diff --git a/.gitignore b/.gitignore
index de3f382f..fe0cf02a 100644
--- a/.gitignore
+++ b/.gitignore
@@ -5,4 +5,4 @@ build*
 *.autosave
 slprj*
 *.zip
-*_ert_rtw
\ No newline at end of file
+*_ert_rtw
diff --git a/ft_sensor/README.md b/ft_sensor/README.md
index d420c270..80217158 100644
--- a/ft_sensor/README.md
+++ b/ft_sensor/README.md
@@ -1,7 +1,7 @@
 # 1. Introduction
 This folder contains the Simulink model for the Force-Torque sensor firmware.
 
-<img src="assets/model001.png" width="600px">
+<img src="assets/model.png" width="600px">
 
 The model is used for code generation.
 
@@ -16,6 +16,7 @@ The model is used for code generation.
 
 # 3. The model
 The model implements the conversion from ADC raw values to Force-Torque values in the Strain2 sensor.
+The calculations inside the model are performed in single-precision (4 bytes).
 
 ## 3.1. Input
 ADC raw values are an array of 6 values, 16 bit each, type: **unsigned**.
@@ -26,9 +27,11 @@ The output is back-compatible with the current YARP implementation.
 
 ## 3.3. Parameters
 - Conversion matrix dimensions 6x6, 16 bit each, type: **integer**.
-- Offset dimensions 6x1, 16 bit each, type: **unsigned**.
+- Offset dimensions 6x1, 16 bit each, type: **integer**.
 
 # 4. Code generation
+
+## 4.1. Base operations
 In order to generate the C++ code, open the `Embedded Coder` application on Simulink tab.
 
 <br><br>
@@ -41,10 +44,39 @@ Set the correct parameters in `C/C++ Settings --> Code generation settings --> S
 
 Finally, use the `Generate code` button.
 
+## 4.2. Advanced operations settings
+
+### 4.2.1. Matrix calculation for rows
+This setting in essential for the correct code generation for matrix multiplication.
+In the firmware memory at the moment the matrix is stored as a raws array.
+
+Select the following options:
+
+From *c++CODE->Settings*
+
+<img src="assets/conf-02.png" width="600px">  
+
+
+From *Code Generation->Interface*, select `Row-major` in **array layout**
+
+<img src="assets/conf-01.png" width="600px">
+
+From *Math and data Types->Interface*, select `Use algorithms optimized ...`
+
+<img src="assets/conf-03.png" width="600px">
+
+### 4.2.2. Disable optimization for cast problems
+Select the following options:
+
+From *c++CODE->Settings*
+From *Code Generation->Optimization*, select `Minimum (debugging)` in **Level**
+
+<img src="assets/optimization.png" width="600px">
+
 # 5. Note
 
 ## 5.1. Public parameters
-The parameters needs to have public access.
-To specify that, click on the `Code Interface` button of tha App and then the `Model parameters` option.  
+The parameters need to have public access.
+To specify that, click on the `Code Interface` button of the App and then the `Model parameters` option.  
 
 <img src="assets/public-parameters.png" width="600px">
diff --git a/ft_sensor/adc2ft.sldd b/ft_sensor/adc2ft.sldd
index b33237cb..dfaf54f4 100644
Binary files a/ft_sensor/adc2ft.sldd and b/ft_sensor/adc2ft.sldd differ
diff --git a/ft_sensor/assets/conf-01.png b/ft_sensor/assets/conf-01.png
new file mode 100644
index 00000000..ebad193f
Binary files /dev/null and b/ft_sensor/assets/conf-01.png differ
diff --git a/ft_sensor/assets/conf-02.png b/ft_sensor/assets/conf-02.png
new file mode 100644
index 00000000..deb9c391
Binary files /dev/null and b/ft_sensor/assets/conf-02.png differ
diff --git a/ft_sensor/assets/conf-03.png b/ft_sensor/assets/conf-03.png
new file mode 100644
index 00000000..79597cc4
Binary files /dev/null and b/ft_sensor/assets/conf-03.png differ
diff --git a/ft_sensor/assets/model.png b/ft_sensor/assets/model.png
new file mode 100644
index 00000000..d21b5d1b
Binary files /dev/null and b/ft_sensor/assets/model.png differ
diff --git a/ft_sensor/assets/model001.png b/ft_sensor/assets/model001.png
deleted file mode 100644
index 0c06460d..00000000
Binary files a/ft_sensor/assets/model001.png and /dev/null differ
diff --git a/ft_sensor/assets/optimization.png b/ft_sensor/assets/optimization.png
new file mode 100644
index 00000000..98040671
Binary files /dev/null and b/ft_sensor/assets/optimization.png differ