FXLS8471Q sampling has a delay

[1730315381]

Dear friends:

  Hi,I use the FXLS8471Q for sampling acceleration sampling, the host is a 5675k microcontroller.
The following figure shows the hardware write initialization procedure and experimental result diagram of the accelerometer.
Compared with the experimental data measured by the ESC on the vehicle, the measured lateral acceleration of the vehicle has a significant delay of more than 100ms.
Very confused, please help me answer, thank you.

[ernestina46]

Yes there's a delay between the input of the system (physical acceleration) and the digital stream representing the measurement, mainly due to A/D conversion and some filtering. This delay should be about 1-2 x sample period (1/ODR).

So if you need to minimize the delay, you shall use highest output data rate (ODR = 800Hz). At that rate, Low Power Mode will give you fastest response time. On the downside, noise will be slightly higher. kissanime

[TomasVaverka]

Hi,

I am afraid I need more information to get a better understanding of the issue you are facing,

1. Does the red curve correspond to the FXLS8471Q and the blue one to the ESC?

2. Have you tried changing the ODR and/or OSR? See table 39 of the datasheet for more information.

3. How do you synchronize the data acquisition from ESC and FXLS8471Q?

Best regards,

Tomas

[1730315381]

Hello Tomas,

Thank you for your reply, I will explain my problem in detail below.

Note: The red curve in all the figures below corresponds to the FXLS8471Q, and the blue curve corresponds to the ESC on the car.

The figure below is the circuit schematic of the FXLS8471Q.

1. Yes, the red curve in the figure corresponds to FXLS8471Q, and the blue curve corresponds to the ESC on the car;
2. I refer to Table 29 in the datasheet and tried to change the ODR and OSR.
(1) When changing ODR=100Hz and 50Hz, the experimental results are as follows:

(2) When changing OSR=4 and 2, the experimental results are as follows:

As can be seen from the above figure, the effect of changing the ODR and OSR is not obvious.

3. My goal is to change the hardware settings of the FXLS8471Q to:
The curve using the FXLS8471Q is slightly ahead of the curve measured by the ESC on the car, because the ESC on the car still takes some time to process the data.

In addition, I used the FXAS21002C to collect the yaw rate signal on the car, and the effect is better, as shown below.

I am very confused, why use FXAS21002C can be used normally, and FXLS8471Q will not work, there will always be a delay, this delay makes my subsequent work can not progress.

Best regards,

Impression.

[JackITB]

Hi Gao,

Yes there's a delay between the input of the system (physical acceleration) and the digital stream representing the measurement, mainly due to A/D conversion and some filtering. This delay should be about 1-2 x sample period (1/ODR).

So if you need to minimize the delay, you shall use highest output data rate (ODR = 800Hz). At that rate, Low Power Mode will give you fastest response time. On the downside, noise will be slightly higher.

I have a couple of questions to better understand your plots:

- what's the exact time scale of the X axis ?

- what is the ODR setting for FXAS21002C ?

Regards,   Jacques.

[1730315381]

Hello   Jacques,

FXAS21002C is set to ODR=200Hz and it can be used normally.

I can understand that the accelerometer you mentioned has a delay of 1-2 sample periods. As you said, I use ODR=200Hz, then there will be a 5-10ms delay in theory, but from the current experimental results, the accelerometer produces a delay greater than 100ms.

Best regards,

Impression.

[JackITB]

Hi again,
Thanks for the quick feedback.
The delay you are reporting for the accelerometer is really excessive so I would suggest you carefully check the way sensor data are collected and displayed along with the ESC data, to identify possible extra time-lag.
What's surprising me also is that the delay looks almost the same between 100Hz and 50Hz ODR, whereas it should double, so there maybe other contributor besides sensor.

[1730315381]

Hello Jacques,

Thanks for your quick reply.

The ESC data should be correct and the model number FXAS21002C (gyroscope) can be used normally in the car.
So, the problem should still be on the FXLS8471, and I am also trying to  change its configuration continually and try to eliminate the delay.

Best regards,

Impression.

[1730315381]

The host and peripheral data transmission mode is polling mode, and the interrupt technology provided by FXLS8471 is not used.