I've included a photo of a force hammer (Kistler 9722A2000) that I have at my disposal for use in impact testing. I have inquired on this forum before about it's use and have gotten some really good information. I have now reached a point where I can set up the analyzer for performing an impact test and as far as I know, I have all the settings right. The only question I have now is which end is the sensor and which end do I impact with? Is it the left end or right end in the photo? It looks like the end on the left is the sensor, but with having no experience with this hammer, I have no way of knowing. I don't want to damage it, so I figured I ask you guys and see. I have gone to the Kistler website and have verified the different frequencies that I can excite with each tip, but that is all the info that I have.

BWZE,
Looks to me like the left side has a threaded hole which would be where you screw in your different tips and would be the impact side. I'm not familiar with this type hammer but I would think the hexagonal shaped part (left side) is your load sensor.Does it have a BNC connector coming from the bottom of the handle. I use one from one of the 3 letter anacronym companies that has a double head. One side has a load sensor and the other doesn't. The BNC comes out the bottom of the handle for connection to the data collector for force in measurement.

Sorry, I should've stated that both ends are threaded with the same size threads.

That probably would have been a good piece of info to include in my original post. It would have explained why I didn't know which end was which.

On the left in the picture. The end you hit with is the one with the force transducer on it.

If both ends are threaded, then I wouldn't think it mattered which end you used as long as you had one of the tips on it. You would use the load sensor end if you wanted to see what kind of force you are introducing to the system and use the non-sensored end for "plain" bump test.

That makes good sense. Why didn't I think of that???

Bwze,

I have a similar hammer. Screw the aluminum thread adapter into the sensor threads. Screw the colored impact tips into the aluminum adapter. Use soft tip (Grey) for low frequencies (say below 200 Hz) and harder tips for higher frequencies. Hit the machine/structure with face of impact tip. You can plot Force Vs Frequency to see how force drops off depending on tip hardness, or look at Coherence to be sure it is high at the frequencies of interest.
The steel cylinder is a mass (weight) extender that srews into the other side of the hammer head. Do not hit with this side of hammer head!! Use the mass extender to increase hammer force. I suggest you tape the cable to the hammer handle as a strain relief, since cable and connector can take a beating (or get loose) and give bad data.

Free Tip: I bought replacement colored tips from McMaster-Carr that are much cheaper than from the vibration company.

Contact me if interested in onsite training.
Walt
w_f_strong [at] msn [dot] com

Thanks Walt.
I was wondering about the mass extender. I figured that it was a way to increase the force imparted to a structure while performing the impact test, but wasn't sure. I'll definitely tape the cable up too the save on the possibility of errors.

I am enjoying all the answers. Just adding that the hammers function must be understood. Why is there a sensor in the head and what is its function? This can maybe bring some understanding to that:

I have this hammer above and one bigger using the same soft tips. Since hammers are really expensive, I have also taken a small 3 lb sledge and placed a thread for a rubber door stop (various lengths to vary the softness and hence the freq. range) and an accelerometer at the back side. It does the same. How???
Just reconsider what you need from an impact test: One signal that is proportional to the force, right? And you know force is equal to mass x acceleration. So you can "calibrate" this poor mans modal hammer so you know from its weight and the acceleration of that mass what force you impacted with. I have made numerous comparisons and use the cheap hammer a lot since it is no problem to drop it to the floor (accel has a protection hub).
The force sensor in that modal hammer is just an accelerometer internal crystal where the crystal has a glued plate and thread on each side. So when you hit the object with the proper modal hammer, that force sensing crystal is hit by the intertia of the hammer head. Connect it live on the vibration channel and tap on it with your finger tip and you can see what I mean.

What is the purpose with a modal hammer as compared to a 4x4 whacker or a foot or a fist as impacter? The scientific touch of your report, you can charge much better! If you are just out to modify a specific structure or machine and have no goal in a systematic redesigning like many have in developing new product lines, there is a lot of efficiency in a simple foot or 4x4 impact. As soon as you want to get systematic and want to make comparisons, it becomes valuable to be able judge tests made by different people and in different situations. Taking a measure of the actual force and dividing with it means that you know the structures response to unit of force, not just a vague estimation like your strong fist or Jim´s big foot.

I like that you ask all these questions, since even if they are elementary and basic and some readers can feel smart commenting about people not reading manuals, I am just the same. First making all the mistakes, then ask around, then read the manual, then maybe understand some more but not always. However at least confused on a higher level than before. Keep on asking! All who are here and are responding are your friends and all these questions is what you have friends for!

Arne's comments on users not reading manuals is correct. My company makes software and hardware that come with manuals and we find it's unusual to find a customer who actually reads the manuals. What they typically do is what Arne describes.

This message has been edited. Last edited by: Duncan Carter,

As I p.s. to my previous post, I'd like to pass along a comment by the former sales manager of Data Physics who told me that he considered it a failure on his company's part if a customer needed to use a manual to operate his company's products.

I have no problem with reading manuals whatsoever. The only problem in this particular instance is that I don't have one. I too agree that the best course of action is to read first, search second, and when all else fails ask last.
On a note about the manual for this hammer, it would seem logical to me for a company in this day and age to have manuals readily accessible to customers in some sort of electronic format on their websites. In this situation however, it is not so. Seems as if Kistler either doesn't have one available in the previosly described format or maybe you just have to call and get one ordered.
In any case, I truly do appreciate all of the responses to this post. It's nice to have a place to bounce ideas and get help from people who have a lot experience in this field.

I might be wrong, but the gut feeling from reading many papers and books on modal analysis and listening to many doing it for years, there is a slight gap beween modal fellas and the 4x4 ones. A modal person has an urge to disregard a single channel bump result such as 16 Hz as unworthy since it must be followed by a statement of 0.99 coherence and you need two channels to calculate it. The modal guys report often has a result 16.06789 Hz and the soft has calculated that the mod suggested will change the resonance to 23.78894 Hz. Sad thing is that the 1000 RPM motor (in this case, I read from an actual report) is only concerned with the sad fact that the resonance at 16 Hz was just moved 0.6 Hz up when that nice mod was made in reality. Why was the error so large when the 5-digits approach seemed so reassuring? The mod itself was surely fine, but the structural facts of the real machine in that model was just too rough. Barely half a digit accurate. With a few tests on site, without any calculations, the foot as exciter (tall concrete pillar) the new mod was moving the resonance 8 Hz.

I do not want to discourage anyone to learn modal analysis, since it is valuable in the right situation. But as with many trades, much more important is to get routine from real world work. Modal work is just a tool, not something to do as a mainline without getting dirty hands. I hope these points of views are not valid in the US, just over here.

The reason your modal hammer did not come with a manual can be the same as the reason why a carpenters hammer also do not have one. If you get a hammer for nails you are supposed to have aquired needed knowledge from your background may it be father, mother or apprentice training, whatever. Same here I think.

Just to agree / expand upon Arne's points made above;
Yes it seems that full modal analysis is often applied when not necessary. A simple bump test to get the frequency, and a plot of amplitude versus position to get the shape will do it (if your plot of amplitude versus position shows a decrease to very small magnitude then an increase, you don't need phase to figure out that's a node.
About theoretical model versus the real structure: many years ago did some lab work on tubular beams. One detail was to build a "clamped" end condition. Took three tries to get close to that behaviour!! - first cuts were not sufficiently clamped. And that's in a lab!!
Further on overkill: it's now happening with Computational Fluid Dynamics. Have seen cases where a flow distribution problem / solution has gotten the full treatment, self-pat-on-the back, and a paper published. Knowledgeable, experienced fluids person could have done it all in half an hour with sketches on the back of an envelope.

Possibly, I have misunderstood the use of this hammer. I am well aware of it's uses in the world of modal analysis, but I was under the impression that it could be used to correctly verify a natural frequency of a structure to a higher degree of accuracy. I just thought that with a force hammer and a two channel analyzer that you could more or less definitely determine natural frequencies better than you could with a straight forward bump test.

BWZE - you have not misunderstood. A tuned force hammer and a response channel (accel) connected to a 2 channel analyzer will give you very accurate natural frequency data. The added benefit (as Arne pointed out) is that you can confirm your natural frequency by using coherence or corresponding phase data. (there will be a phase shift at resonance which you can see on your data collector)

Can anyone solve this mystery to me: How can the spectrum be evaluated to show a resonance with better accuracy just because you have hit with a modal hammer rather than a fist or a foot or a 4x4 inch wood piece?

I think that is a myth supported by the modal establishment. The poor accelerometer that receives the bump vibration can not possibly know how to react differently from a modal hammer or foot impact.

Yes, the forces spectrum is important. you have to cover the range of interest. Different heads of the modal hammer (or various length of the dooor stop) and different feet, fists with different gloves and other ways to change the force spectrum are important matters to keep track of. But you can do that with the single channel response accel placing it on the "hammer", at least to get an idea. You should train on "known" machines a lot to have a routine when hitting the problem machines. Figure out the differnce between a hard head and a soft head on a known machine. You will soon find out.

Normalizing to get a vibration per unit force using the second force channel in the hammer is just a way to get several tests to be comparable from level point of view, not frequency.

Modal software have curve fitting algorithms to figure out resonance peaks that are tricky to separate when they are near to each other. That can improve accurcay at a certain congested spectrum out of many on the same machine. But a resonance is visible on almost all spectra, just different levels. So also here I cannot see how a resonance is determined more accurate. This improvement is not a part of the CSI adv 2-ch software.

There are a few pending traps in bump testing that however are important. One is that a machine normally stands still. So bearing oil films and loads in teeth, etc are not established. Therefore resonsnce in the rotors can be a lot off from real running world. Forces during operation in an aggregate can stiffen or loosen joints. This can tune resonances much much more than the 4th decimal claimed from the modal soft results.

In general, let me make the claim that very rarely I read a report where the analyst has taken on the job to make a realistical error analysis of her/his results. Maybe you always do that? Then be proud, you are one of a rare species.

Let me try your patience a bit more :-) What is a tuned hammer? Tuning means shifting a frequency by means of a change in mass or stiffness. If you mean the change of the force spectrum by a mod of the head stiffness any hammer using a simple rubber block in various thickness willl also establish a tuned poor mans bump hammer.

The modal hammer has the advantage that the handle is well damped so the hitting does not excite its own resonances to add to the froce sensor. Using a hickory or a tube with a polyurethane cover handle. My 3 lb sledge has a hickory handle! Text says Made in Taiwan and it was 4.50 USD. Must be almost the freight cost around the world? Door stops are 2:50 each. A quality sledge by the way, I had to drill deep into the head to reach steel without hardening to tap the 5/16" hole.

Sure Arne, you don't have to spend big bucks to get the gear you need. We have used a 40 gallon drum filled with concrete to shake a paper machine. Whether or not the data is good? Well, there are so many variables. But we do need a starting point. So, if your spectrum shows a possible resonance, the 2 channel will give you a phase relationship whereby you can be a little more certain that it is resonance. A good program will have a force channel set up where you "tune" the hammer by practicing hitting the machine until you feel you've got the right force and location to get a good response. I'm sure you can do the same just watching the live spectrum and hitting the thing with a 4x4. Spruce for low freq and Maple for high end stuff OK?

quote:

How can the spectrum be evaluated to show a resonance with better accuracy just because you have hit with a modal hammer rather than a fist or a foot or a 4x4 inch wood piece?

Maybe I've misunderstood all these years, but my thinking is the purpose of a force/impact/modal hammer is NOT to impart any special excitation to a machine or structure, but rather to accurately measure the FORCE that is imparted, so that when the resulting RESPONSE (acceleration) is also measured, you can determine how RESPONSIVE the machine is to input.

In other words, using a simple "bump test" with a 2x4 do determine that an 1800 rpm fan has a resonance at 1870 cpm is not enough information. If it takes 10,000 lbs. of force to generate a significant response, then the resonance is probably not a problem. But if it takes only 100 lbs. of force, then it most certainly will be a problem.

Identifying the mere existence of a resonance is only slightly helpful. You need to be able to characterize the force/response functions. For that you need a 2-channel analyzer with appropriate software.

This message has been edited. Last edited by: rustythevibeguy,