Experimenting with Vibration Sensors Part 2: The Journey

27 Jan 2021

Difficult times

Yes, everyone know this, but this doesn´t have to be a stopper. I'm facing with problems to access to the laboratory with the autonomous vehicles.

So, I'm going to show my progress and in the future I'll make another blog with the rest of the experiments.

Car test

The first real test in the field it was in my personal car, I attached the sensor to and ECU that is under the steering wheel. The sensor Z axis was in alignment with the earth's gravity.

In the picture above we can see three sections, the first one is on a stone paved road, there's a lot of vibration with high acceleration, with a peak of 29.19ms² and −32.56ms²

Then I made an stop when the vehicle almost reached the noise floor of the sensor. After that, I was driving on an asphalt road, in this section we can see that there's less acceleration. But in certain moment, I hit a pothole and we can see it with a initial peak of 18.28ms² and the another peak of 19.72ms² , I guess the first peak was when the tire enter in the pothole and the second one when it leaves the pothole.

Autonomous vehicle test

Before the lab was closed due to covid, I was able to save a log. In the next picture we can see the results. The sensor was attached to one of the ECUs that I designed. In this case, the sensor Z axis was in alignment with the earth's gravity too, but the ECU components are in the X axis, so, is required to do more measurements in this axis. I adjusted the Y axis of the chart to have the same that the first chart, in this way we can see that in this test we have a much less acceleration, with similar values as the ones from the asphalt road, but with smaller frequency.

With this initial results, we can say that the vibration is less in the autonomous vehicle (smaller acceleration and smaller frequency). But this data is not enough to conclude that the ECUs will survive in this environment.

The Code

I used STM32Cubeide to coding and STM32Cubemonitor to log the data. So, it was really easy to configure the ADC using this tools.

This is the part where the conversation is done, it is inside an infinite loop.

/*## Start the conversion process #######################################*/
   if (HAL_ADC_Start(&AdcHandle) != HAL_OK)
   {
     /* Start Conversation Error */
     Error_Handler();
   }


   /*## Wait for the end of conversion #####################################*/
   /*  For simplicity reasons, this example is just waiting till the end of the
       conversion, but application may perform other tasks while conversion
       operation is ongoing. */
   if (HAL_ADC_PollForConversion(&AdcHandle, 10) != HAL_OK)
   {
     /* End Of Conversion flag not set on time */
     Error_Handler();
   }
   else
   {
     /*## Get the converted value of regular channel  ########################*/
     uhADCxConvertedValue = HAL_ADC_GetValue(&AdcHandle);
     res = (float)(uhADCxConvertedValue);
     res = res * 3.3 / 65536;
     res = ((res*1000)-1650)/20;
     Vibration = res;




     if(Vibration >= 3000 || Vibration <= 1000){
    BSP_LED_On(LED1);
     }else {
    BSP_LED_Off(LED1);
}
   }

The result of the conversation it was converted to volts using this formula ADC·3.3V65536 , then I converted the voltage to acceleration using this formula
((Volts·1000)−1650)20mVms²

This gives us the result of the measurement in ms²

The log

STM32Cubemonitor can show us the variable result in a nice chart, but we can log this data in our computer, we only have to enable the log option

This will save the log in a weird file ".stcm", basically this is a json file. So to be able to see the data we have to convert it using this flow.

[{"id":"cc0c8ae8.c69f98","type":"tab","label":"Convert stcm","disabled":false,"info":"This flow is used to convert stcm file in csv file\n\nThe file name must be entered in the \"set file name\" node\n\nThe csv output format is variable Name, x, y\nNote : the data may be processed by group, so the x value may not be in time order. The csv file may be sorted on the x value to ensure order coherency. "},{"id":"a9caceb2.eb94d","type":"file in","z":"cc0c8ae8.c69f98","name":"Read file","filename":"","format":"lines","chunk":false,"sendError":false,"encoding":"none","x":280,"y":360,"wires":[["99d027a1.b92d98"]]},{"id":"9350e2b8.9aa3e","type":"inject","z":"cc0c8ae8.c69f98","name":"Start","topic":"","payload":"","payloadType":"date","repeat":"","crontab":"","once":false,"onceDelay":0.1,"x":110,"y":160,"wires":[["3371f5e6.b3b50a"]]},{"id":"99d027a1.b92d98","type":"json","z":"cc0c8ae8.c69f98","name":"Parse each line","property":"payload","action":"","pretty":false,"x":480,"y":360,"wires":[["badf8efa.bacee"]]},{"id":"badf8efa.bacee","type":"function","z":"cc0c8ae8.c69f98","name":"Format data","func":"let output=\"\";\nmsg.payload.variabledata.forEach (data=>output=output+ msg.payload.variablename+ \",\"+ data.x + \",\"+ data.y+\"\\n\");\nmsg.payload= output;\nreturn msg;","outputs":1,"noerr":0,"x":690,"y":360,"wires":[["38fb106e.c00f"]]},{"id":"d2859af6.e93368","type":"file","z":"cc0c8ae8.c69f98","name":"Write csv","filename":"","appendNewline":false,"createDir":false,"overwriteFile":"false","encoding":"none","x":920,"y":580,"wires":[[]]},{"id":"3371f5e6.b3b50a","type":"change","z":"cc0c8ae8.c69f98","name":"Set file name","rules":[{"t":"set","p":"filename","pt":"msg","to":"C:\\GIT\\test\\test3.stcm","tot":"str"}],"action":"","property":"","from":"","to":"","reg":false,"x":330,"y":160,"wires":[["a9caceb2.eb94d","593d0b11.8d3654"]]},{"id":"f3d3a57a.c56e48","type":"comment","z":"cc0c8ae8.c69f98","name":"Enter your filename here","info":"","x":330,"y":120,"wires":[]},{"id":"38fb106e.c00f","type":"change","z":"cc0c8ae8.c69f98","name":"Set csv extension in filename","rules":[{"t":"change","p":"filename","pt":"msg","from":"stcm","fromt":"str","to":"csv","tot":"str"}],"action":"","property":"","from":"","to":"","reg":false,"x":700,"y":580,"wires":[["d2859af6.e93368"]]},{"id":"b1ba9fe.177b06","type":"file","z":"cc0c8ae8.c69f98","name":"Eraze csv","filename":"","appendNewline":true,"createDir":false,"overwriteFile":"true","encoding":"none","x":780,"y":160,"wires":[[]]},{"id":"593d0b11.8d3654","type":"change","z":"cc0c8ae8.c69f98","name":"Format csv header ","rules":[{"t":"change","p":"filename","pt":"msg","from":"stcm","fromt":"str","to":"csv","tot":"str"},{"t":"set","p":"payload","pt":"msg","to":"Variable, x, y","tot":"str"}],"action":"","property":"","from":"","to":"","reg":false,"x":590,"y":160,"wires":[["b1ba9fe.177b06"]]},{"id":"796c6b18.2aab34","type":"comment","z":"cc0c8ae8.c69f98","name":"Process the stcm file","info":"","x":490,"y":320,"wires":[]},{"id":"c2ec88a.7368678","type":"comment","z":"cc0c8ae8.c69f98","name":"Write the csv file","info":"","x":820,"y":540,"wires":[]}]

I took this from this site.

To use it, go to menu import , copy paste the text, and select "new flow".

Then for each conversion :

1) in the box "Set file name", enter the name of file to convert (absolute path + extension : "C:\log\test1.stcm")

2) press start. The file will be processed and stored with the csv extension.

Conclusion

My first impression is that this sensor has a lot of potential. But I want more information in the datasheet and also characterization data. For example, some output curves at different temperatures.

I was able to do some measurements in two different environments and with the results I can say that the vibration in the autonomous vehicle is less in strength and less frequent. This gives us the impression that If we are design our ECUs as we're going to use it in a normal car, it will survive a less dangerous environment. But this is not the end, we can do a lot more measurements and have a better conclusion. Hopefully this covid madness ends soon and I can continue with this journey.