12-03-2017 12:33 PM
Hello,
I am using the attached VI and trying to save data of a sewing machine - Sound and 3xaxis vibration using NI 9234 and cdaq USB NI 9171.
The values started high and then went down.
I am wondering which one is the right sound pressure level for the sewing machine.
Which the right Accelerometer value for X-Axis , Y-axis and Z_axis.
Can anyone explain this values for me, am I doing sampling rate wrong ?
I am taking continues samples at 1000 Hz.
Thanks.
Solved! Go to Solution.
12-05-2017 02:14 AM
With 1kSPS you might get resonable values up to 450Hz ...
On the first run I always try to sample as fast as possible (and if possible monitor on a scope) to get an idea of the signals...
Next thing: A sensor always report the values he feels, that might or might not be the values of your intrest. Or in other words: What do you want to measure and where and how is your sensor mounted?
And a correct setup (known sensor/channel/scaling) is mandatory 😄
12-05-2017 12:02 PM
Dear Henric,
Thank you for your reply.
I am using IMI (PCB) 3-axis Accelerometer, and PCB 378B02 Microphone.
I am trying to differentiate between the sound level of Sewing machine and vibration in 3-axis and in different conditions as follows:
1- Sewing machine running full speed without fabric.
2- Sewing machine running full speed with single layer fabric.
3- Sewing machine running full speed with two layers fabric.
4- " " " " three layers fabric.
5-........................etc.
The microphone is mounted as near as possible from the sewing needle and feeding dog of the sewing machine, while the 3-axis accelerometer is nears to machine shaft and next to the motor side.
Please inform me if you need further information.
Warmest regards.
12-08-2017 09:15 AM - edited 12-08-2017 09:17 AM
Review of your data indicates sensors settling after IEPE and AC coupling have been enabled. I recommend that your measurement software wait for settling before saving measurements to file (or save twice as much data and clip the data that is unusable). Generally, and forgive my very loose math, the cutoff frequency of the AC coupling filter determines a relevant time constant for settling (0.5 Hz --> 2 s), and I would wait 5 time constants (5*2 = 10 s) for a settled measurement after configuration changes.
This settling is only expected the first time you connect IEPE sensors up to a 9234. So long as the measurement system is used for IEPE sensor measurements and remains powered, I expect that subsequent runs would not show this behavior.
Some other recommendations:
Configure the DAQ Assistant for your sensors. Instead of using the default sensor sensitivities, at least use nominal sensitivities for each sensor. Better yet, use calibrated sensitivities.
Configure the Write to LVM to include a time channel.
12-08-2017 09:34 AM - edited 12-08-2017 09:38 AM
Oh, and I definitely agree with Henrik. Sample faster.
You may not need the full bandwidth of the 9234, but signal components with frequencies up to 12.8 kHz may contribute to loudness. Normally, vibration frequencies that contribute to user safety/control/comfort are lower, but the extra bandwidth will only serve to give greater confidence that you are not attenuating signals of interest. Sample at 51.2 kHz. If data size becomes an issue, only save measurement data after the data has settled and reduce the duration of the file. Also, if you find that your signals don't contain components above 11.5k, then you can confidently decrease the sample rate to 25.6 kHz.
12-08-2017 09:42 AM
Dear Doug,
Thank you for your reply.
1- Would you please explain how can I make my "measurement software wait for settling before saving measurements to file "???
2-Do you mean add frame structure at the beginning with delay functions?
Or
Should I run the first measurement session for 10 seconds and neglect the first reading , like flushing the pipes when changing a new chemical material in container?
sorry for the example but really I need confirmation from you.
Warmest regards.
Ihab
12-08-2017 12:25 PM
simple solution could look like your suggestion, or it could look like acquire and don't save first ten blocks:
or you could add a test to see whether the measured level had settled and save some finite number of blocks:
12-08-2017 02:25 PM
Dear Doug,
Many thanks.
I am really appreciated.
Ihab