Change font size
It is currently Sun Aug 18, 2019 9:20 pm
www.macrofibercomposite.com
Board index How-To's

How-To's

About the power frequency signal measured by MFC sensor

Question and Answers for making the MFC work

Moderator: DRK

About the power frequency signal measured by MFC sensor

Postby zilanqingyun » Mon Mar 25, 2019 10:08 am

Hello

I have a MFC sensor(2814 P1) for vibration measurement. However, when I connect the MFC sensor to my charge amplifier, the power frequency signal (50Hz,100Hz,150Hz,200Hz......) are very prominent. As a result it can not meet the requirement of signal-to-noise ratio.

Do I miss any equipment for the MFC sensor or is there any requirement for the charge amplifire (to remove the power frequency signals)?

Thanks!
zilanqingyun
 
Posts: 7
Joined: Fri Mar 08, 2019 3:52 am

Re: About the power frequency signal measured by MFC sensor

Postby TPD » Mon Mar 25, 2019 1:25 pm

Hi There!

did you bond/glue the MFC to a structure so it is properly strained? Or better how do you strain the MFC?
Normally the generated charge should be much higher than the environmental EMs.
Did you use a shielded wire to feed the charge amplifier?
How long is your cable?

If your strain signal is very weak (<1µm) you might want to attach a copper foil for shielding on top of the MFCs.

In general we recommend the P2 type MFCs as strain sensor, because it has about a 10times higher capacitance for the same size MFC than the P1 type. This means a lower impedance and a better SNR in general.
Also check out this publication here in this regards https://www.smart-material.com/media/Presentations/MFC-strain-sensor-ISPA-2009.pdf
TPD
Smart Material Support Team
---------------------------
"That's not right, it's not even wrong."
Wolfgang Pauli, 1945 Nobel Laureate - Physics.
TPD
 
Posts: 98
Joined: Tue Aug 11, 2015 12:46 pm
Location: Florida, USA

Re: About the power frequency signal measured by MFC sensor

Postby zilanqingyun » Mon Apr 08, 2019 4:06 am

Thanks for your answer!

Yes I bond MFC to a steel plate by a 3M double sided tape, so I think it is properly strained. Besides, the wire to feed the charge amplifie is shielded since when an accelerator is connected to the charge amplifire, the power frequency signal (50Hz,100Hz,150Hz,200Hz......) is very weak. Length of the wire connecting MFC sensor to charge amplifire is less than 1 metre.

I want to use MFC sensor to measure the vibration of a host structure. I wonder whether the charge amplifire is not suitable for MFC sensor or do I miss any instruments for a MFC sensor?

Thank you :lol:
zilanqingyun
 
Posts: 7
Joined: Fri Mar 08, 2019 3:52 am

Re: About the power frequency signal measured by MFC sensor

Postby TPD » Mon Apr 08, 2019 7:33 am

Please measure the signal with an oscilloscope directly
and disconnect the charge amplifier for this measurement.
So connect the MFC directly with a standard > 10MOhm probe to your oscilloscope. Let me know what you measure.
This is to confirm what the MFC generates as a signal first.
There are multiple reason why the charge amplifier
might not work correctly. So let's verify first what the output from the MFC looks like.
TPD
Smart Material Support Team
---------------------------
"That's not right, it's not even wrong."
Wolfgang Pauli, 1945 Nobel Laureate - Physics.
TPD
 
Posts: 98
Joined: Tue Aug 11, 2015 12:46 pm
Location: Florida, USA

Re: About the power frequency signal measured by MFC sensor

Postby zilanqingyun » Tue Apr 09, 2019 5:20 am

Thanks TPD

I connect the MFC sensor to an oscilloscope directly while the sensor is bonded to a cantilever beam. However, the signal seems to be too weak to be detacted. I just saw some cluttered waveforms, which look like impact waveform on the screen. When I gave an excitation(sinusoidal signal) to the cantilever beam, the graph shown on the oscilloscope did not show a corresponding sinusoidal signal. Maybe the charge of MFC sensor is too weak to be detected.
zilanqingyun
 
Posts: 7
Joined: Fri Mar 08, 2019 3:52 am

Re: About the power frequency signal measured by MFC sensor

Postby TPD » Tue Apr 09, 2019 12:59 pm

Please send details about the plate (thickness, material),
if it is properly clamped on one side, where is the MFC located, how much did you bend the beam (deflection)?
TPD
Smart Material Support Team
---------------------------
"That's not right, it's not even wrong."
Wolfgang Pauli, 1945 Nobel Laureate - Physics.
TPD
 
Posts: 98
Joined: Tue Aug 11, 2015 12:46 pm
Location: Florida, USA

Re: About the power frequency signal measured by MFC sensor

Postby zilanqingyun » Tue Apr 09, 2019 9:16 pm

Size of the beam:
length:530mm width:50mm thickness:2.3mm
Material: structural steel

The MFC is located close to the clamped side, where the distance is about 177mm from the clamped side.

I gave a sinusoidal excitation to the actuator,which is located at the clamped side of the cantilever beam. But sorry that I can not measure the deflection directly. However, I bond a accelerator back to the MFC sensor and under the same excitation, the acceleration level could reach 80dB(Reference acceleration=1e-6).
zilanqingyun
 
Posts: 7
Joined: Fri Mar 08, 2019 3:52 am

Re: About the power frequency signal measured by MFC sensor

Postby TPD » Wed Apr 10, 2019 9:35 am

Well,
your setup reads correct. A 2.3mm thick steel beam will properly not bend very much, depending on your excitation actuator. The bending of the beam under the MFC location is the critical part. The bending does strain the MFC which generates the signal. It is not the acceleration at the MFC location. I don't know about your excitation actuator but can you sweep the frequency until you hit the resonance mode of the steel beam? This should generate a sizeable signal. Resonance is were you should see deflection with your eye at the end of the beam, opposite the clamp location.
IF this all does not produce a good signal I would suspect your bonding is not done correctly. Did you cure the 3M adhesive properly as outlined in the datasheet?
TPD
Smart Material Support Team
---------------------------
"That's not right, it's not even wrong."
Wolfgang Pauli, 1945 Nobel Laureate - Physics.
TPD
 
Posts: 98
Joined: Tue Aug 11, 2015 12:46 pm
Location: Florida, USA

Re: About the power frequency signal measured by MFC sensor

Postby zilanqingyun » Thu Apr 11, 2019 8:08 am

Thanks for your kind reply TPD!

I made some progress today.

1--The power frequency signal(50Hz,100Hz,150Hz...) could be eliminated by attaching a piece of silver paper on the welding spots. Also the silver paper is connected to the ground for shielding. Carefully measures stated above are needed for the welding spots are exposed in the air.

However, although the power frequency signal is removed, I find that the background signal is not perfect for vibration measurement. Some frequency peak value exists before excitation.
When I gave the white Gaussian noise to the structure, resonance frequencies can be obtained. I found that some peak values in the background signal are same with the resonance frequencies. Is any method available to get stationary background signal?

2---I do not know how to set the sensitivity in the charge amplifier if I want to measure the acceleration of a structure.
zilanqingyun
 
Posts: 7
Joined: Fri Mar 08, 2019 3:52 am

Re: About the power frequency signal measured by MFC sensor

Postby TPD » Thu Apr 11, 2019 8:53 am

1) How about you do a FFT on the the signal or Auto Correlation to filter the noise? Good oscilloscope have these software function as addons.

2) If you only want to determine acceleration, why don't you use a relatively inexpensive accelerometer? The MFC does only measure strain and you have to derive acceleration by calculating.

3) Also try to use the M2814P2 for this application. The lower impedance should give a better SNR
TPD
Smart Material Support Team
---------------------------
"That's not right, it's not even wrong."
Wolfgang Pauli, 1945 Nobel Laureate - Physics.
TPD
 
Posts: 98
Joined: Tue Aug 11, 2015 12:46 pm
Location: Florida, USA


Return to How-To's

cron

User Control Panel

Login



www.macrofibercomposite.com

Who is online

In total there are 2 users online :: 0 registered, 0 hidden and 2 guests (based on users active over the past 30 minutes)
Most users ever online was 248 on Fri Nov 30, 2018 6:46 am

Users browsing this forum: No registered users and 2 guests

Birthdays

No birthdays today