angular accelerations through accelerometers

[pierluk]

I have to measure angular acceleration in a vehicle and since I haven't gyrobox, I have to use two accelerometer placed at known distance. At work I have only accelerometer with +/- 50g full scale. Are they good for the measurements I have to carry out? I'm doubtful, they have a low sensitivy for my need.

 

[IRstuff]

Seems to me that you have a more serious issue in that you're trying to measure angular acceleration with a linear accelerometer.

TTFN

FAQ731-376

 

[GregLocock]

It can be done, but whether those accelerometers are appropriate is another matter. The maths is easy, why not work out whether they'll be ok?


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pierluk]

Yes it isn't the best way to measure the angular acceleration, but it is possible for small angles. In my application I think the maximum angle is 6 degrees. However I think my accelerometers have low sensitity and so it should be difficult to measure very small acceleration difference, isn't it?

 

[Tmoose]

What are your accelerometer specs?
What are the accelerations you are interested in?

Is it something like a general purpose accelerometer used for vibration analysis? Noise floor of a a few dozen micro-gs depending on frequency, but unsuitable for "low frequency" events.

https://www.ctconline.com/fileup/3DNEWPDF/AC102series_datasheet_3D.pdf

Then there are these kind of interesting comsumer product accelerometers.

http://www.sparkfun.com/commerce/tutorial_info.php?tutorials_id=167

A word like "Low" applied to sensitivity in some ways reminds me of my wife's response to my inquiry of how much some some purchased item cost. Her reply is almost certain to be "oh, not very much."

I used to think I simply didn't know any more about the item's price after hearing that answer than I did before. But now I know what it really means is that in a few years the equity line on the house is going to be all tapped out.

 

[pierluk]

I'm using entran A2 and entran D2 accelerometer with +/-50g full scale. In this link you can find the spec

http://www.meassensors.com/manage/sensortypes/EGCS.pdf

has you can see my accelerometers has a sensitivity of 40mV/g.

 

[Strong]

Pierluk,

What is the shaft speed?
What is the frequency range?
6-degrees is angular displacement, but what is the angular acceleration you expect (max and min) to measure?
Why do you want to measure angular acceleration instead of velocity or displacement?
Not only do the accelerometers have to produce adequate voltage, but the summing amplifier and telemetry system have to handle the signals with minimum noise. I would consider alternative methods such as an angular accelerometer or encoder-F/V system.

Walt

 

[pierluk]

Strong,

I have to measure vehicle roll and pitch accelerations. I don't know what should be the maximum roll and pitching angles (but I suppose less than 6 degrees for both of them) and their acceleration values. I need to measure the angular acceleration for finding a correlation with the trasmitted driver acceleration. In the company I work, we haven't an angular accelerometer. What do you mean for encoder f/v system?

thanks

 

[IRstuff]

The next relevant question is what accuracy are you looking for?

While it's certainly not impossible to use linear accelerometers, it's nontrivial, since you do not know the center of rotation.

TTFN

FAQ731-376

 

[GregLocock]

It would have been helpful if you'd told us the application rather than just playing guessing games.

Yes 50g accelerometers are completely inappropriate for that application. 2-3g is more like it. You will also need good low frequency response, 0.3 Hz cutoff sounds about right.

The good news is that three or four accelerometers will measure pitch and roll succesfully.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pierluk]

Dear Greg,

for my measurements, it is better to use the accelerometer in DC coupling or in AC coupling?

I haven't understood, what do you mean for " 0.3 Hz cutoff sounds about right". Which cutoff frequency do you means? Do you mean the accelerometer natural frequency?

 

[Strong]

Pierluk,

Disregard most of my comments, since I made the incorrect assumption that you were measuring angular acceleration (torsional vibration) of a rotating shaft.
"for my measurements, it is better to use the accelerometer in DC coupling or in AC coupling?" You have to understand what the measurements are for; need static acceleration (0 Hz) or not.
Forget about the two low sensitivity accelerometers you have and get the appropriate sensor. Here are two expamples:

http://www.endevco.com/product/Product.aspx?product_id=83

http://www.kistler.com/br_en-ex/13_...e-sensor-triaxial-measuring-range-100-s.html>

My advise is to spend some time understanding what is needed (requirements and end use) for the measurements before making the measurements.

Walt

 

[pierluk]

Greg, I know that with DC coupling I can measure the acceleration at 0Hz, but the question is: with AC coupling will I lose some spectral emplitude error close to 0?

Thanks for the suggested sensors, but I believe we haven't these on my company, I will try to ask a collegue.
However probably I can find some accelerometers with +/-5g of full scale, do you think with those ones can I measure the vehicle pithing and rolling accelerations with a good accuracy? Moreover I don't need to know the centre of rotation, but only the distance between the two accelerometer, isn't it?

 

[pierluk]

I have got an idea: I suppose a car equipped with ESP system, should have a sensor for measuring roll angle o roll rate. So can I use that sensor? Maybe I can find a wasted car, to get that sensor almost for free, or that sensor has a too low sensitiviy? In my experience sometimes vehicle embedded sensors aren't very accurate.

 

[pierluk]

sorry Strong, I have committed a mistake refering the reply to Greg, but it is for yor

 

[GregLocock]

If someone asks for 'good accuracy' what the hell does that mean? Engineers use numbers. If you come up with a sensible spec then someone might be prepared to put the time in to understanding your requirements.






Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pierluk]

Yes Greg you are on right.

ISO road vehicles standard reccomeds these instrumentation maximum errors and full scale:
max error full scale
lateral acceleration +/- 0,15m/s^2 +/- 15ms^2
roll rate +/- 0,2°/s +/-20°/s

for pitch rate I have found nothing.
So do you think whit +/-5g full scale accelerometer it is possible to perform good measures?

I have found the following sensor, it is cheap, do you think it is good enough?

http://www.analog.com/static/imported-files/data_sheets/ADIS16354.pdf

 

[pierluk]

I have other ISO standard with further recomended instruments properties. It is written it is possible to measure the angle and calculate the angle rate or acceleration through a derivate. The other ISO standard recomends:
full scale maximun error
for angles: +/-15° +/-0.15°
for angles rate +/- 50°/s +/- 1°/s

so If I have two accelerometers with a maximum error of 0.25g and the distance between the accelerometers is 0.85m, so the maximum error of angle acceleration should be 0,21°/s^2. So how can I asses if the maximum error of my instrumentation is good enough? I hope I have explained better my need.
thanks

 

[IRstuff]

0.25g*sqrt(2)/.85m = 233deg/s^2

TTFN

FAQ731-376

 

[pierluk]

sorry IRstuff, I don't understand your calculation, may you explain better?