SparkFun Forums

[sebmadgwick]

I have found that the MPU-6050 accelerometer sensitivity is half that stated in the datasheet for all ranges; i.e. 1g output is ~1024, ~2048, ~4096 and ~8192 for ranges 16, 8, 4 and 2 g respectively. This is the case for two parts I have tested.

To make things more confusing, the accelerometer output still saturates at 32768 which would mean that the accelerometer has a 32g range!

The gyroscope and thermometer outputs are as expected.

Has anyone else had this issue?

[sebmadgwick]

I think I found the answer in the Invensense Developers forum:
http://www.invensense.com/developers/fo ... hilit=half

It seems that our parts are not production parts but are engineering samples that have sensitivities values half that of the production parts and are only intended to operate for FS_SEL=2 (8g) only.

I guess our parts were not purchased from SFE.

[nahid5692001]

Thanks. I was able to get it to work and applied my algorithms for yaw pitch and roll successfully. Now I am good to go.

I was playing around with the IMU i built and for my research project, I need to determine if the IMU is on a wall or flat on a table.

Determining if it is parallel with the ground is simple because the Z axis can tell me that. But determining if it is on the LEFT, RIGHT, FRONT, or BACK wall is getting to be daunting because the Z axis is perpendicular ( roughly 0 ) and the X,Y accelerometer values give numbers but they are roughly the same on ALL the walls.

My question is, is there any way I can tell if I am mounted on a wall or flat on a table with an IMU (gyroscope + accelerometer)?

I appreciate everyone's time. Thank you.

[Mee_n_Mac]

A little OT ?!?!

[sebmadgwick]

Invensense finally got back to me and said: The parts you have are an early engineering sample (hence the ES on the package). The sensitivity was halved, but for the production version this has been corrected.

I read in the Invensense developers forum that this part should only be used with full-scale range selected as ±8 g (AFS_SEL=2). Another significant difference of this part is that the accelerometer may have very large bias errors.

The correct datasheet for this part is:
Document Number: PS-MPU-6000A-00
Revision: 2.3

[SkikoVienna]

I have found that the MPU-6050 accelerometer sensitivity is half that stated in the datasheet for all ranges; i.e. 1g output is ~1024, ~2048, ~4096 and ~8192 for ranges 16, 8, 4 and 2 g respectively. This is the case for two parts I have tested.

To make things more confusing, the accelerometer output still saturates at 32768 which would mean that the accelerometer has a 32g range!

The gyroscope and thermometer outputs are as expected.

Has anyone else had this issue?

I don't know if you still read this post, but I've the same issues here. I get values way over the measurment range with my engineering sample. Did you solved them?

[sebmadgwick]

My above posts conclude the topic. The part is an engineering sample and has a different datasheet.