v1.0 - SEP Step 5 - Detailed Design - Data Acquisition (DAQ)

Added by Jeremy Wright about 10 years ago

While documenting the Shepard v1.0 DAQ on the wiki, I noticed a major oversight regarding the FSR. The FSR that was selected has an effective measurement range of 0.1 to 10 Newtons. This is a big problem due to the fact that some of the D and E motors that we've tested have a peak thrust in the range of 20 to 30 Newtons. The oversight was not noticed by anyone who saw the FSR spec document, and there was a lack of due diligence on my part to double check the specs against the peak motor thrusts and do the proper conversions.

This has become a moot point now though as the lack of accuracy and drift that we've seen in the FSR make it unsuitable for even a level 1 kite. We'll be evaluating other options soon. Even so, in the interest of transparencty and "sharing our failures" I've documented this oversight here and on the wiki.

Replies (1)

RE: v1.0 - SEP Step 5 - Detailed Design - Data Acquisition (DAQ) - Added by Jeremy Wright about 10 years ago

We're seeing very low sample rates of 14 to 15 samples per second on the finished Shepard v1.0 DAQ setup. I suspected that maybe we were waiting for the thermocouple amplifier since it uses a SPI interface and does 14-bit analog-to-digital conversion internally. After looking through the documentation for the MAX31855 chip on the thermocouple amplifier we're using, I noticed that it has an typical conversion time of 70 ms with a max of 100 ms. That would put us right in the range of 14 to 15 samples per second if we were waiting on this amplifier.

To increase the sample rate maybe we can use an analog thermocouple amplifier like the AD595 . The analog input reads on the Arduino are orders of magnitude faster than waiting for the MAX31855 to finish its conversion.

We could sample the temperature less often as well, but with the MAX31855 that still means that every time we do sample the temperature we're going to have a lag in our FSR sample rate of 70 to 100 ms.