Hi there guys,
I know that there must be a loads of topic's with similar problems where I could find a few answers and mix them up to suite my meeds but, please don't kill me, I'm making it as a part of my brachelor thesis in mechanical engineering and (of course..) I'm veeery much off schedule so i'm sitin' my ass writing about rocket engines and didn't have time for extensive search..
Here's meritum of the thing, Im building laboratory station which is measuring thrust of modelar rocket engine (solid fuel) attached to the little table moving in one horizontal axis, at one side there'll be spring which will not allow the tabe to run for its live and at the other there'll be a tensometric bar.
Signal is taken from it by whitston'e bridge so changes of tensometer resistance is transform into analog signal of changing voltage. this signal I want to put in analog input pin in arduino.. and here's the problem because, as I said before, I'm trying to be a mechanical engineer so I'm kind of week with programming.. not completely but still it's to big for me now so I'm asking for your help.
To be clear, I'm not asking you to do the dirty job and write a whole program for me, I'm asking for something like exact big hints.
Second thing is that I'would want my arduino to measure and also save results on for example sd card and i'm not completely sure if my arduino nano with atmega328 will manage to do both things..
If you'll help me I promise to send you right here some photos and movies from eventual firings
tensometric bar is limited to 10kg It's even to much for me so I'll use max half of it so we've got 5kg range. Ill make signal from witestone's to be max 1 volt so I think we can use internal analog reference voltage 1.1V.
If my math is right we should have accuracy of 2g but i dont need that much, accuracy like 20-50g on step would be excelent
Second thing is that I'would want my arduino to measure and also save results on for example sd card and i'm not completely sure if my arduino nano with atmega328 will manage to do both things..
Rocket engines i have are burning out in less than 3 seconds so i think that bottom line to get some acceptable amount data is 1MHz, I would like it about 10 MHz but Im not sure if it stands.. Is it?
Edit:
I forgot about number of measurements couse focused on time and freuency.. If I assume that maximum time of measure will be 1 minute (we need to turn it on before igniting to get sample of 0 thrust state and turn off a few seconds after) with probing in 10 MHz freuency so it gives us about 600k measurments
Your high speed and hight volume data acquisition needs will probably be taking you away from "Arduinoland" and into some customized hardware.
The standard 10 bit analogRead() takes a bit over 100microseconds and gives about 9600 samples/sec. You can study the Atmega 328 data sheet, and figure out how to change the ADC clock prescaler, use free running mode, and switch to 8 bit resolution but I'm not sure you can get over 150K samples/second. You will have some research to do to maximize the conversion rate of the Arduino ADC. Heres an article to get you started ATmega ADC tutorial | Open Music Labs
Once you get the data, storing or off loading the data is another major challenge.
Specialized hardware for high speed data acquisition and storage is an other topic.
It's lucky you are a rocket scientist, because this project is not simple.
No no no no.. we have a major missunderstanding here.. I think that i mislead you with that "600k measurments" I was thinking that you want me to tell you TOTAL number of measurments during a test for calculating needed space in inner buffer or something like that...
If we speak about measurments per second then i need at least 1k/sec. and my desire value is 10k/sec.
Maby I did get something wrong because I was thinking that frequency telling me how many measurements is made in second.
Edit:
Now I did read that link to atmega adc, did you guys was thinking about clock frequency when you did asked me for "frequency"?
TSNG:
Rocket engines i have are burning out in less than 3 seconds so i think that bottom line to get some acceptable amount data is 1MHz, I would like it about 10 MHz but Im not sure if it stands.. Is it?
Edit:
I forgot about number of measurements couse focused on time and freuency.. If I assume that maximum time of measure will be 1 minute (we need to turn it on before igniting to get sample of 0 thrust state and turn off a few seconds after) with probing in 10 MHz freuency so it gives us about 600k measurments
MHz means "million cycles per second". Please learn that. We are not misunderstanding, you are.
you're right, sorry, I'm mesed up a bit right now..
so now, with freuency of probing at 10kHz I assume it's is possible. So do you have some hints for me how to program that to read analog input and write it to sd card for example in txt format?
So do you have some hints for me how to program that to read analog input and write it to sd card for example in txt format?
High speed logging of data to an SD card requires buffered writes, and is somewhat complex. The best place to get started is with the SdFat library. (not the standard SD library which comes with the IDE) One of the examples is called "AnalogBinLogger" and is set up to write 5000 samples from the Arduino ADC to an SD card. I would expect that it could be pushed faster. Using a Mega will help overcome some problems with SD card latency in that it can hold more buffered data.
You may also want to look at pushing your data into a computer.
The fastest baud rate in IDE 1.6.6 is 250,000 baud. You may be able to go faster with other terminal programs.
At 250000 baud it takes 10 bits to get a byte so that is 25,000 bytes per second so for a 16 bit word that will be 12,500 words a second. If you go to 8 bit ADC resolution your back to 25,000 data samples over serial.
High speed data acquisiton and storage is not an area I know much about, so I'm not sure where you put all this data.
