks up the vibration, you can use an Op Amp to feed the inverted vibration onto your accelerometer inputs to get a cleaner signal.

You usually need to integrate the accelerometer data to get changes in velocity and position, so you just need to isolate the error induced by the vibration.

This is a common problem on airframes as they fly through buffeting winds.

DAB

Thanks! I'm far more confident now knowing the physics of the process aren't hopeless. I'll probably end up testing both an accelerometer and gyroscope just to see what I can see.

Hello Streuzucker,

First, let's get your requirements straight; you ask for positional information, AND about whether the needle is horizontal or vertical. The first is position, the latter orientation (angle of object with respect to fixed world).

Getting position requires you to set a 'zero' point, after which you have to integrate an accelerometer twice (first integration gives velocity, second integration gives position). However, this will drift tremendously, especially if your vibration causes higher frequencies than what your sample rate can cover.

Getting orientation is mostly done by integrating the gyroscope once and/or using the accelerometer to find the angle with respect to the gravity vector. Finding gravity vector works fine for slow movements, and you could filter out the vibrations if you're only interested in low speed movements.

Finally, I'd like to suggest two options for you if you're interested in orientation:

• look into RC (Radio Control) groups, as a lot of people there use orientation sensing by MEMS sensors, using various filters with differing complexity
• Have a look at the MPU9150, which is a Gyro / Accelero / Magneto combisensor. In Germany it's distributed by Scantec. There's arduino code available to get orientation output

Thank you for the reply. My requirements are vague, that's true. Maybe I can explain myself a bit better.

The concept behind the project is to measure everything I can about the tattoo process, then use those metrics to create some secondary visual art (primary being the tattoo itself). So I wouldn't exactly use the accelerometer/gyroscope data for the pinpoint telemetry. I'm mostly interested in getting a few stable data sources that I can map to some aspect of the visual output. For example, I could use processed values from a 3-axis gyroscope to inform the RGB values of the shape I'm currently coloring.

So in that respect, I have a lot of slack when it comes to things like calibration and drift since my final output is going to be more or less arbitrary anyway.

And thanks for the part recommendation! I was looking at the MPU9150 last night, it seemed like it fit my needs perfectly. And now that you've seconded the nomination, I think I'll pull the trigger and get one!