SmartMotion® MEMS Motion Sensors

I have called

dmp_get_pedometer_step_count(pedo);
dmp_get_pedometer_walk_time(time);

However both pedo and time always keep in 0. The value never updated no matter I shake the sensor how many times.

The defined address is following:
#define D_PEDSTD_STEPCTR (768 + 0x60)
#define D_PEDSTD_TIMECTR (964)

I have checked that I can write these register and I also can read back the value I written.

Are the addresses wrong or I need to enable anything?
I can read the gyro data, so the dmp should be enabled.

So I've been stumped trying to connect my windows 7 setup to the mpu 9250 through the ev_invarm_g eval board. I installed the driver included in Motion6.12 but have not been able to connect with the device with IDL. The data sheet's suggestions: 1) reinstall the driver (I did). 2) "Flash" the memory and reupload the firmware. Trying this requires additional hardware and software, but I had access to a SAM-ICE debugger and downloaded the Atmel Studio software.

We're seeing magnetometer reading spikes at random intervals in our MPU9250 based design

The spikes appear to be a corruption of the sign bit as the value always spikes +ve to -ve or visa versa.

The spikes appear if we read the mag axis in bypass mode and also in normal mode. In bypass mode we check for data ready and over flow.

I2C bus speed makes no difference.

Anyone else come across this ? We're scratching out heads.

Rob

Hi,

Hi,

I am trying to use ICM-20608 in my project, i would like to know what technical changes i have make to implement ICM-20608 into my project.
I am currently using MPU-6050 would it be possible to upgrade it to ICM-20608. please provide me with information to take note of while implementing it in my c programming source code be it changes in register map etc.Thank You.

I am using the DMP engine in the MPU9250 part to output quarternions using the eMD library. I use a dmo fifo rate of 20 Hz (default).
I notice that after power on and in stationary state the value of the quaternions being outputted is not (1,0,0,0) - (indicating zero rotation).

The initial values instead show an non trivial existing rotation eg: (0.81,0.62,0,0)

Can you explain why this is the case?

Also when quaternion data is read from the fifo, what is the order of the coordinates? I am seeing the order as: x,w,z,y

Is this the expected order?

Hi,

I'm using the MPU9250 with msp430.
Sometimes (during debugging and power on/off) the MPU9250 hold the SCL line down and don't release it.
Studying the datasheet I found that nCS pin for I2C operation is connected to VDD (Figure 2 [a]).

I have two questions:
1. Regarding the nCS pin: should it be connected to VDD or I can keep it floating for i2c operation?
2. Is there any legal way to "reset" the MPU9250 (or to recover the I2C module) without re-powering it?

Regards,
John.

Greetings,

It's my understanding that ultrasonic welding is a high-risk assembly technique due to the welding frequencies being near the mechanical resonant frequencies of the sensors (most notably the gyro). I would like to understand risk-mitigation options in the event that ultrasonic welding is the only choice.

I’m running the Embedded MotionDriver 6.12 software on a STM32F DISCO board. Everything seems working fine, as shown by the Python display. My question is about understanding of the MPL. More specifically, from user’s (applicate code) point view, where the MPL functions are called from? I can see these functions / features are enabled in main() before the while(1) loop, but I am expecting they are called regularly by the user application code. It seems the MPL functions are running invisible to the user.

Thanks in advance!

Xipu

We are looking to do some HWIL testing and are looking for a good injection point for accel / gyro data. Ideally, it looks like the analog signal coming from the sensors would be a good spot for injection gyro/accel data. However, not sure these are available at the pins. Am half hoping the NC might not be NC at the pinouts. If not where is the easiest place to inject gyro / accel data and turn off the magnetometers...

Thanks in advance.