What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

[matrixofdynamism]

The datasheet for the ADXL330 states on pg.12:

​

The selected accelerometer bandwidth ultimately determines the measurement resolution (smallest detectable acceleration). Filtering can be used to lower the noise floor to improve the resolution of the accelerometer. Resolution is dependent on the analog filter bandwidth at XOUT, YOUT, and ZOUT.

​

The output of the ADXL330 has a typical bandwidth of greater than 500 Hz. The user must filter the signal at this point to limit aliasing errors. The analog bandwidth must be no more than half the analog-to-digital sampling frequency to minimize aliasing. The analog bandwidth can be further decreased to reduce noise and improve resolution.

​

The ADXL330 noise has the characteristics of white Gaussian noise, which contributes equally at all frequencies and is described in terms of μg/√Hz (the noise is proportional to the square root of the accelerometer bandwidth). The user should limit bandwidth to the lowest frequency needed by the applica-tion to maximize the resolution and dynamic range of the accelerometer.

I am confused between resolution and bandwidth. I think that resolution is the smallest movement acceleration produced by spatial movement that the ADXL330 can detect. On the other hand, since movements can be repetitive like a vibration, the bandwidth is the maximum continuous motion that the ADXL330 can detect and provide output for. Is this correct?

If so, how/why are resolution and bandwidth related? Why does one really need to use capacitors at the output of the ADXL330 which shall form a low pass filter together with resistors inside the device?

 

[KlausST]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

Hi,

Noise is a mixture of random frequencies.
This noise is added to the acceleration signal.
Therefore the true acceleration signal is somehow invisible inbetween the noise.
If you pick a single sample value it has a lot of uncertainty caused by noise.

Now comes the bandwidth thing..
Limiting the bandwith means to suppress high frequencies with a low pass filter.
If you are interested in low frequency acceleration signals, then you filter away all the (useless) high frequencies...usually noise.
Remainig is the acceleration signal with less noise...means less uncertainty...means higher acceleration signal quality.
= better signal to noise ratio.
As an effect you get better resolution of the acceleration signal.

Klaus

 

[matrixofdynamism]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

OK, so with lower noise (since most higher frequencies were filtered out), we are able to actually successfully detect much smaller changes in the measured variable called g (acceleration due to gravity or otherwise). Is that correct?

I find this IC a bit strange and messy. I would expect this type of IC to have a digital interface so we can communicate with it via SPI, I2C or some other interface.

By the way, the Mikroe have a daughter board with this IC on it here. If you look at schematic in Figure 3 of the user manual of this board found here, once notices an opamp connected to this accelerometer.

Initially, I thought that the opamp must be amplifying the signal. But then I found that it is merely being used as a buffer.
(1) Why is a buffer required since the accelerometer datasheet said nothing about needing a buffer?
(2) What purpose do the 47 ohm resistors serve??

 

[Warpspeed]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

Buffer is often required because the accelerometer itself probably has a fairly high impedance output, and is not really capable of driving much of a load.
And high impedances are also much more susceptible to noise pickup.

So if this sensor has to (for example) drive a long screened cable, a buffer is pretty essential.

If it has some kind of amplifier or active filter situated right next to the accelerometer chip it should not need the buffer.

 

[FvM]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

I find this IC a bit strange and messy. I would expect this type of IC to have a digital interface so we can communicate with it via SPI, I2C or some other interface.

If you think so, you have chosen the wrong chip. ADI has both, accelerometers with analog and digital interfaces, simply review https://www.analog.com/en/products/mems/mems-accelerometers.html#mems-accelerometers

Nevertheless there are many possible reasons to use an analog accelerometer.

 

[matrixofdynamism]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

Actually the ADXL330 came along with a development board that I bought sometime ago.

(1) What reasons would one have to use analog accelerometer? Since the actual design shall ultimately use a microcontroller or microprocessor (or a PLD of some sort), this shall just make things inconvinient by requiring that an extra ADC be added into the design isn't it?

(2) According to the datasheet, the output buffers of this chip have 32kOhm resistors in series. They are used to form low pass filter by connecting shunted capacitors at the output pin. Is 32kOhm output considered "high impedance"?

(3) Why are high impedance outputs more susceptible to noise pick up?

(4) What is the purpose of the 47ohm resistors here:

 

[Warpspeed]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

What makes you think every possible application for this sensor always must have a microprocessor ?

Yes 32K is relatively high impedance.

Stray voltages and currents coupling into a high impedance have a much greater effect than if coupled into a very low impedance.

The 47 ohm resistors are probably there to limit the fault current if you accidentally short the output to ground.

 

[FvM]

Re: What does the term "bandwidth" mean for the ADXL330 3-axis accelerometer?

The 47 ohm resistors are probably there to limit the fault current if you accidentally short the output to ground.

Also isolation of possible load capacitance, e.g. cable capacitance. Connecting it directly can cause OP instability.