Good post, OldSS. The OEM was Arvin. It was in the '60s. Put a closed pipe off the main exhaust pipe and tack it further down so it runs parallel. Arvin called it the Arvinode. It was a simple quarter-wave tuner. It's very effective at targeting a single frequency band. A Helmholtz tuner is a similar concept, with a volume combined with the tube. It widens the attenuation band and works better.

I don't think Harley was successful in receiving a patent for their sound. I think the Patent Office realized how stupid the idea was. I could be wrong, but I don't think so.

Exhaust tuning is 1 part science and 1 part black art. The guys who are really good at it have been doing it for years. They can look at an engine and pipe configuration and give the OEM an 85% good solution on the first try from prototype. The money is in the remaining 15%. Computer simulations are lucky to get halfway there on the first try.

Lots of things influence the sound coming out the tailpipe such as: Engine configuration (more cylinders, higher firing frequencies), pipe diameters, pipe lengths, unequal primary lengths (worst case is some cast manifolds), unequal y-pipe lengths, distance to silencing units, and of course, muffler volume and configuration (plus any number of other things). OEMs target a particular sound and engineer the exhaust for what they've got.

One thing that sticks out as particularly relevant for Ferrari is that there is minimal pipe length from the engine to the silencing unit. Mid-engine setups have that problem. It's important because the pipe length influences how effective the muffler is. Too little and the muffler isn't in the right place for problem frequencies. The Ford GT is a nightmare to tune because of this. I don't know exactly what the Ferrari setup looks like, but I can guess a little. Relatively small displacement, high-revving V8, longitudinally mounted, tubular headers (probably close to equal length), short intermediate pipes (again, close to equal length) into a tuned, stamped muffler. Tune the muffler for lower frequencies and not the higher ones, so when you rev, you've got the full sound in the upper RPM range. Result is higher frequencies predominate, the odd-sounding engine orders are cancelled, and it sounds clean and pretty distinctive. There is also some induction noise that mixes with exhaust to make the overall sound. That is tuned as well.

It would take a lot of work, and fabrication to make an F-body approach the sound of a Ferrari, and I don't think you'd ever really get there. Too many differences upstream of the tailpipe. You could make a very nice sound with equal-length headers and downpipes, an H or X pipe and the proper mufflers, tuned, not just some bullet resonators hanging there. Packaging is a nightmare though. You'd have to give up a lot of clearance on the left side (remember, equal length is important). You'd probably need to give up a few horses to get the right tuning as well if you wanted to make it sound great instead of just loud. Oh yeah, and rev it to 8500.

Another thought on the quarter-wave tuning (organ pipe) OldSS mentions. The frequency being attenuated is related to the diameter and length of the pipe. Larger diameters need longer tubes to be effective for a given frequency. Short tubes, small diameter= narrower frequency band. Larger diameter, longer tube=wider band attenuation at the same frequency. Too small=not effective. Same can be said for the tube in a Helmholtz tuner, but you now have the chamber volume that must change as well.

Enough typing... My $.02

 

Good post Nate.

I was telling OldSStroker "I wonder when Nate will chime in on this post?"

We have to get you two together with a few beers and design some exhaust systems, at least one for my 64 GTO and his Lakes Modified.

Bret

 

Great post, Nate. Many good points that I didn't think of.

Maybe even before Arvin's work, or along with them, a Pontiac development engineer spent a lot of time on this project. It was '63-'65 if I recall. I wonder if any patents were ever issued. The engineer was Tom Cassel (spelling?). He'd be in his 70's now.

How critical would it be to tap into the main exhaust tube art a positive pressure node with the engine running at a particular rpm, say 65 mph cruise? That seemed to be important as I observed Tom's work. I was an engineering student then, and spent a lot of time observing and learning. FWIW, I was fortunate enough to work with/for some really sharp guys back then.


So how about an exhaust processor with a Helmholtz resonance chamber with a variable tuning throat. The tuning frequency of the exhaust processor changes as the outlet opening from the variable tuning throat into the Helmholtz resonance chamber changes. Plug it into the PCM to control the sound. Think it would work?


Thanks again, Nate.

Jon

 

Proof that I have no business correcting anyone.

They should fire that team of programmers.

Programs can and have been extremely accurate in the most dizzying array of simulations. GIGO, gotta have sound mathematical models if you want to get there though....

Very interesting response!

The OEM guys have it tough though... they have to design an exhaust system that doesn't hamper performance too much, yet meets stringent noise control criteria. If I worked in that field, I'd be fired cause they'd all come off the block sounding like a 426 hemi..... rickity-tickity-rickity-tickity.

-Mindgame

 

Jon,
I'm sure there were patents issued, but I don't know to whom. The timeframe you mention is exactly when that activity was highest, so you're right on with the recollection. I did see mention of an ArviNode on a '64 Fairlane. I don't know for sure, but I don't think too many actually made it to production. Probably due to cost and durability issues.

Location of the silencing element is one of the most critical things to consider. It's also one of the most ignored in OEM circles because of packaging (you know spare tires, fuel tanks and such in the way). If you are targeting a specific frequency, you need to place the silencing element at the quarter-wave point of that frequency, or as close as possible. It can be measured from the engine or from the tailpipe, but a lot like to measure from the tailpipe and place the box with as much pipe behind it as possible. That's why tuning a mid-engine car is so difficult. There is little pipe length before or after the muffler, so getting the tuning elements placed properly is next to impossible. Given that the optimal never occurs and designers have to target noise in several areas, it's a testament to how well they get things to work. It's amazing to me sometimes that they get some of these things legal.

With the organ pipe, placement is critical. It could be completely ineffective if put in the wrong place.

If you know what RPM you want to target, you can get the fundamental frequency. From the freqency you can get the wavelength and you can place the pipe. If you have an unequal-length y-pipe though, the problem might not be at the fundamental firing frequency. It might be some odd order (like 1.5 times or 1.625 times firing frequency) and that would change the requirements. The 80's and 90's Pontiac V-6 blatty sound is an example. If you have equal-length downpipes and an h or x pipe, the frequencies you hear will be predominately fundamental firing. Think Mustang GT for example.

I think your PCM controlled system would work for a narrow range of conditions, but it would never make it past the pass-by test. The latency associated with changing the tube length would mean it wouldn't work for the first part of an abrupt throttle change. The pass-by test involves approaching a trap at a set speed (30 mi/hr or 50 km/hr, depending on the test) and going to WOT at a pre-determined point and going through the trap.

You're touching on the active noise cancellation that we talked about a few weeks ago in another thread. Doable, but very expensive and not too durable.

Interesting note: Many years ago, development people used to use a "kwinkie tube," that was a variable length tuning throat, exactly like you describe. They'd use it to quickly identify the problem frequency by tapping into the exhaust and changing the tube effective length until they got some cancellation. Then they'd tune the prototype muffler to hit that frequency. Nowadays, a frequency sweep with respect to RPM on a dyno roll lets you see immediately all the problems at any point in the RPM band.

Mindgame,
I know just how you feel on how it should work, and I agree. The problem is there is still not enough information on everything that contributes to and affects the noise. The big things are easy, but other stuff is difficult to quantify. How about quantifying the acoustic effects of an aluminum engine block vs. an iron block. Does it affect sound out the tailpipe? Pipe thickness, material? Hanger locations? Cam timing events (obviously, but quantify it)? Calibration and strategy in the ECM? Interactions between any of the above? It gets seriously complex. The software is admittedly better than it used to be, but it still has a hard time with a lot of configurations. There is also the aspect of time. It can take someone days to set up the whole simulation properly, with the system modeled in the computer. The development engineer can get prototypes very quickly if he has to. It's most often faster to order a batch of prototypes and test them than to try to model all the muffler internals separately and run simulations. Believe me, simulations do figure prominently in a lot of development, especially when you don't have a test vehicle or engine, but when it comes down to the end of development, the final parts are designed, prototyped and tested by a person. It happens often enough when the simulation suggests one thing and actual testing shows different. No one is giving up on computer simulations though. It's going to happen more and more. The models get more and more accurate every year.

Sorry to post so long.

 

Post as much as you want Nate... we'll read it.

WAVE is an extremely accurate simulation program that can do just what we're proposing. I don't work in the automotive field but I know of WAVE from my dealings with Chuck Riddeck last year. Maybe he'll chime in.

Here's the information from Ricardo's website.

What amazes me these days is the amount of information out there. Here we are having a discussion on exhaust systems and how to create the sound you're after..... almost by accident I found this article on the new GTO. Fascinating... cause it goes right along with what everyone here has been saying about creating a "characteristic" sound. The engineers went through a bit of effort to do just that.

http://ultimategto.com/monaro13.htm

-Mindgame

 

Wave is the big player in the exhaust simulation field. It is well used. It's very cool, but the truth is that no one dares to take the results of a Wave simulation and make production tooling decisions based solely on them, especially late in the program, when tool changes are serious $$$$. You always test to see if there is a difference and there nearly always is one, even if it's a five percent variation in some areas. Then, you need to tweak the design using experience and trial and error to get it right. That's on a simulation that comes out really good. There are others where the real-world results aren't really close to agreement with the model, and a human could make a first shot much closer. It becomes a time thing, which is faster?

Wave is used in the industry to try to narrow the designs down in the absence of a vehicle. We do development without vehicles more and more early on in a program. It is very useful, but it isn't dead on and you always correlate the data. It is, as I said, getting better. It isn't cheap either.

 

Worked as a program director for a while. I can tell you about the cost of developing accurate software applications.

You're right Nate, the program is not going to give you a perfect model, but it does create a good starting point. There are always going to be a number of prototypes based off that original concept. The importance here is getting that "original concept".

The GTO team I referenced in my last response... same situation there.

The beauty of software goes much further than original design. You have a design team working on the project. An original concept is developed, that concept is used to create a computer model. That computer model is associative, in that any changes in say a chassis component will automatically update the exhaust design should it fall out of certain constraints. From original design, you have an accurate model for the simulation. That model is used for simulation and it is used for manufacturing. Speeds the entire process.

I guess the point I'm trying to make here is, it goes further than getting that first design. Nothing cheap about it when you look at initial expenditures, but the return on investment can be substantial for a company that does this continually.

-Mindgame

 

Yup, that's how it's supposed to work, and sometimes it does. What do they say about it works that way in theory, but not always in practice? There's a catchy saying there somewhere that I can't remember. Jon probably could...

I need to comment that several of you demonstrate remarkable understanding in a wide variety of areas. Jon, Mindgame, you guys seem to know a lot about a lot, even though the topic may not be in your primary area of expertise. You, and several others, specifically Bret, Ken, Fred, and more make it a pleasure to browse the forums. Thanks...

 

Advanced Age helps a lot. Old Farts like me have just experienced more, read more, talked more(!) than you young guys. Some of it is retained. Keeping the mind active seems to help retain more of the stuff that has been accumulated over the decades.

BTW, I don't have a primary area of interest. That's too boring.

Thanks, Nate.

 

The difference between theory and practice in theory is much less than the difference between theory and practice in practice

That the one you're looking for?

Not as old and wise as some around here but extremely eclectic. To myself and probably alot of other gentlemen here... living is all about learning.

Now if I could just motivate my nephew....

-Mindgame

 

WOW
I know this were the place to ask for.
I love the corsa sound, this C5 sound are just what I looking for my Z but they donīt sound even close in the C5 to the LT1 or LS1 Z.

C5 http://www.corsaperf.com/flashc5.htm

LT1 http://www.corsaperf.com/flashcam-fire.htm

LS1 http://www.corsaperf.com/gm/cfflash.htm

 

This kind of makes me wonder about those tuning for power and sound. Do guys just add spacers to the headers in small incrememnts? What size are these increments (its gotta be based on wavelenth but I don't know it).
Also, this reminds me of the polarizing experiments i did in an old physics class and the comments of a chemistry teacher.
Has anyone ever tried using a grating with very small slits in only one direction? It'd be very restrictive so you'd have to make the diameter of the tube alot bigger at that point.
It'd let through only one plane of waves, and you could have another grating set up further down that you could adjust by turning it.

edit: also, to expand on his question, what is it that made the old trans-am series cars sound so real?

 

Yeah, the c5 corsa does sound pretty amazing, i have to wonder if the differance is in the exhaust itself or the engines, the ls1 and c5 have simular engines if not the same so u would thing they would sound simular???

 

That's an extremely high revving (7200+ rpm) LS1 in the C5 soundfile. I wonder if they "enhanced" it.

Here's what a 6800 rpm shift sounds like:

http://kraest.com/vette/2ndgear.mov