Q:
How can I use the ADIS16228 evaluation tools to determine if this product meets my performance needs?
-----------------------------------------------------------------------------------------------------------------------------------
A:
This post illustrates an experimental approach of observing the vibration signature associated with a example system, using the ADIS16228 and its evaluation tools. In this example, we decided to find out if we can use the ADIS16228 to determine the operational state of a compressor, which produces cold air in a temperature control chamber. For this compressor, there are three different states, which appear to have different vibration profiles: OFF, IDLE and ON. The combination of the ADIS16228/PCBZ, ADISUSB, a laptop and the ADIS16228 evaluation software provided a convenient platform for analyzing the vibration on the compressor's enclosure. Since we didn't want to drill holes in the enclosure (at least to start with), we attached the ADIS16228/PCBZ to the top side of the enclosure (in the middle), using double-sided tape. In order to minimize the mass of the sensor, we used a 6" ribbon cable to connect the ADIS16228/PCBZ (J1) to the ADISUSB (J1), as shown below. For complete setup guidelines, please see the ADIS16228-ADISUSB User Guide Wiki.
Once we secured the ADIS16228/PCBZon the surface of the enclosure, we plugged the ADISUSB in to the PC-USB port and opened the ADIS16228 Evaluation Software package. Under Main Control, we established (or verified) the following settings:
The following plot showed us that there appears to be vibration, but at levels that are much lower than the 20g range.
After reviewing this plot, we clicked on the Scale(g)button to scroll down to a range, which better represented the FFT data.
After observing that the peak vibration energy was at lower frequencies, we changed to the the Sample Rate setting of SR1 and selected a Range of 0 to 5g for this Sample Rate. From the following plot, we can observe the impact of finer frequency resolution.
We also experimented with using a Sample Rate setting of SR2, a Range of 0 to 5g and 8 FFT Averages, in order to reduce some of the variation in the vibration signature. Again, after all settings were complete, we hit Startto trigger the capture/FFT analysis process. In the plot, we observed a large peak at bin 193, which was ~0.1288g (see marker).
Using the same settings, we put the compressor into IDLE mode and ran another FFT (observed a large change in the peak energy. As we can see, the energy at bin 193 was greatly reduced. From this information, we have a clear difference in FFT signature, between the ON and IDLE states.
In order to find a signature difference between the IDLE and OFF modes, we used the Scale(g) button to scroll down to a lower range option. In doing this, we found "peaking" at bin 80, which was ~0.0277g.
When the compressor was in an OFF state, we hit Start again and found that the energy at bin 80 reduced by a large amount.
From this simple experiment, we can conclude the following:
In concluding this post, we also want to note that using double-sided tape was helpful for this quick demo, but would expect that a complete investigation would want to consider a more rigid attachment approach. The next step in this process will be to determine the appropriate spectral alarm settings for detecting these conditions. More on that in our next post!
The EVAL-ADIS is now the preferred evaluation tool for the ADIS16228. Check out the following link to download the Vibration Evaluation Program:
http://www.analog.com/en/evaluation/EVAL-ADIS/eb.html#SOFTWARE_AND_TOOLS
Check out the following link for the ADIS16228 PC-Evaluation Wiki Guide:
http://wiki.analog.com/resources/eval/user-guides/inertial-mems/imu/adis16228-pc-eval
We will be conducting similar exercises, using the newer evaluation tools in the near future.
Correct me if i am wrong, interleaved sampling would give data for a FFT out to about 2-2.5kHz?
I need to think about what implications that decimation could give.
So we need minimum about 11kSPS per channel for our use. Olov
Yes 20.48kSPS 3 ch would be useful in our current product range. I tried to email further but it bounced. You can email olov.li AT vtab.se
Olov
The ADIS16228 only supports continuous, real-time data acquisition at 20.48kSPS on one channel at a time. You can switch between channels, but you need to throw out the data from the first three data cycles, after switching to a new axis. This limits the effective data rate to 5.12kSPS per axis, when interleaving the data reads. We didn’t market this heavily, since the core sensor resonance was in the same region of this sample rate. Would having 20.48kSPS available on all three channels be helpful only during your design process or could you envision that being part of your final product implementation? Feel free to email me if this information is better communicated in private.
Best,
NevadaMark
Ok, looking forward to that. Don´t forget to take a look at the frequency Hz calculation of the marker reading, it is not making sense to me depending on AVG setting and what Sx is used. In our custom hardware we can interrupt on the DIO pins, can we read data from X, Y and Z in streaming mode? I believed you only could read one selected at a time. Olov