Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Zero offset, zero drift or something else? 1

Status
Not open for further replies.

E2015

Civil/Environmental
Apr 20, 2015
22
I have some strange readings from an impact hammer and acceleromeeters. The hammer is from Piezotronics PCB Model 086D50 and the accelerometers are from Dytran model 3333A and 325M24.

On the first picture called Force you can see the plotted force magnitude in Newtons with number of samples. I have applied in total 4 impact with the hammer and the approx distance between each impact was approx 10 seconds. The force plot like this looks normal.

Figure 1 Force
Force_ss5va5.jpg


However, if you zoom in the force plot close to the hammer response at rest, you can see that there are fluctuations in the data. This you can see on the second attached picture named ForceZoomedIn. Red arrows point on the events between impacts. Between each impact there is a distance of approx 10 sec and for this interval the force from the hammer should be around 0. But why do I have these jumps in the reading from x-axis when the hammer is at rest ? Can you clarify why this happens with the data reading from the hammer?

Figure 2 ForceZoomedIn
ForceZoomedIn_mktkgp.jpg


The same problem occurs with readings from the accelerometers. The attached figures are showing acceleration reading as result of 4 hammer impacts.
On the third picture called Acceleration you can see the plotted acceleration magnitude in m/s^2 with number of samples. The four acceleration peaks are result from four hammer impacts where distance between each impact is approx 10 sec. The acceleration plot like this looks normal.

Figure 3 Acceleration
Acceleration_myysy7.jpg


Again, if you zoom in the acceleration plot close to the acceleration response at rest, you can see that there are drifts in the data over time. Red arrows point on the events between impacts. The fluctuations are present for both accelerometer models. This you can see on the fourth attached picture named AccelerometerZoomedIn.

Figure 4 AccelerometerZoomedIn
AccelerationZoomedIn_bb6i2i.jpg


Between each event in the acceleration there is a distance of approx 10 sec and for this interval the acceleration reading should be around 0 because there is no excitation of the structure. But why do I have these jumps in the noise reading from x-axis when the vibration dissipates out?

Thank you!




Emina B.
 
Replies continue below

Recommended for you

Sample Number means nothing without knowing the Sample Rate and F-max. Why not present your X-axis in Time units? Typically there is only one impact per time record, so why the closely (apparently) spaced impacts? It is hard to tell whether your data and concern about it is significant.

Walt
 
Here are the plots of force and acceleration with x-axis representing time in seconds.
Figure 1 Force
ForceTime_bbjvr2.jpg

Figure 2 Force Zoomed IN
ForceTimeZoomedIn_xxmmqh.jpg

Figure 3 Acceleration
AccelerationTime_ujus2i.jpg

Figure 4 Acceleration Zoomed IN
AccelerationTimeZoomedIn_cp2osx.jpg


Are the jumps in between impacts, or settling in the data for hammer and accelerometers in between impacts normal?
Can I consider these measurements as OK?
Is there a way how I can center the data around 0 on the x axis but for events in between impacts?

Emina B.
 
My first thought was discharge time constant but your low amplitude signal looks quite wavy. Can you share the plots of transfer function as well as the coherence from the 4 averages?


Kind regards,
Jason
 
I have the same model PCB hammer. It has a Microdot cable connector on the end of the wood handle. Swinging a heavy sledge hammer with small cable and connector caused a lot of signal problems until I refitted a larger cable and connector. I can't comment further on the signal "noise", without knowing the entire instrument system and settings. I suspect the data shown is the result of poor test procedure and settings.

Walt
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor