In Search of an Accelerometer(s) with +/- 1 - 1.5 g andMUSThaveNoise inHigh Freq

Hi

Can anyone help me find or locate or show me a link of an accelerometer IC with the following specifications:

  1. (+/-) 1 - 1.5 g
  2. High Sensitivity - At Least 100 mV / g
  3. Most Importantly - Noise should be at the High Frequency Range (must be above 200Hz). This is because I am trying to measure/moodel the non-linear Friction of Hard Disk Drive's Actuator Arm and so I am only interested in the Low Frequency Range (10 -100Hz). Therefore I need accelerometer that has noise in the high frequency range as I would ignore high freq part!

Is there anyone kind enough to share some accelerometer idea/knowledge and suggest me an accelerometer. I am unable to find any accelerometer that has noise in high freq range.

Thanks Tasnim

Take a look at this: http://shop.moderndevice.com/products/mma7260qt-3-axis-accelerometer

I bought it for my music box project and used its 6g range but you can use 1.5g range.

Hi,

But what is the noise frequency of this chip? Can't identify in the datasheet. If it has a low noise frequency (10- 150 Hz) then the chip is useless to me. My requirement is either no noise (that would obviously be the best) or noise in the high frequency range (above 150 Hz).

Waiting in anticipation Thanks Tasnim

What do you mean by noise frequency?! If noise has a fixed frequency, why don't we filter it out entirely and be noise-free?

A little more information would be helpful…

What price range? What have you rejected as unsuitable?

Why not this unit for example?
http://www.directindustry.com/prod/pcb-piezotronics/accelerometers-7413-17415.html

A capacitor on the input (a low pass filter) might help if the unit does not have built in bandpass or low pass.

I have written software low-pass filters as well – that’s what I have working now for accelerometers where I cannot get the hardware to behave as I wish.

Is there a problem with the ADXL335 at Sparkfun – other than the 800mv / g (as I recall). It certainly is lower noise.

I have the Bosch BMA180 in trial – writing code for it now. It has band pass low pass and high pass filters built in and is settable. Again check at Sparkfun

hth

Tasnim
I am not sure you understand the concept of noise. Because a sensor has noise in the high frequency range it does not mean it has no noise in the low frequency. By it’s nature noise is spread over a wide frequency range and you can’t just say you want all the sensors imperfections to end up in one place.

There are two types of sensor, analogue output and digital output. The digital ones are easier to use and have had the manufacturer reduce noise as much as possible. They are lower noise than you can get simply with an analogue sensor but not as low as an analogue sensor if you really do a proper design job on the associated electronics.

Can’t identify in the datasheet

There is a reason for this, the manufacturer doesn’t want to tell you. That tells you something about the sensor.

check the BMA180 that got some options for filtering onchip. Otherwise: filter on the micro, so you have complete control on the filter characteristics.

You can program Low Pass filters on that accelerometer, so high frequencies noise are suppressed.

I forgot to mention, I wrote a library for the BMA180 http://www.schwietering.com/jayduino/bma180/

SCJurgen:

I will give it a trial...

If it works out or I spot anything worthy of comment I will start a new thread.

Thank you!!

S. C. Jurgen:

Your library looks excellent! Beyond First Class!

I hope they add it to the distribution! I will comment more later as I work with it! I have already made some minor changes to suit my needs.

WillR: A little more information would be helpful...

[/quote: Is there a problem with the ADXL335 at Sparkfun -- other than the 800mv / g (as I recall). It certainly is lower noise.]

Hi, Do I have to program the ADXL335? I dont want to go into coding now, I need to measure acceleration of the actuator arm of hard disk drive (hard disk drive other parts removed except for the actuator arm)

Grumpy_Mike:

The digital ones are easier to use and have had the manufacturer reduce noise as much as possible.

Hi,

Thanks for your input. But can you tell me any digital accelerometer that has low g, around =/- 1.5g, and high sensitivity like 400 or 500mV/g. Noise should be as low as possible!

Thanks Tasnim

Thanks for your input. But can you tell me any digital accelerometer that has low g, around =/- 1.5g, and high sensitivity like 400 or 500mV/g. Noise should be as low as possible!

As already posted before, try the BMA180: 14bit resolution, programmable to +-1 or +-1.5g, noise is about 150microG/HZ*0.5 (yes, depends on sample rate). It has also a special low-noise operation mode (consuming more power). You can program the data rate, so you ulterior lower the noise by using the internal 2-pole LP-filter (10,20,40,75,150Hz...) which should match exactly your problem if you use the 150Hz or even 75Hz mode. Bit resolution is about 0.18 milliG if you use +-1.5G, which is really sensible. It does not make much sense to pretend a digital accelerator, requiring 400mV/G resolution if you don't consider the resolution of the builtin ADC.

Absolutely consider using the smp_skip sample mode as it reduces noise (and MCU load).

I hope this helps

Hi Everyone,

Thanks for your input. Though the BMA chip is good, it requires programming and I have shortage of time. My application is too simple to go into that direction. By the way, do you guys think the ADXL103/203 is suitable for my purpose? http://www.sparkfun.com/datasheets/Accelerometers/ADXL203.pdf

With High sensitivity of 1000mV/g, even if I have noise, it is not a problem as long as I am able to identify the actual signal from the noise. Here maximum BW of 2.5 KHz will result in a rmsNoise of 7mg (is this a considered a high noise level?)

Any comments or thoughts, am I thinking the right way?

Regards,
Tasnim

The BM180 with Jurgens Library is a far better choice. You should be up and running in 20 minutes on a Mega2560 -- that's what it took me - I soldered on a connector so I could plug it into my breadboard or attach a cable-- wired it to the arduino and ran his example... It did not work that well till I added the required Interrupt line to digital pin 2, but that was my goof. :-) Then you can set the filter to 150HZ -- the function is included.

Page 10 will give you the bits of resolution.

However, if you are going to get the ADXL203 breakout board from Sparkfun you should have linked to that. Linking only to the PDF is misleading.

See the schematic and compare the caps on the schematic to the table on page 10 of the PDF. You will have to change the caps from .01uf to .05 uf -- do you have the equipment to solder the surface mount caps?

hth

Anyway -- it's your project.

It did not work that well till I added the required Interrupt line to digital pin 2, but that was my goof.

Interrupts are a cool way to get things done on a uC but you are pretty much right, i should add a non interrupt driven polling example, I simply forgot that! ;-)

One question though, did you attach the BMA180 at 5V levels???

scjurgen:

It did not work that well till I added the required Interrupt line to digital pin 2, but that was my goof.

Interrupts are a cool way to get things done on a uC but you are pretty much right, i should add a non interrupt driven polling example, I simply forgot that! ;-)

One question though, did you attach the BMA180 at 5V levels???

The routine is fine --- I just was not paying attention. My fault! Not yours!

I have made a few modifications to your routine. One was a simple addition to the notes in the headers.

As for the 5V I am hurt that you think I can't read -- -oh wait I just proved that I can't... :cold_sweat:

I used 3.3 v from my Mega2560. I do note that it specifies 2.4V It occurs to me that I should hook it to 5V with a voltage divider (as in divide by 2 and add a filter cap to the resistor where the 2.5V is present...)

Now the values are reading a bit high -- not that I really care that much as I display "A/C" waveforms in my data collection -- centered about Zero by subtracting the average.... BUT I would also like to calibrate for determining actual force and pressure on the object attached to... One thing at a time though....

The library does need a MODE to sent the Noise/Bandwidth command. (Location 30H, bits 0:1) I dummied up a mode command in my program -- if you do not add a mode command I will, and then send it to you. You saved me quite a bit of time -- it's the least I could do! And your library structure makes it simple.

Now, considering your obvious expertise, we just need you to look at the BM085 library that is floating around -- it gives weird results on some settings...

You have me thinking that you work for BOSCH since you seem to understand their sensors so clearly!

You have me thinking that you work for BOSCH since you seem to understand their sensors so clearly!

lol, hell ]:) no, i am not working for BOSCH! I am into embedded systems and DPS's for music (hobby) and medical systems (work) ;-)

The bma180 came handy for expression control for midi devices and is pretty versatile.

Now, considering your obvious expertise, we just need you to look at the BM085 library that is floating around -- it gives weird results on some settings...

for the weird results, usually happens when you don't respect the specs ;-) !!! A wild guess (i did not have a look into it), that someone reads a status bit, and discards the value and reads it again to have the complete value. But then the values are shadowed, and then, yes, you get weird readings! That is also the reason to use an interrupt, because when the sensor triggers the interrupt it means the values are actually ready, hence no false reading or shadowing.

WillR: However, if you are going to get the ADXL203 breakout board from Sparkfun you should have linked to that. Linking only to the PDF is misleading.

Hi, what do you mean by link, here? The circuitry needs to be as light as possible as I am placing on the actuator arm.

Tasnim:

WillR: However, if you are going to get the ADXL203 breakout board from Sparkfun you should have linked to that. Linking only to the PDF is misleading.

Hi, what do you mean by link, here? The circuitry needs to be as light as possible as I am placing on the actuator arm.

That is indeed a different issue. I was not certain of that before.

I presume that the head will still "fly" with the added weight of the sensor and the wires? ???