Library compatibility with SAMD architecture

[TinManAkshay]

Hi,

We are going to have a custom board (SAMD51 MCU) which LSM6 module on it and I plan to use this Arduino based library having the development platform to be Arduino IDE. I am thinking to use the same example on the repository but little modified to extract the Accelerometer values. I am getting this warning """"WARNING: library LSM6DSR-master claims to run on stm32, avr, sam architecture(s) and may be incompatible with your current board which runs on samd architecture(s).""""

Is it true that it is not compatible with SAMD arch. In that case, do i have modified something in the libraries? Can anyone please suggest since its critical and deadline is getting closer for our project.

Thanks, Akshay

[fpistm]

Hi @TinManAkshay As far as I can tell, I think it could be compatible as the library seems only used default Arduino API.

@cparata any feedback on this ?

[cparata]

Hi @TinManAkshay , it should be compatible, because the library uses only standard Arduino APIs. Unfortunately, I do not have a device based on SAMD platform to test the library with it. So, I suggest to test the library with your platform and if it is fine, we can update the library.properties file in order to support also samd platforms and remove your warning. Let me know, how it works on your side. Best Regards, Carlo

[TinManAkshay]

Hi @cparata & @fpistm , Alright. Once we have the boards in the house, I will test the libraries and will let you know and if its working smoothly.

Also, we are not using any of the Interrupt pins (INT1/2) of the LSM6 module. I hope libraries will still work on getting the Accelerometer values right?

Thanks -Akshay

[cparata]

Hi @TinManAkshay , yes, the driver should work also if you do not use the interrupt pins. The only thing, if you are connecting the component to the I2C bus, I suggest to mount a Pull Down resistor on INT1 pin in order to disable the I3C support in the component that could interfere with the I2C protocol. If you are connecting the sensor to the SPI bus instead you do not need this patch. Best Regards, Carlo

[TinManAkshay]

@cparata ,

Yes we have I2C connection with the module. But I have a concern that we don't have interrupt connection to the MCU, so how would you suggest in that case?

-Akshay

[cparata]

Hi @TinManAkshay , As I said before, it is not an issue if you do not connect the interrupt pins to the MCU. I would just add only a Pull-Down resistor on INT1 pin of the sensor in order to disable the I3C support on the sensor as you can see in the Datasheet of the component. Best Regards, Carlo

[TinManAkshay]

Alright!! Thanks a lot for the info. Appreciate it.

[TinManAkshay]

Hi,

I have got this https://www.digikey.com/catalog/en/partgroup/lsm6dsr-adapter-board/89058 LSM6 dev board to test your library. So far I have been testing this sensor with SAMD51 micro-controller and I have kept INT_1 of the sensor open since our custom board also doesnt have anything connected to this pin of the sensor, not even pull-down resistor.

This is the result I am getting (attached)

[TinManAkshay]

Hi @cparata @fpistm ,

I am only getting values at accelerometer[0] which are varying rapidly but [1] and [2] are '0'. I dont know why is this showing behavior like. I should receive at those Y and Z axis. Any suggestion could be great help?

Thanks, Akshay

[cparata]

Hi @TinManAkshay , could you share the sketch that you are using? Best Regards, Carlo

[cparata]

Anyway, the output that you showed is a bit strange if the board is still. If the component is still on the table (without any inclination), I expect 2 components with a value close to 0 (the range between -50 to +50 should be fine) and the other component with a value close to +/-1000 (the range between +/-950 to +/-1050 should be fine).

[cparata]

I also suggest to connect INT1 of the component to one pin of SAMD51 (let's say D4 of the Arduino header) and add at the beginning of your sketch in the setup function:

pinMode(D4, OUTPUT); digitalWrite(D4, 0);

In this way, we are sure that the I3C interface of the component is disabled. Could you make also this test?

[TinManAkshay]

Hi @cparata ,

I have tried connecting INT_1 to A5 as mentioned in the example as well as to D5 (since D4 is not there). But still behavior is still same. I am getting first value and '0' other values. I am moving the accelerometer breakout board, but yeah same behavior.

I am using this sketch:

#include <Wire.h>
#include <LSM6DSRSensor.h>
#include <lsm6dsr_reg.h>

#define INT_1 A5

// Components
LSM6DSRSensor *AccGyr;
uint16_t windowLen = 25;

void setup() {

//  pinMode(5, OUTPUT);
//  digitalWrite(5, 0);
  pinMode(LED_BUILTIN, OUTPUT);
  // Force INT1 of LSM6DSR low in order to enable I2C
//  pinMode(INT_1, OUTPUT);

//  digitalWrite(INT_1, LOW);

  delay(200);

  // Initialize serial for output.
  Serial.begin(115200);

  // Initialize I2C bus.
  Wire.begin();

  AccGyr = new LSM6DSRSensor (&Wire, LSM6DSR_I2C_ADD_L);
  AccGyr->Enable_X();

}

void loop() {

  // Read accelerometer and gyroscope.
  int32_t accelerometer[3];
  uint32_t tstart;

  AccGyr->Get_X_Axes(accelerometer);

  digitalWrite(LED_BUILTIN, HIGH);
  // Output data.
  //Serial.println("LSM6DSR: | Acc[mg]: ");
  Serial.println((float)(accelerometer[0]/1000));
  Serial.println((float)(accelerometer[1]/1000));
  Serial.println((float)(accelerometer[2]/1000));

  Serial.println("****************");

  delay(2);
 /* Serial.print(pitch2);
  Serial.print(",");
  Serial.print(roll2);
*/
  digitalWrite(LED_BUILTIN, LOW);

}

[TinManAkshay]

Sometimes i get 2nd value as a steady value, as in, no change in the value at all after moving the accelerometer. See down below the result:

[cparata]

Hi @TinManAkshay , how did you connect SA0 pin? To GND or to Vdd? I guess to GND if you are using LSM6DSR_I2C_ADD_L. Can you try to print the integer values of the Axis without the float casting?

Serial.println(accelerometer[0]); Serial.println(accelerometer[1]); Serial.println(accelerometer[2]);

[cparata]

Can you also specify the wire connections between the adapter and the SAMD51 board just to clarify your hardware setup?

[TinManAkshay]

@cparata ,

This is the schematic i am following to connect to the MCU and you will see INT_1 is kept open, so does in our custom board.

This is the result you have asked w/0 float casting. I noticed one thing in serial monitor is that 2nd value is incrementing but 3rd value (244) is steady.

[cparata]

SA0 is connected to VDD, so you must use LSM6DSR_I2C_ADD_H instead of LSM6DSR_I2C_ADD_L. Try to create the instance of the sensor in this way:

AccGyr = new LSM6DSRSensor (&Wire, LSM6DSR_I2C_ADD_H);

Best Regards, Carlo

[TinManAkshay]

I did the same actually. But the same output.

[cparata]

Could you try this sketch after connecting INT1 to A5?

#include <Wire.h>
#include <LSM6DSRSensor.h>

#define INT_1 A5

// Components
LSM6DSRSensor *AccGyr;

void setup() {

  pinMode(A5, OUTPUT);
  digitalWrite(A5, LOW);

  delay(200);

  // Initialize serial for output.
  Serial.begin(115200);

  // Initialize I2C bus.
  Wire.begin();

  AccGyr = new LSM6DSRSensor (&Wire, LSM6DSR_I2C_ADD_H);
  AccGyr->Enable_X();
}

void loop() {

  // Read accelerometer and gyroscope.
  int32_t accelerometer[3];

  AccGyr->Get_X_Axes(accelerometer);

  // Output data.
  Serial.println(accelerometer[0]);
  Serial.println(accelerometer[1]);
  Serial.println(accelerometer[2]);
  Serial.println("****************");

  delay(2);
}

[fpistm]

Do you have pull up on I2C lines ?

[TinManAkshay]

@cparata @fpistm ,

I tried connecting INT_1 to A5 and put it to low. Also, there are pull-up resistors connected to I2C lines. But output is still same. This is critical for me to get the accurate results for our project since deadline is very near. I dont know what has been going wrong here.

[cparata]

Hi @TinManAkshay , unfortunately, I do not know what could be the issue. I'm sure that this library works with STM32 because I tested it with a Nucleo board. Could you also share the schematics and Data Sheet of the SAMD MCU? Best Regards, Carlo

[TinManAkshay]

@cparata ,

I am using this dev board to connect to the LSM6

https://www.adafruit.com/product/3857?gclid=CjwKCAiAv4n9BRA9EiwA30WNDxEE0tUVJvLwKvxD4ZTIkDXRwtAhT99AqwqPrq-KR4IZ75vjzSROxhoC3hsQAvD_BwE

[fpistm]

All info here: https://learn.adafruit.com/adafruit-feather-m4-express-atsamd51/downloads

[cparata]

Hi @TinManAkshay , could you also check that your wire connections are like in this file: https://www.st.com/content/ccc/resource/technical/layouts_and_diagrams/schematic_pack/group1/98/bb/ee/ca/4d/4d/4c/da/STEVAL-MKI194V1_SCHEMATIC/files/steval-mki194v1_schematic.pdf/jcr:content/translations/en.steval-mki194v1_schematic.pdf

So: Pin1 -> Vdd Pin2 -> Vdd Pin4 -> GND Pin5 -> GND Pin22 -> Vdd Pin21 -> SDA of SAMD Pin20 -> SCL of SAMD Pin19 -> Vdd Pin14 -> A5 of SAMD putting it LOW Pin13 -> GND

All other pins can stand unconnected.

Best Regards, Carlo

[TinManAkshay]

@cparata,

Every pin's connection is up to the mark. Still same behavior. I am getting only 1st value which is changing as I move the board. 2nd value is incrementing and 3rd value is just 244.

[cparata]

Hi @TinManAkshay , could you compile my sketch and send me a snapshot of the output in the time? I would like to understand better the values that you are seeing. Best Regards, Carlo

[cparata]

For example you can find below the output that I can see with my sketch using the LSM6DSR and a Nucleo-L476RG. Unfortunately, I do not have a SAMD board to test the component also with it.

[TinManAkshay]

Hi @cparata ,

I have two MCUs ..SAMD21 and SAMD51

1 . With SAMD21, the snapshot of the output is below:

2 With SAMD51, the snapshot of the output is below:

[cparata]

Hi @TinManAkshay , do you have a Nucleo board to test the behavior of the LSM6DSR with it. In this way, if you get the same results that I see, we can exclude that the LSM6DSR board is broken. I would like also to look at your hardware setup. Could you provide a picture of that? Finally what are the values of the Pull-Up resistors that you are using on the I2C lines? Best Regards, Carlo

[TinManAkshay]

@cparata ,

No I dont have that nucleo board. I am sending you the picture of my HW setup and I have 4.7k ohms pull-ups

[cparata]

Hi @TinManAkshay , It seems to me that you connected the I2C resistors between the pins of the MCU and the ones of the LSM6DSR. In this way they are not pulled-up. Try to connect one of the pin of the resistors (for both resistors SDA and SCL) to 3.3V and give a try to this configuration. Obviously you need also to link directly the I2C pins of the MCU to the respective ones of LSM6DSR. Best Regards, Carlo

[TinManAkshay]

@cparata ,

Oh my goodness, now its working. I can notice the accurate change in the movement of the board in +- X, Y and Z axisi. I made very silly mistake and you were right that pull-ups were not pulled-up correctly. These libraries are working. What's the unit of the values? Are they in "mg's"? I'd need float values in g's for my further calculation in my project's firmware.

You can now remove the warning about the incompatibility with SAMD architecture.

Thanks a lot for your help. Appreciate it!!

[cparata]

Hi @TinManAkshay , Great! Yes, the values are in mg. In order to convert them in float values in g, you can apply a cast like this:

(float)accelerometer[0]/1000.0f ...

Or even better in order to have more accuracy, you can call the Get_X_AxesRaw and Get_X_Sensitivity doing something like that:

int16_t acc_raw[3]; float acc_in_g[3]; float acc_sensitivity;

AccGyr->Get_X_AxesRaw(acc_raw); AccGyr->Get_X_Sensitivity(&acc_sensitivity); acc_in_g[0]=((float)acc_raw[0]acc_sensitivity)/1000.0f; acc_in_g[1]=((float)acc_raw[1]acc_sensitivity)/1000.0f; acc_in_g[2]=((float)acc_raw[2]*acc_sensitivity)/1000.0f;

I will update the library properties to support also samd arch. Best Regards, Carlo

[TinManAkshay]

Got it!. In order to enhance the performance, what if I increase the sample rate as well as +-2g to +-8g?

[cparata]

You can use whatever full scale and output data rate you want according your application requirements. In order to change the full scale to 8g, you can call in the setup:

AccGyr->Set_X_FS(8);

In order to change the Output Data Rate (you can set it up to 6667Hz) you can call in the setup:

AccGyr->Set_X_ODR(6667.0f);

Best Regards, Carlo

[cparata]

Hi @TinManAkshay , I updated the library adding the samd architecture. Now, you should not see any warning during compilation. Best Regards, Carlo

[fpistm]

Hi @TinManAkshay I would just point that if you had shared you wiring setup as requested by @cparata 2days ago this would save us lot of time.

More information you provide better and faster is the support.

[TinManAkshay]

@cparata - Thanks a lot again!

@fpistm - Yeah I totally understand. I am really sorry for making a silly mistake. I really appreciate all the support. Thanks again!