Home » Why Engineers Use $1000 Hammers To Hit Prototype Cars So They Sound Good And Don’t Explode

Why Engineers Use $1000 Hammers To Hit Prototype Cars So They Sound Good And Don’t Explode

Nvh Ts
ADVERTISEMENT

As a noise, vibration, harshness (NVH) engineer, I get to play with some neat tools. From lab-grade microphones and accelerometers, anechoic chambers, acoustic cameras, to high tech data acquisition systems. This equipment is important, as cars and trucks are complex machines in a market that is more competitive than ever. Despite thousands of explosions occurring in the engine every minute, whirring harmonics from gears and bearings, the road battering the suspension, and wind rushing across and through the body, your vehicle’s interior must be a pleasant place to be. Manufacturers aren’t just looking for quietness; the sound quality you experience from when you slam the door to when you rev the engine must match the character of the brand and vehicle you are driving.

The tool-of-the-trade that surprises most people is the modal hammer. This isn’t something you can buy at Home Depot, it’s a precision-tuned instrument that costs thousands of dollars. Every vehicle, from the complete system down to the component level has been hammer tested throughout its development. The data gathered while tapping away with these hammers not only ensures that your vehicle sounds and performs as it should, it also confirms that it won’t fail in a catastrophic explosion. Here’s how.

Vidframe Min Top
Vidframe Min Bottom

[Editor’s Note: Everyone please welcome Steve Balistreri, a fellow enginerd and Detroiter. He’s an expert on NVH, and since I’m fascinated by the topic, you can expect to see lots more from him soon! -DT]. 

Screenshot 2023 09 08 At 9.22.22 Am Copy
Bruel & Kjaer

A huge aspect of nailing the NVH performance of a vehicle is managing resonances. An example you are familiar with is how your vehicle’s mufflers are tuned to cancel out specific frequencies your engine makes. There’s a reason deleting your muffler often creates a loud drone at common cruising speeds; my 2001 S8 with a rear muffler delete is a prime example that gets really loud right at 45 mph. The OEM mufflers were tuned deliberately for this.

Resonance Can Make Cars Ridiculously Loud And Uncomfortable

What do I mean by “resonance”? Well, everything has frequencies that it naturally wants to resonate at. Body panels, subframes, engine blocks, driveshafts, the air cavity in your vehicle’s cabin (the Supra wind buffeting issue is a prime example of this). In the case of musical instruments, they are tuned to resonate at specific frequencies or notes that are pleasing to the ear. The components of a vehicle are specifically tuned to avoid frequencies they are likely to be excited by, such as engine firing, gear harmonics, or forced inputs from your vehicle’s suspension. The reason is best shown in the animation below, plotting vibration amplitude vs frequency.

ADVERTISEMENT

You can see that below the *resonant frequency* the response is pretty flat, meaning the part will vibrate about the same amount as what is put into it. However, as you approach the resonance, amplitude increases exponentially to the point where your part is vibrating at 10x the amplitude as the vibration you are putting in. As you can imagine, this is a bad thing.

University of Utah

A classic example is the opera singer breaking a wine glass with her voice. The physical force from the sound of her voice is miniscule, and the glass is unaffected (unless of course the glass is some sort of opera connoisseur). Once that special frequency is hit, the once-rigid glass starts undulating in impossible ways until it finally explodes. The video below from the excellent Slow Mo Guys channel is a great visual. Trying to deform a wine glass like that with your hands would be impossible, but a speaker at the right frequency can turn that rigid glass into jello.

Screenshot 2023 09 08 At 9.23.56 Am

In a car this can cause a few things to happen. Most often a specific gear will suddenly get loud at a certain engine RPM range as a mount or subframe resonates at that gear’s frequency (the frequency of the powertrain depends on the engine RPM, which is set by the gear when traveling at a steady highway speed), or your car can sound very “boomy” as different road surfaces or engine frequencies excite resonances in your vehicle body.

Resonance Can Literally Cause Things To Explode


Other times this concept of resonance can cause catastrophic failure. A common example is driveshafts. Some vehicles are speed-limited so that a once-per-rotation vibration caused by imbalance doesn’t intersect with the driveshaft’s first resonant frequency. There were several instances of V6 Mustang owners deleting their speed limiter, then getting the rude surprise of their driveshaft exploding as they flew down the drag strip at over 130 mph. There are videos on YouTube of boneheads doing this on public roads. As you can hear, it’s a catastrophic failure (and you can see the aftermath here).

ADVERTISEMENT

Screenshot 2023 09 08 At 9.27.19 Am

OK But How Do You Prevent These Resonance Problems?

The company I work for, HBK, is based in Denmark and produced training materials in the 80s with delightful hand drawn illustrations. Here’s one showing an example of how resonances can amplify frequencies you don’t want. In the bottom left graph (commonly called a waterfall graph) the diagonal “waves” are engine orders (firing frequencies based on the number and layout of the cylinders that increase with RPM). That engine order crosses a 30 Hz resonance at 100 kph, where the acceleration (Z axis) increases dramatically. The top graph shows that engine order sliced out and plotted as acceleration (vibration) vs speed. As you can see the resonances dramatically increase the level of vibration.

Screenshot 2023 09 08 At 10.16.30 Am
Bruel & Kjaer

But how do you calculate resonant frequency? If you went to engineering school prepare to have some flashbacks. Below is a simple mass / spring diagram which represents the simplest, ideal structure. The resonant frequency equation is below and is based on the mass m and spring constant k (or stiffness in a non-spring), and these are what you need to change if your resonant frequency is in the wrong place.

Think of stretching a guitar string to change its pitch. This is a very simplified example, like comparing the driving in Police Quest with Forza. In the real world, parts have internal damping inherent to their structure, the shapes themselves are complex, and the equations get more complicated. Adding damping flattens the peak and lowers the frequency slightly. You can imagine this like a tuning fork which has very low damping; it will ring forever at a very sharp tone if you tap it with a metal rod. While if you tap your desk with a metal rod, it’s a much broader set of tones that go away almost immediately. This is the difference damping makes.

Screenshot 2023 09 08 At 10.18.58 Am
Georgia State University

Let’s Talk Hammers

All this is modeled out in CAE/FEA (Computer Aided Engineering/Finite Element Analysis) software early in the vehicle program. Engineers do what is called modal testing with these hammers on prototype parts to help correlate these models and confirm these frequencies. But why do we need a hammer that costs as much as a used car? Can’t you use a regular hammer, your knuckle, or a petrified French fry you found between the seats, and just record the sound with a microphone? Technically you can, and sometimes for quick and dirty troubleshooting we may just tap around to find a part that rings around the problem frequency, but that only gets you half of the equation.

ADVERTISEMENT

To get the frequency response or sensitivity, you need to know your input and your output. You can tap a driveshaft with your knuckle all day, but how much force are you putting in to get it to ring that amount, and what frequency range are you exciting? These hammers have force transducers built into the business-end that records the force you are inputting into the system to get that input to output ratio. They also come with interchangeable tips with different hardness to excite different frequency ranges. For example, tap your coffee cup with the fleshy part of your finger and again with the tip of your nail. Sounds different right? Soft rubber tips will set off any low frequency modes, while harder nylon and even steel tips put high frequency energy into the system.

Having the force and response gets you a Frequency Response Function like you see below, displaying how much vibration you get per Newton of input force vs frequency. The peaks are the resonant frequencies. You usually test multiple points on the part which is why there are several different traces on the graph.

42452 2019 806 Fig6 Html
Springer Nature

The hammers come in different sizes depending on the part being tested. There are tiny baby hammers smaller than a Bic pen for testing small parts like hard drives, and large sledge hammers for giant structures like bridges or the NASA launch pad I got to do structural testing on. The 8206 hammer is our workhorse (note, images below are not to scale).

Hammers Copy
Bruel & Kjaer

Here is a photo of a modal test being done on a car mirror. You can see the accelerometer is attached to the mirror glass. They are probably confirming that the parts in the mirror won’t be excited by vehicle vibrations, which would cause the mirror to shake excessively, blurring the reflection and causing a safety issue.

Impacttesting Rearviewmirror
M+P International

The hammers are also modally tuned, so the hammer itself doesn’t have resonant frequencies in the range you are testing. You want the part to be as isolated from outside interference as possible, so if the hammer was resonating along with the part, you’d have bad data. This is also why we isolate the parts as much as we can, suspending them with bungee cords if they are small enough.

ADVERTISEMENT

Large parts like vehicle bodies get suspended from the ceiling with rubber straps or are mounted on airbags to isolate them from the ground. It is an odd sight walking into an anechoic chamber and seeing a full pickup body, instrumented with hundreds of accelerometers and strain gauges, hanging from the ceiling with some poor engineer underneath hitting it with a hammer. It’s a tough job, but somebody has to do it. Below is another drawing from our company’s training manual showing a vehicle body being isolated from its environment. In this photo they are using a shaker instead of a hammer but it’s the same idea.

Screenshot 2023 09 08 At 10.39.42 Am
Bruel & Kjaer

So the engineer will map out several points of interest on the part in question to glue accelerometers to, and then do a modal survey (basically tap the part in a few different spots to see what works best). They will then tap that special spot with their hammer, which will trigger the data acquisition software to record the hammer force and the resulting vibration from the accelerometers. Typically, it’s five hits per location to average out any inconsistencies. The hits must be quick and clean — no double rebound hits or any scraping movements, because it messes up getting that clean initial impact. This can be tough when your arm is snaked in an engine bay, hitting some transmission mount you can’t even see. You don’t need a whole lot of force to do this — light taps will do. There is a story of a new guy who was testing a car’s roof and was hitting it hard enough to dent the roof. Not only does that give you bad data, it screws up the hammer. He learned his lesson pretty quick.

An example test is shown in the illustration below. The accelerometers record the resulting vibration from your hammer hit, the data goes into our giant mainframe, tape reels do whatever they do, and it spits out the results. We print the graph out with our extremely loud dot matrix printer, give it to the head engineer and he gets mad.

Screenshot 2023 09 08 At 10.41.01 Am
Bruel & Kjaer

For a full modal test, the measurement point coordinates are imported into our software to create a wireframe model. The test data is then imported and processed, and we get the resonant frequencies as well as the mode shapes. The mode shape is like the shape the wine glass was deforming into during the above video, each frequency has its own shape, and they get more complex as you increase the frequency. An example of a vehicle mode shape is below (the colorful one). This is CAE data; our experimental data has significantly lower resolution as it’s just a wireframe of a few dozen/hundred points, more like the image below that, but you get the idea.

Screenshot 2023 09 08 At 10.42.44 Am
Bruel & Kjaer

 

ADVERTISEMENT
EMFAD

A Look At Data On A Pickup Truck Body

Below is a screenshot of our analysis software doing some curve fitting and extracting mode shapes from a test on a truck body. As you can see, there are several modes between 12 and 36 Hz. The mode shape at 17.6 Hz is shown in the geometry window and appears to be the rear of the truck bed wagging back and forth like a dog’s tail.

Advanced Modal Analysis 1
Bruel & Kjaer

The mode shape in the CAE animation above most likely happens at a very low <40 Hz frequency and would be heard and felt as a booming noise over certain road surfaces or bumps. It may be hard to believe that your heavy steel vehicle body is wobbling like a giant Jello jiggler but that is actually happening. The part needs to vibrate in some way to create the sound waves and vibration that eventually reaches your ears and your butt. The displacement in these animations are greatly exaggerated so you can visualize what the shape is, as it is typically invisible to the human eye, but not always.

We tested a prototype industrial engine which had a fuel line that would resonate around 450 Hz, vibrating so much it would look blurry at high RPMs. We glued an accelerometer to it and measured a whopping 122 g’s of vibration! As a reference the engine block rarely exceeds 2 g’s. At 450 Hz that fuel line is moving almost 300 millimeters per second which is pretty incredible. This would quickly fatigue the line, causing fuel leaks and fires in their engine dynos. We fabricated some brackets that increased the stiffness, shifting the resonance up and away from where it would be excited by the engine, which dramatically reduced the vibration and fixed the problem.

Automotive engineers have gotten a lot better at NVH over the past decade or so. Much of it is due to advances in CAE, and automakers making NVH considerations much earlier in the design process. There was a running joke that NVH stood for “not very helpful.” Early prototype vehicles are rattly, loud pieces of junk, so NVH concerns didn’t make themselves known until engineers got their hands on the later, more refined prototypes when most tooling was finalized. So solutions to NVH problems were either to lessen the source excitation ie. make the combustion process less explodey or the gears less geary (not really feasible); redesign part in question (extremely expensive and time consuming after the part/tooling has been finalized); or treat the path to the driver which involves adding mass, stiffness, or damping treatments (like tuned mass dampers) which all add weight and cost. Not what a manager wants to hear when they are looking for an easy fix. Thankfully these surprises late in vehicle development are becoming more rare as good NVH has become a higher priority.

Screenshot 2023 09 08 At 10.56.01 Am
Bruel & Kjaer

Fascinating Ways To Masking Problems Late In The Vehicle Development Timeline

In the modern days auto companies have gotten creative for masking these problems when they can’t be bothered to fix them. One example is ANC or active noise control, which uses cabin microphones and your stereo speakers to cancel out problem frequencies. These work a lot like your noise canceling headphones but are only effective for low frequency boom sounds in your car. It works quite well and is much cheaper than significantly changing the vehicle body structure.

ADVERTISEMENT

Another more humorous solution is to edit out specific RPMs, literally making them disappear. A common boom detection test is to do a slow (45 second) engine runup from idle to 1500 RPM while the vehicle is parked. This will cause the engine frequencies to excite any problem areas in the idle range. One particular car would not let me get a smooth ramp, the tach needle avoided 1100 RPM like a baby swatting away a spoonful of steamed spinach. I asked an engineer and it turned out the engine vibrations set off a resonance somewhere in the vehicle at this RPM, so they programmed it to simply skip past this engine speed so the resulting boom is barely noticeable. Once again adding a couple lines of code is much cheaper than changing the tooling for a major structural component, so they went for it.

Our future with EV’s presents unique challenges and opportunities. For one, you don’t have an engine, transmission, and driveshaft flooding the vehicle with vibrations and noise. The problem with that is engine noise helps mask a lot of wind, road, and accessory noise so any issues there become more prominent. Also the heavy engine block acts like an isolator for things like your A/C compressor and water pump, which now have to be mounted on an isolated frame of some sort. Electric motors and inverters create less NVH comparatively, but it is at much higher frequencies than an engine which creates its own host of problems. Some automakers are experimenting with different materials for things like motor mounts that are less sensitive to high frequency vibrations.

There is also the problem of a lack of vehicle character, which relied heavily on the noise and vibration of the powertrain. Besides certain restrictions from federal pedestrian safety noise requirements, automakers have a blank slate to create an artificial noise profile that imparts a character to the vehicle. This takes a lot more work, getting into psychoacoustics and sound quality and perception for different customer groups, in my humble opinion we have a way to go with that. We will save that for a future article.

So next time you go for a drive, enjoy the solid thunk of your car door closing, the lack of loud engine booming, intrusive gear noise, buzzing interior bits, shaky mirrors, and the absence of exploding parts. Thank the humble NVH engineer, who spent days holed away in an anechoic chamber covered in glue, tirelessly tapping away with their expensive hammer, tuning your vehicle like a musical instrument to be as safe and enjoyable as possible.

Share on facebook
Facebook
Share on whatsapp
WhatsApp
Share on twitter
Twitter
Share on linkedin
LinkedIn
Share on reddit
Reddit
Subscribe
Notify of
84 Comments
Inline Feedbacks
View all comments
Hgrunt
Hgrunt
7 months ago

Welcome to the site, Steve!

Hope to see more deep dive articles like this this. I’ve been fascinated by NVH and psychoacoustics ever since I learned about how much engineering effort goes into the sound a car door makes when it’s closed. At one point, I had even considered buying an SAE whitepaper on psychoacoustics

Steve Balistreri
Steve Balistreri
7 months ago
Reply to  Hgrunt

Thanks! I’ve actually done a lot of those door closure measurements. Surprisingly the hard part was getting the door to close at a very consistent speed each time.

GertVAG
GertVAG
7 months ago

I really had no idea this was involved in the conception so heavily ! Great stuff to read about.

Ben
Ben
7 months ago

So what you’re saying is that for NVH engineers, when all you have is a $1000 hammer every problem looks like resonance?

Torque
Torque
7 months ago

1st Excellent Article!
2nd NVH from ICE engine is the one of the main reasons why I’ve thought EV would make so much sense for high cost luxury vehicles (Bentley, Rolls. To your point a lack of an ice engine removes all sorts of sources for nvh & these vehicles pride themselves on the refined quiet ride.
Also the high entry price point & limited quantity sold + not really used by the consumer for road trips, thats what the jet is for…
would make for an excellent real like experimentation zone for new vehicle propulsion technology, especially when said propulsion tech is naturally nvh friendly

Steve Balistreri
Steve Balistreri
7 months ago
Reply to  Torque

Thanks! That’s definitely true. Plus with luxury vehicles they can absorb the cost and weight of any extra NVH treatments needed to keep the road and wind noise down.

Citrus
Citrus
7 months ago

One of the most amusing stories of unexpected consequences of resonant frequencies is when a friend of mine put some overpowered subwoofers in his Dodge Dynasty and the trim fell off whenever the bass hit.

Phuzz
Phuzz
7 months ago
Reply to  Citrus

In a Mk3 VW Polo the stereo is at the top of the dash. At around 75-80mph it hits a resonance which starts shaking the stereo if it’s not attached at the back, enough to cause CDs to skip. The only solution was to go faster!
Had it happen in several different Polos with different stereos.

Torque
Torque
7 months ago
Reply to  Citrus

Or the license plate (and anything else), mirrors, door cards, ect… vibrates & rattles like hell

67 Oldsmobile
67 Oldsmobile
7 months ago

this was awesome,thanks.More on this please.

CSRoad
CSRoad
7 months ago

Thanks for the great read.
I’m always terrified with new vibrations in a vehicle and usually think it is a precursor of something coming apart.
The mention of the Mustang V6 driveshaft rang true (-; . I chased the phenomena in my drag race Chevy Nova blaming everything else until we approached a specialist who identified the problem as resonance and built a new drive shaft which lasted well. Losing a driveshaft at over 100 mph is a nasty experience. There is a good reason that NHRA requires a drive shaft loop in anything half serious.

Last edited 7 months ago by CSRoad
Elanosaurous
Elanosaurous
7 months ago

Great article! Fascinating to see the level of detail this is taken to. Looking forward to more of the same in the future.
Steve Mold on YT has a fascinating video of cameras that can capture motion in an amplified form to spot otherwise invisible resonances in large structures: https://youtu.be/rEoc0YoALt0?si=AbSYBDdYM-DtjO-L

Back in the 90s I worked at a computer manufacturer in Silicon Valley (back when they were actually manufactured there) and we got a demo from a UK company doing active noise cancelling (our servers were very loud…). One of their clients was Ford, and their frustration was that Ford at the time didn’t have a spec for the proper Ford exhaust note – there was just a guy who knew what it should sound like, and he signed off on new ones or didn’t. This made doing the analysis to cancel the sound out rather tricky! Presumably different these days…

Steve Balistreri
Steve Balistreri
7 months ago
Reply to  Elanosaurous

Thanks! Several years ago I saw a demo of a camera that used a scanning laser accelerometer that did what you mentioned. It was very cool but extremely expensive.
On the Ford thing, they definitely take the mustang exhaust note seriously, and I think it’s still a very subjective target vs one based just on numbers.

Geekycop .
Geekycop .
7 months ago

Nah. Gimme the straight cut gears, whiny supercharger, and back breakingly stiff ride. It gives the car all of the character, plus at this point I have enough hearing loss and damage from more than 30 years of cabinet shops, straight pipes, and firearm qualifications to not really care anymore about the sound.

Doug Schafer
Doug Schafer
7 months ago

Welcome Steve–loved this! The level of detail is both enjoyable and really informative. And on top of that, it is well written with a light, pleasant tone. More Autopian goodness!

Steve Balistreri
Steve Balistreri
7 months ago
Reply to  Doug Schafer

Thanks I really appreciate it!

Dar Khorse
Dar Khorse
7 months ago

Fascinating article about modal hammers! Do sonic screwdrivers next, please.
As you pointed out, this is even more important with EVs, since there is so much less noise to mask minor sounds caused by harmonics that would go unnoticed in an ICE vehicle. I’m astonished at how quiet my Polestar 2 is when gliding down the highway. Even my much cheaper FIAT 500e, which was just a hastily thrown-together compliance EV, performs surprisingly well when it comes to NVH. Thank you NVH Engineers!

Last edited 7 months ago by Dar Khorse
Lokki
Lokki
7 months ago

This is very, very cool. MORE please!

The F--kshambolic Cretinoid Harvey Park
The F--kshambolic Cretinoid Harvey Park
7 months ago

1) has a career removing noise
2) knows how to use “damp” and not “dampen”

Are you single?

Cheap Bastard
Cheap Bastard
7 months ago

This kinda reminds me of the old woman I was told about who worked at Siemens. They made a particular (I think) MRI machine and it was her job to hit the resonance chamber with a special hammer juust so and dent it juust right so it would work properly.

She was the only one who could get it right every time.

TOSSABL
TOSSABL
7 months ago
Reply to  Cheap Bastard

I hope they told people she was the Fonz of MRI machines

Cheap Bastard
Cheap Bastard
7 months ago
Reply to  TOSSABL

Aaayyyyy….

Steve Balistreri
Steve Balistreri
7 months ago
Reply to  Cheap Bastard

That’s really cool, those MRI machines are fascinating. It definitely takes some finesse to use these hammers.

AC2DE
AC2DE
7 months ago
Reply to  Cheap Bastard

In the satellite comms and microwave radio community, they used to tune antennas for optimum performance by selectively denting the waveguide that carried the signal from the transmitter to the dish. Sounds a lot like the MRI tuning you mentioned!

Fir me, hammers are a precision instrument, but much cheaper than the article’s example. I’m a machinist, and I often have to align a part or fixture using a hammer and a dial indicator. For the really fussy stuff, I sometimes have to get something square within 0.0002 inches over a 6 inch sweep. (Some applications are even more picky; thankfully I don’t usually have to go that far!)

84
0
Would love your thoughts, please comment.x
()
x