Data collection with micro controllers-Suggestions?

[wob2015]

Hi,

I'm using an accelerometer to record an impact, and currently Im using a data acquisition device to do so. The problem is that it is large and expensive.
What I would like to do is collect data from an accelerometer and do calculations on it through either C or assembler. The results of those calculations would then sent wirelessly to a tablet or laptop.
As far as i can tell, my best bet would be to draw out a design for a board which uses something like the ADUC7061. This micro controller has an ARM 7 with multiple 24 bit ADC inputs to handle the output form the accelerometer, along with the usual flash memory, DAC outputs, etc.
That being said this is the first time i've attempted to design something with a micro controller, so any suggestions or other opinions on how to accomplish this would be very much appreciated!

Thanks,

Wesley

 

[MacGyverS2000]

Have you ever programmed a blank microcontroller? If not, there's a learning curve to get over... bootloaders, peripheral setup, etc.

Is the current DA device doing any signal pre-/post-processing? If so, you'll need to recreate that. If the current DA is bulky/expensive, there may be a reason for that.

If you do go down this road, you'll probably be better off purchasing a dev kit and using a COTS board rather than trying to spin your own.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[wob2015]

MacGyverS2000 thanks for the reply.

I cant say I have, but i am aware that I have quite a learning curve ahead of me. I believe this micro controller has a dev kit i can purchase along with it, which includes an eval board, Kiel IDE, some PLA tools, and provides interfacing for the on board emulator. As for the bootloader, I believe this is taken care of through the UART serial connection to the controller so I can work on it while it is in circuit, but i may be mistaken.

The current DA device is not doing any signal processing, the accelerometer is IEPE and therefore does not need to be conditioned. It has a number of features that I do not need, such as signal generation, 4 inputs, simultaneous capture, etc.

Ill take a look into those COTS boards.

 

[MikeHalloran]

Are you gathering data, or are you hoping to sell accelerometer boxes?

I.e., if you're sticker-shocked over the cost of a tool, maybe you can find a less finished tool, like a tested board assembly that's not packaged, linearized, etc.

The danger of designing a board yourself is that debugging a new design can suck up your time to such an extent that you make no progress on whatever you were collecting data for. ... even if it works on the first try.





Mike Halloran
Pembroke Pines, FL, USA

 

[MacGyverS2000]

As for the bootloader, I believe this is taken care of through the UART serial connection to the controller so I can work on it while it is in circuit

This is fine for debugging, but you need a bootloader of some form for when the board is not attached to a computer... you said you wanted something smaller than what you have now, so hauling around a laptop connected to the board via USB or RS-232 cable is probably not ideal. Depending upon the processor, such a bootloader may be very simple or very complicated.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[IRstuff]

I think there may be Arduino or Parallax boards that could do the job with possibly a second card specifically for SD memory, ala:

http://phillipecantin.blogspot.com/2012/01/arduino-and-micro-sd-logging.html

or:

http://www.parallax.com/tabid/768/ProductID/597/Default.aspx

Both of these controllers are designed for standalone operation, and in the case of Arduino, you can do your programming on the PC and the GUI will download and burn your app into the board.

TTFN
faq731-376

 

[wob2015]

Thanks for the responses guys,

I am actually a mechatronics eng co-op op student and as you can tell from my name im still far off from graduating, so the help is appreciated.


I'm collecting accelerometer data as part of a testing system for our product. The issue is not price, it is more of the size of the DAQ and as far as i can tell I cannot transmit analog signals wirelessly with the accuracy i need. My solution to it was to convert the signal to digital on the testing unit, then transmit the digital signal to the laptop. However since i have it going through a processor and loading it into memory already, I figured i may as well do some of the calculations on board the rig.
Can I expand the processing power of the arduino if need be? I have algorithms that involve numerical integration that need to be run, and id rather run them on this board instead of sending the entire data set across wireless and have the laptop do it.

Wes

 

[IRstuff]

There's no easy way to do that. Perhaps you should be looking at a single PC, which would have sufficient memory and processing power to do numerical integrations

TTFN
faq731-376

 

[wob2015]

hhmmm ok. Maybe ill leave the main calculations to the laptop then. I would imagine there would be enough processing power to do a simple comparison? All i really care about is the data which is non-negative. Since im recording an impact, that would account for 50 data points out of the 50 000. If I could find the 50 data points and only transmit those, it would greatly reduce the transmit time.

 

[IRstuff]

That should be doable.

TTFN
faq731-376