Query Regarding Accelerometer Resolutions: Discrepancy Between "G" and "No G".

[DaviPrata]

Hello!
I would like to ask for your help regarding a question, please.
I'm conducting some experiments that involve using the accelerometer, and I've noticed a discrepancy in resolution between experiments in certain cases. Today, upon closer examination, I think I've figured out why.
When I use the "Acceleration (no g)" sensor, the experiment's resolution is around 200 samples per second. However, with the "Acceleration with g" sensor, the resolution is doubled, at 400 samples per second.
Does anyone know the reason behind this? I was using the "no g" option because I found the data easier to work with, as all three axes have the same level, but the double resolution would be very beneficial to me.
Is this difference normal? Could it be that different sensors are being used? Is it possible to achieve a 400-sample resolution with the "no g" option?
The device I'm using is a Samsung Galaxy S20 FE. In the device information, it mentions the Accelerometer sensor (type 1) LSM6DSO, Resolution: 0.0023928226 m/s², but it also lists several other sensors that I'm not familiar with, labeled as unknown.
Any suggestions or tips would be greatly appreciated.
Thank you all for your attention.

[Jens Noriʇzsɔɥ]

The linear acceleration (without g) is a so-called fusion sensor of typically accelerometer, magnetometer, and gyroscope. Checking https://phyphox.org/sensordb for SM-G781 (that should be your device‘s model family), the magnetometer is at 100Hz. I thought that the fusion sensor could not go beyond that, so you get double of what I would have expected…

[DaviPrata]

The linear acceleration (without g) is a so-called fusion sensor of typically accelerometer, magnetometer, and gyroscope. Checking https://phyphox.org/sensordb for SM-G781 (that should be your device‘s model family), the magnetometer is at 100Hz. I thought that the fusion sensor could not go beyond that, so you get double of what I would have expected…

Hi Jens!
First, thanks for the reply.
In this case, the device I'm using is a SM-G780G/DS.
So, for me to have the pure accelerometer I would have to use the accelerometer with g? Can you tell?
Could it be that if I discount the gravity value on the Z axis (or whatever axis it is influencing), I will have a value similar to "Without G"? But with 400 Hz?

[Jens Noriʇzsɔɥ]

Yes, you get –sort-of– raw data from the accelerometer “with g” at the highest possible rate (there might be still some processing by the sensor or smartphone firmware). If you definitely know the direction of Earth's acceleration g, you could obtain the linear acceleration by subtracting g. So, the smartphone should not rotate in the slightest.

[DaviPrata]

Yes, you get –sort-of– raw data from the accelerometer “with g” at the highest possible rate (there might be still some processing by the sensor or smartphone firmware). If you definitely know the direction of Earth's acceleration g, you could obtain the linear acceleration by subtracting g. So, the smartphone should not rotate in the slightest.

Thanks a lot for your help, Jens.
This information will help me a lot.