Does the Mass of an Accelerometer Affect Vibration Frequency?

[Thien]

Hi everyone,

Is it normal if the acceleration measured is different with different instrumentation?
I have measured the acceleration of shaker with three different accelerometer and I didn't obtain the same result.

Thanks for your help!

 

[Simon Bridge]

It is normal for different instruments to disagree - the same instrument, if it is sensitive, will disagree with itself when repeating the same measurement - if they are properly calibrated, though, they should agree to within their uncertainty limits.
If all measurements agree, it is usually suspicious.

 

[Thien]

It is normal for different instruments to disagree - the same instrument, if it is sensitive, will disagree with itself when repeating the same measurement - if they are properly calibrated, though, they should agree to within their uncertainty limits.
If all measurements agree, it is usually suspicious.

Thank you for your response!

So, how can we know which intrument is reliable to calibrate the others?

Sincerely

 

[Simon Bridge]

You don't use an accelerometer to calibrate other accelerometers. You should use the manufacturers specifications for calibration.
In general, you calibrate instruments using known values that you obtain by other means ... i.e. use known accelerations (measured using a stopwatch and an accurate ruler say). It helps to decide how accurate you need the measurements to be.

If you knew that one accelerometer was the most accurate - wouldn't you'd just use that one and not bother with the others?
I'm guessing you are using the accelerometers in a device like a phone - and you have different kinds of phone... is that correct?

Note: the most reliable accelerometer is the one that returns the same value most consistently each time for the same test... reliability is not always helpful: an accelerometer that has two readouts "accelerating" and "not accelerating" will probably be very reliable... ("reliable" can also mean "robust").

 

[Thien]

You don't use an accelerometer to calibrate other accelerometers. You should use the manufacturers specifications for calibration.
In general, you calibrate instruments using known values that you obtain by other means ... i.e. use known accelerations (measured using a stopwatch and an accurate ruler say). It helps to decide how accurate you need the measurements to be.

If you knew that one accelerometer was the most accurate - wouldn't you'd just use that one and not bother with the others?
I'm guessing you are using the accelerometers in a device like a phone - and you have different kinds of phone... is that correct?

Note: the most reliable accelerometer is the one that returns the same value most consistently each time for the same test... reliability is not always helpful: an accelerometer that has two readouts "accelerating" and "not accelerating" will probably be very reliable... ("reliable" can also mean "robust").

I use an MEMS accelerometer of Analog Device, a laser telemeter and an PCB piezo accelerometer to measure at harmonic regime.

Moreover, the acceleration level is so small about hundreds "mg". It's difficult to calibrate accurately. And I can't find the manufacturers specifications for calibration.

 

[A.T.]

So, how can we know which intrument is reliable to calibrate the others?

I would use a turn table to spin them at some precisely known radius and angular velocity, then compare their reading with the calculated centripetal acceleration. But note that you have to know the exact position of the actual sensor for this.

 

[Thien]

I would use a turn table to spin them at some precisely known radius and angular velocity, then compare their reading with the calculated centripetal acceleration. But note that you have to know the exact position of the actual sensor for this.

Thank you for your response!

Do you mean that the value measured is depend on the position of each accelerometer?
I'm not clear about radius and angular velocity which you mentioned, can they be measured?

 

[A.T.]

I'm not clear about radius and angular velocity which you mentioned, can they be measured?

I think that measuring the radius and angular velocity in a rotating setup is more reliable than accelerating the sensors along a linear track, to calibrate them.

 

[nasu]

How did you measure acceleration with a laser telemeter? Recorded position versus time and integrated to get acceleration?

 

[A.T.]

How did you measure acceleration with a laser telemeter? Recorded position versus time and integrated to get acceleration?

You would have to differentiate it twice, which for noisy data usually means fitting some smooth function.

 

[nasu]

Sure, you are right.
We had some discussion with measuring acceleration and wanting to get position by integration and remained in the "integration mode". :)

 

[Simon Bridge]

There is still the question of how accurate the measurements need to be ... for that matter, how different are they?
Are the measurements consistent within the manufacturers uncertainty limits?
Perhaps if we had an example of the data with a description of the measurement process?

However - the controlled situations suggested above should help calibrate the equipment.

 

[Thien]

How did you measure acceleration with a laser telemeter? Recorded position versus time and integrated to get acceleration?

Hi nasu!
I carried out two measurements with the same laser telemeter: 1/ displacement measurement then integrated to get acceleration. 2/ acceleration measurement directly. There is a difference between two measurements and this ecart type relative is 27%.

 

[Thien]

I think that measuring the radius and angular velocity in a rotating setup is more reliable than accelerating the sensors along a linear track, to calibrate them.

I agree with you but I don't have equipment to do that..
Note: the acceleration in Z-axis is useful for my test.

 

[Thien]

There is still the question of how accurate the measurements need to be ... for that matter, how different are they?
Are the measurements consistent within the manufacturers uncertainty limits?
Perhaps if we had an example of the data with a description of the measurement process?

However - the controlled situations suggested above should help calibrate the equipment.

Hi Simon!
The shaker is excited with a function generator and then the vibrating part is measured with three accelerometer. The ecart type depends on frequency and tension excitation. For example, the Analog device and the laser telemeter has 19% and 200% ecart type at 30Hz and 100Hz, respectively.

 

[Simon Bridge]

Hi Simon!
The shaker is excited with a function generator and then the vibrating part is measured with three accelerometer. The ecart type depends on frequency and tension excitation. For example, the Analog device and the laser telemeter has 19% and 200% ecart type at 30Hz and 100Hz, respectively.

You just left out a great deal of information - assume I have not looked at your setup and do not know what you are doing or what equipment you have.
It sounds like you have a machine for shaking something - what? how? What sort of shaking do you expect?
Are you trying to measure the acceleration in three different axes directions at the same time or just at random orientations about the thing being shaken or what?
19% and 200% of what? 30Hz and 100Hz what?
You still have not said how accurate you need them to be...

I think we've gone as far as we can without knowing what you are trying to do.

 

[CWatters]

Is the mass of the accelerometer changing the frequency of the vibration?

 

[Thien]

You just left out a great deal of information - assume I have not looked at your setup and do not know what you are doing or what equipment you have.
It sounds like you have a machine for shaking something - what? how? What sort of shaking do you expect?
Are you trying to measure the acceleration in three different axes directions at the same time or just at random orientations about the thing being shaken or what?
19% and 200% of what? 30Hz and 100Hz what?
You still have not said how accurate you need them to be...

I think we've gone as far as we can without knowing what you are trying to do.

The acceleration is recorded in z-axis at the same time. To be more clearly, I want to measure precisely acceleration in z-axis at low frequency with a shaker. The accuracy could be less 10%.

 

[Thien]

Is the mass of the accelerometer changing the frequency of the vibration?

No, it doesn't change the frequency but it changes altitude of the acceleration.
But the measurement condition is unchanged.