TivaWare sensorlib example compdcm_mpu9150 bug

Hello Chris,

Yes, we are aware of this issue. And we did clearly miss it during the last TivaWare release but I have now filed an internal bug on the same.

Thanks.

Regards
Amit

My fix is here, it is not the greatest solution most likely but much simpler / lighter than rewriting UARTprintf to parse floats.

This replaces the loop in compdcm_mpu9150.c that decomposes the float into integers and the list of UARTprintf statements that compose the table with the same net result (without the 0 to -1 bug).

            const uint8_t table_offset[16][2] = {
            		{5,17}, {5,40}, {5,63}, // Accel
			{7,17}, {7,40}, {7,63}, // Gyro
            		{9,17}, {9,40}, {9,63}, // Mag
			{14,17}, {14,40}, {14,63}, // Euler
			{19,14}, {19,32}, {19,50}, {19,68} }; // Quat

            uint8_t neg_flag = 0; // flag for numbers between 0 and -1

            for(ui32Idx = 0; ui32Idx < 16; ui32Idx++)
            {
                //
                // Conver float value to a integer truncating the decimal part.
                //
                i32IPart[ui32Idx] = (int32_t) pfData[ui32Idx];

                //
                // Multiply by 1000 to preserve first three decimal values.
                // Truncates at the 3rd decimal place.
                //
                i32FPart[ui32Idx] = (int32_t) (pfData[ui32Idx] * 1000.0f);

                //
                // Subtract off the integer part from this newly formed decimal
                // part.
                //
                i32FPart[ui32Idx] = i32FPart[ui32Idx] -
                                    (i32IPart[ui32Idx] * 1000);

                //
                // make the decimal part a positive number for display.
                //
                neg_flag = 0;
                if(i32FPart[ui32Idx] < 0)
                {
                    i32FPart[ui32Idx] *= -1;
                    if (i32IPart[ui32Idx] == 0)
                    {
                    	neg_flag = 1; // fix for number between 0 and -1
                    }
                }

                //
		// Print the acceleration numbers in the table.
		//
                if (neg_flag)
                {
		   UARTprintf("\033[%d;%dH -0.%03d",
			table_offset[ui32Idx][0], table_offset[ui32Idx][1],
			i32FPart[ui32Idx]);
                }
                else
                {
                   UARTprintf("\033[%d;%dH%3d.%03d",
                	table_offset[ui32Idx][0], table_offset[ui32Idx][1],
			i32IPart[ui32Idx], i32FPart[ui32Idx]);
                }
            }

Hello Chris

Thanks again for posting the code. But we would like to make sure that a permanent change in most concise manner after testing is incorporated.

Regards
Amit

Amit,

Yes I understand, thank you. This is just a fix to get people by until your permanent change comes.. Thanks for addressing this.

Regards
Chris