Piezoelectric and normal force sensor 1

[lekours5]

Hi,

I would like to know if someone has references or ideas about the use of a piezoelectric force sensor in parallel or in series with a normal force sensor.

The piezoelectric force sensor can sense dynamic forces but loses its charge. On the other hand the normal force sensor can't measure small forces but keeps the static force.

My idea would be to sense very small dynamic forces while holding a large static force.

Do you think it would be possible and does someones has references on this ?

Thank you very much !

Alex

 

[BobM3]

How long do you need to "hold" the charge? As I recall, piezo electric conditioners can have time constants over 10 seconds.

 

[lekours5]

In fact there would be a static charge that will be constant. Let's say 1000N. The dynamic force amplitude would be maybe 5N. It is why I can't use a normal force sensor. The piezo force sensor would loose the static 1000N charge and could then sense the small forces. The problem is that the dynamic force can be constant for about a minute. I know this is too much time for the piezo.

What I was thinking is by having a model of the lost of charge in the piezo and combining with the normal force sensor, I guess I could know the actual small forces.

I am then trying to see if it is possible and if there is literature on this.

Thank you very much!

Alex

 

[BobM3]

I don't think you would mount them side by side so they share the static and dynamic loads. The load sharing would be sensitive to tolerances and stiffness and temperature changes. I'd think you would want the two inline so the same force is passing through both. Then the challenge is to sense the small force with the large force present (for the piezoelectric transducer) which I believe is something the piezoelectrics are good at. Then you have two output electrical signals you'd probably want to combine somehow.

I'd talk to the people at Kistler (piezoelectric) and Interface (strain gage) and see what they say.

With a good strain gage load cell, conditioner and cabling, I've seen noise levels on the scope less than 10 mVpp for a full scale output of 10 Vpp. So with a good strain gage setup you should be able to cover nearly a 1000:1 range.

 

[rb1957]

this is a continuation of a previous thread, not that that's a bad thing. but i thought it was mentioned before that the signal is very noisy in the first place (i think that means that the static load varies significantly compared with the dynamic load that's trying to be measured). if the static load fluctuates by more that the dynamic force i don't know how you can detect the dynamic force.

 

[lekours5]

Hi,

I haven't seen any posts on that subject before...I am new but I searched the database without any results...

The static load does not vary...it is static. The signal from a normal strain gage would be be noisy in comparison of the dynamic force because for a 1500N strain gage sensor there would be a lot of noise (in the range of the dynamic force). (For example the 1500N force is a mass and while moving there is vibrations). It is then hard to see if a 1N dynamic force is real or vibrations or sensor noise.

I will look for Kistler and Interface and give news !

Thank you all !

Alex

 

[lekours5]

Excuse me, later I said that some vibrations could mask the dynamic force. That is not true. If it was, it would not be possible to sense this force. The static force is really static. The problem is that a 1500N straing gage sensor is not sensible enough to sense a 1N force. In general it is very noisy I think.

Alex

 

[rb1957]

ok, but i remmeber this type of problem, with exactly these numbers, though it looks like the thread has been removed ('cause my replies don't show the thread i'm thinking of).

anyways, please clarify "in general it is very noisy". if the static force is very static (ie stable) what's the noise ? does the dynamic force a very short wavelength/period ? does the sensing sytem need to respond very quickly, and measure a small signal very accurately ??

i know you can modify a wheatstone bridge to increase the sensivity of the strain gauge, get a bigger ouput for a samll signal. however i suspect that doing this increases the reaction time of the system, decreases the ability of the mesuring system to detect short wavelength signals.

 

[lekours5]

Hi, thank you for your interest !

The noise does not come from the static force. Normally a strain gage sensor with a 2000N range will by itself have a lot of noise. It is why it is hard to obtain a small force. A lot of filtering implies bigger reaction time.

The dynamic force can vary, it can change very fast (10ms response maybe or more, not sure) but also be static during a minute. The sensing system would have to respond in the order of 10 ms. The small signal accuracy is not that important, 1.2 N instead of 1N is not dramatic...


I will look for the Wheatstone bridge, that looks promising. Do you think the effect of a low-pass filter would be the same ??

For the previous thread, maybe someone with who I work posted it...Do you know why it was deleted ?

Thank you very much !

Alex

 

[rb1957]

how will a sensor have/make/contribute to noise ? the sensor may not be very accurate, or not detect a fast changing signal, but to me niether of these are noise.

so the dynamic signal can be very brief, or quite long in duration.

are you trying to detect the dynamic load, by measuring it's effect on some structure ? that looks to me doomed, depending on the structure's response rate.

talk to the strain guage folks ...

 

[lekours5]

Hi,

There is noise in a sensor. Even with no forces on it, there is noise for a hundred reasons, mainly in the cables going to the computer, small vibrations, etc...For large scale sensors, this noise is more important. No I am not trying to measure the dynamic load with it's effect on the structure. It could be though. Because the force I want to measure can be measured with a normal low range force sensor. The problem come from the static force and all the problems coming with it.

 

[BobM3]

Yes, this particular case was in another post a week or so ago. But before the OP wanted to sense 5N with a static force of 1500N.

Do you really mean to group "small vibrations" and noise together? I think your concern is electrically induced noise in the wires, right?

 

[lekours5]

To be honest, I know that it is difficult to sense a small force over a great force but i am not 100$ sure why. Forget about the vibrations, and consider only electrically induced noise in the wires.

A -200 to 200N range sensor provide a signal between -5V and 5V. With a noise of for example 0.01V after filtering, that makes a 0.4N noise.

A -2000 to 2000N range sensor provide the same signal -5V to 5V signal with the same 0.01V noise. But now it translates in a 4N noise, after decent filtering. This is why we can't get the small 1N force out of a big sensor.

I think that's the explanation. Now the idea was to use a piezo and a strain gage together and to mix their properties. I will look with the companies and if you have more informations, thank you to share !

Thank you all !

Alex

 

[rb1957]

but isn't you shield EMI out of the system ? shielded wires ?

 

[lekours5]

Yes the cables can be shielded but it can be very expensive and in an industrial application, there can be other noises. Also, shielded wires are not perfect and there will always have noises.

 

[Compositepro]

One way to measure small changes in a large force are to use force balance principles. An example would be a mechanical lever balance with one ton on each side of the fulcrum. You can use a one pound load cell to measure the force imbalance. There are analytical balances that use this principle to get very sensitive weight measurements. Counterweights are added
until very close to balance and then a very sensitive force detector gets the final weight. An additional benefit of counterweights is that vibrational noise can also be canceled out to some extent.

 

[lekours5]

Thank you for your answer,

I have been thinking of balancing the weight but the problem is that with a counterweight it would be very very cumbersome and I would have to move this extra weight! Also, the static force can change (I would know this change and I don't want to measure the force while this change).

I also have been thinking about balancing with spring but there is too much vibrations with these.

Thank you !

Alex

 

[BobM3]

I don't think you have any choice but to use shielded cables between the transducer and the amplifier. The output of bridges is usually no more that 3.0 mV per volt of excitation at full load.

Out of curiosity I just looked at Detecto's web site and some of their digital scales. They claim to have a 2000:1 ratio (20 lbs FS, .01 lbs resolution) on some of their scales.

You might want to ask the electrical guys (Circuit Design Forum) here what they think is possible.

Let us know what the Kistler and Interface guys (or whomever) say about your application.

 

[lekours5]

Hi,

Yes I will use shielded cables but there will still be noise. What I meant is that extra shielded cables would be expensive. I may have 3 meters cables so there is a lot of noise accumulations. I looked at Detecto but they do balances where they have a lot of spaces in the balances for signal conditioning but more importantly they do not have long cables...the display is at the balance. There is nothing to do with the noise except normal filtering. It is why I was looking at other solutions as a piezo and strain combine or balancing solutions as Compositepro solutions.

Still no answer from the companies, I'll let you know.

Thank you all for your answers !

Alex