Measuring with multiple hammers

One way to measure the dynamic characteristics (FRFs) of a component or assembly consists of exciting it with an impact hammer and measuring the responses with sensors. When hitting with an impact hammer, a long pulse in the time domain results in a short or narrow frequency spectrum, while a short pulse results in a wide frequency spectrum. The pulse length depends on the material of the hammer tip and the mass of the impact hammer.

Selecting the right hammers

When deciding which hammer tip to use for your measurement campaign, you must consider whether it will input enough energy into the system to measure the entire frequency range you are interested in. Sometimes, it is necessary to use multiple hammers with different tips and filter their signal. For example, you could use a soft tip to excite the low-frequency range and a harder tip for the higher one. DIRAC will then merge the signal of the hammers.

In this article, you will find a workflow that helps you understand whether you need one or more hammers for your measurement and how to merge their signals properly to get high-quality test-based models.

Whenever you make an impact with the hammer, it excites a wide frequency range. You want to hit strong enough to put a lot of energy into the system but still ensure that your impacts are repeatable and do not generate overloads. For a good measurement, you want the input spectrum curve to be “flat” and without significant drops or zeros. However, the signal will decay at a specific frequency and eventually reach 0. When this happens, the energy generated with that hammer is insufficient, so the structure is not excited at those higher frequencies. You also want to be sure that the input spectrum does not exceed the frequency range you are interested in. When this happens, it might cause sensor overloads. For this reason, you need to check whether the hammer and tip combination you have chosen can excite the entire frequency range of interest, both at high and low frequencies, with enough energy but not too much.

Checking the maximum frequency you can measure with a hammer means checking the quality of the signal-to-noise ratio over the frequency range. You can study this in the Measure module of DIRAC.

Step 1: Add the hammer and make some impacts

The first step to understanding whether you are using the correct hammer tip consists of making some impacts with the chosen hammer. We generally recommend hitting hard enough to put enough energy into the system. However, if you put excessive strength, you could generate overloads and your repeatability will decrease. To start checking if you have chosen the correct hammer, you have to:

1. Add the hammer to the DIRAC project.
2. Connect to the DAQ.
3. Choose one excitation point.
4. Make a couple of impacts at that point.

In the Graphing card, the top two plots show the input signal in both the time and frequency domains. To understand up to which frequency the hammer inputs enough energy, you have to look at the force over the frequency graph. The area below the curve corresponds to the energy added in the system.

Step 2: Check the maximum frequency excited

You must now understand whether there is good signal-to-noise at the maximum frequency range you are interested in. To do so, check the coherence graph of the FRFs measured at the sensor farthest from the impact location. You can select this sensor in the 3D viewer or the drop-down on top of the accelerance graph. In the coherence plot, find the frequency where the coherence starts decreasing. Where the coherence drops below 90%, your signal-to-noise is not excellent, so that is the maximum frequency the hammer can excite properly.

Checking the coherence is the best method to identify the maximum frequency where you have good signal-to-noise. If you struggle, you can use the following rule of thumb which is less accurate and more conservative.

1. Check the maximum force value in the input frequency graph (top right in the Graphing of Measure). You can read it by hovering with your mouse on the curve.
2. Divide it by ten and find the point on the curve with this force value.
3. From that intersection point, draw a vertical line to the frequency axis.
4. Check the frequency value.

This value is the frequency up to which that hammer inputs enough energy in the system that results in good signal-to-noise. If this value is higher than the frequency range you are interested in, you can proceed to step 3. If not, you should replace this hammer with a harder tip and return to step 1.

Step 3: Checking overloads

It might seem that adding a very hard tip (e.g., steel) that excites the system up to very high frequencies is a quick and intelligent solution. That is not always the case. You want your hammer to excite just above the maximum frequency you are interested in and not much more. When you use a too-hard tip, you will likely introduce overloads, especially to the sensors located close to an impact point. For this reason, you must check whether the sensors are overloaded. You can do so by following the instructions in this article.

If some sensors are overloaded, the chosen hammer tip is too hard, so you must switch to a softer one and go back to step 1. If that is not the case, the hammer you have chosen is fine, so you can check how it performs at the lower frequencies.

Step 4: Checking the low frequency range

After finding the highest frequency your hammer can measure, check whether that hammer excites the system at low frequencies with enough energy. You do this by checking the coherence at low frequencies (indicatively between 0 and 200 Hz) of all sensor channels. Go to the Analyze module to check it for all sensors simultaneously.

1. Select the Measured preset.
2. Set the mode to Coherence.
3. Select a low-frequency range (e.g., 0 – 200 Hz).
4. Select each sensor channel and check the coherence plot.
5. Check the average coherence by hovering on the matrix cell with the mouse.

If you see low coherence (below 90%), your hammer does not input enough energy in the system. Add an additional hammer with a softer tip to the project, do some hits and repeat step 4. If the coherence is high enough, you can proceed.

If you use a single hammer, you have done all the checks and you can start testing. If you need multiple hammers, you have to filter their signals so that each of them only excites part of the frequency range. Go to step 5 to learn how to do that.

Merging the hammers signal

You just realized that you need multiple hammers for your measurement campaign. You can add the additional hammer(s) in the Prepare module, accessing the sensors database, or directly in the DAQ window, by clicking on the + symbol in the 2 Measurement Settings step.

Step 5: Finding the merging area

The input energy is the area below the curve in the force-frequency plot. Looking at the picture you can see that at low frequencies the rubber hammer inputs more energy than the nylon one. At higher frequencies, the nylon hammer is better. You want to use the signal from a hammer at the frequency where it inputs the most energy and then switch from one to another at the intersection. The intersection corresponds to the cut-off frequency. Since you do not want an abrupt jump in the results, you need to merge the signal of the two hammers. The frequency range in which the two hammers merge is called the roll-off frequency.

You now need to filter and merge the signal of the hammers, so that each of them is used for the desired frequency range. This is done in the Step 2 Measurement Settings of the DAQ module.  When applying the filtering, DIRAC will filter the signal of the two hammers according to the settings and merge the resulting FRFs.

1. Add a low pass filter to the hammer exciting low frequencies.
2. Add a high pass filter to the hammer exciting high frequencies.
3. Set the cut-off frequency (frequency where the signal is 50% filtered) at the frequency where the two hammers intersect.
4. Set the desired roll-off frequency (the frequency range where the filtering will take place). We recommend setting 100 Hz at least.

Measuring with multiple hammers

Connecting

After setting up the hammers, you need to connect both of them to the hardware, at two different locations. If you only have one hammer (in which you will change tips), you still need to use two different connectors, plugging the hammer first into one and later into the other. You also need to map the hammer channels correctly in DIRAC. You can do this in the DAQ at step 3 Map Sensors. Here, you drag and drop each hammer to the correct connector location on the hardware.

Performing the measurement

When all is set up, you can start measuring. For each impact point, perform some impacts with each hammer. DIRAC automatically detects which hammer was used for the measurement. If you only have one hammer (in which you will change the tip), you first do all impacts with it. Then you change tip and connect it to the next connector, before performing all impacts with the new tip.

The measurements of multiple hammers are merged automatically whenever there is data of multiple hammers. If for an excitation there are only measurements with one hammer, it is used for all frequencies, independent of the filter settings of the hammer.

Checking the quality of the hammer merge

It is very important that the merge of the hammer signals is done accurately. You have to check the quality of the hammer merge at the beginning of your measuring campaign, after performing a couple of impacts in two or three impact locations. To know how to do this, read the dedicated article here.