Hi

Just to make sure - I hope you are computing Frequency Response Functions (aka Transfer Functions) & Coherence, not simply dividing the response and force spectra with each other?

https://en.wikipedia.org/wiki/Transfer_function

http://blog.prosig.com/2009/10/19/a-simple-frequency-response-function/

They should be the same, but the methods differ in the kind of systematic measurement errors they introduce. They can be quite close though. 

The hammer tip has a mass which is accelerated and influence the magnitude/phase of the force signal. Also, the hammer picks up force before and after the impact & you can get double impacts. A hammer blow never is exactly the same twice, as it varies in strength, orientation and location when landing on the object. Therefore, the rule of thumb is that it does not pay to average more than about 9 hammer impacts.

A force sensor excited by a shaker also has a mass that is accelerated, but this acceleratoin can be measured and compensated for. The shaker places an inertial load on the object, and if it is a thin shelled object, rotational intertia loading can affect results. Depending on the mechanical setup, non-driving dofs of the shaker may add damping to the object you are studiying. 

Comparing FRFs obtained with the respective methods, expect the largerst differences to arise at anti-resonance, as this is where phase errors matter the most. 

For a locally reacting material, i.e. for a material where the response close to the load greatly varies (think of a soft sponge and your loading it with your finger to get an idea of what it looks like when locally reacting), there may be large differences between the methods, partially as the excitation surface (size, shape and loading plate stiffness) affects the local FRF magnitude. Concrete is a locally reacting material. 

Then, of course, you have the question of what happens when/if you move acceleroemeters across the measurement object during acquisition of your FRFs ,

You might want to take a look here as well

http://qringtech.com/learnmore/recommended-reading/

http://qringtech.com/learnmore/why-simulate-measure-correlate-automate/ 

Hope this helps

Claes

Arun,

The hammer test, is supposed to excite all the modes because the impulse consists of all the frequencies.At the same time it is assumed that the shaker test is sweeping the bandwidth of the system and will excite all the modes or give excitation at all the frequency range from zero up to the Nyquist frequency.

The difference between the two methods, in theory, is that the first excites all the frequencies at once, while the second excites one frequency at each time instant. This does not affect the result in theory. But in practice you can not do a 100% accurate sweep with a shaker and you will jump over frequencies that could be significant. Therefore the hammer test will give slightly different results than the frequency sweep test. If your hand is trained to do the hammer test, it should give a very good and reliable FRF.

Best 

Hi Tariq

A shaker can emit transients, chirps, random- or periodic- noise in addition to the sine sweeps you discuss. 

/Claes

Thank you Tariq and Claes for your answers. 

So, just to summarize my understanding of your views:

Theoretically, both the results should be the same, assuming that the noise, inertial masses of the hammer and shaker force sensor are ignored and we assume an infinitely high frequency resolution during the shaker sweep.

Hi Claes Richard Fredö ,

If you don't mind me asking, you mentioned

"The hammer tip has a mass which is accelerated and influence the magnitude/phase of the force signal. Also, the hammer picks up force before and after the impact & you can get double impacts. A hammer blow never is exactly the same twice, as it varies in strength, orientation and location when landing on the object. Therefore, the rule of thumb is that it does not pay to average more than about 9 hammer impacts. "

In this instance am I right in assuming that if you have an automated impact device which can control the force/speed of the impact (highly repeatable) and orientation and location can be obtained from sample testing typical specimens, one could theoretically carry out further tests on similar specimens without the need for a load cell/force sensor? Even the double impact you mentioned can be compensated for by knowing the input force signal characteristics by carrying out tests on similar specimens, I presume.

So basically if you one could characterise the typical input force signal then it could be used (albeit being an estimate) instead of actually measuring it every time to get the transfer function and subsequent FRFs. Is my thinking right?

Thanks in advance,

best regards,

Ravi

Hi Ravi

The rule of thumb applies for free hand impacting. Reproducability can be increased, e.g. by gluing on an easy to hit contact area or as you write, by automation.

FRFs with double hits become strange. The usual trick is to abort the last impact and simply make a new one.

For impact testing, force usually is measured as it tends to vary, i.e. as the reference is not stable. If your reference is very stable, things become easier.

Then, of course, you have neat signal analysis tricks. R.B. Randall showed me that he, using Cepstrum analysis, was able to get FRFs from response only measurements.

Hope this helps.

/C

Hi Claes,

Thank you so much! Your answers were clear, concise and absolutely addressed all my queries. That is absolutely perfect.

Thank you!

best regards,

Ravi

Hi Ravi

Thx. Have fun, because, this is fun. (Swat it & watch it wiggle)

You might want to check out this stuff. https://github.com/openmodal

/C

The hammer excites all modes of vibration. The energy of the coup is distributed along a large spectrum.The natural frequencies have very different intensity of response of several magnitude orders weaker or stronger. So the weaker responses can not be detected by the hammer method because of low intensity. If you use the sweeping frequencies method, you can stop the scanning and you can also increase the excitation to a better discrimination of weak responses.

That will depend on the mass of the shaker head In comparison to the object you are exciting. An impact hammer test is pretty much the contact less response, given that the acclerometer is not loading the structure too much either.

There will be a component of increasing mass load for a shaker with frequency.

And the driver point is not the real contact point between the shaker and the object, but inside the coil of the shaker. The mass of the shaker pin or plate will be part of the excited system. A shaker usually should be able to give more power, but that has a cost in the response curve.

E.g. many measurements done on stringed musical instruments at PTB Braunschweig under J. Meyer was done using a shaker. The response curves are clearly influenced by the shaker pin mass moving the resonance frequencies down and showing a characteristic dropping trend with frequency. The peak to valley distance also seem to drop. There may be some added damping too. But violins and guitars are very light structures, especially the bridges.

There is another effect that is usually forgotten. The non linearity of the system that is under investigation. The FFT will give you a well defined curve around the resonances in any case but the sweep method will show you the instability region with different peak responses if the frequency is sweeped downwards or upwards. The FFT does not catch this effect.

Theoretically both the results should be same but in reality output of both (impact hammer, shaker ) the methods may not be same. Though hammer test will produce all frequencies at once but accuracy of this method is rely on accuracy of impact and in case of shaker it may not sweep all range of frequencies.

Kali-Naresh

That is indeed very theoretically, because

- when doing impact testing, you do not excite as strongly from one blow to the next, ie force amplitude varies and impact hammer tips are not totally linear.

- you never hit the same spot, at the same angle. The rule of thumb is that coherence tends to drop after 8-9 impacts. You can get that number up with a few tricks, though, not by orders of magnitude.

Wrt to all frequencies at once - when scanning for resonance and resonance only, bump (aka ping) testing is faster and more reliable than FRF measurement. You then only record response and bump the system any which way and direction you can.

That said - a shaker does tends not to get that far away from a lab setting. I do most practical work with impact testing and bumps.

Hello Arun,

I would like to add to all the previous comments that you might have a look to the auto power spectral densities of each excitations (hammer vs shaker) to understand if they suite your needs. They might be quite different, with the shaker having much more freedom in shaping the excitation spectra, beyond repeatability of the tests. This might be very relevant when searching for weakly excitable eigenmodes.

Also, please do not forget that the structure responses change when the excitation location varies, therefore any hammer hit is slightly different in the point of energy injection.

You can have some evidence in my extensive works about full field FRFs and derivative quantities in the TEFFMA project:

Article Broad frequency band full field measurements for advanced ap...

Article Competing optical instruments for the estimation of Full Field FRFs

Article Full field optical measurements in experimental modal analys...

Article Chasing the high-resolution mapping of rotational and strain...