I'm on the Formula SAE team here at UA, and they are putting together a muffler design contest for the club members. Basically, we have to design a muffler for our 2007 car, which will be judged by some of the team leaders. The best looking designs will actually be built and dyno-tested, and whichever muffler loses us the least horsepower will be the one we put on the car for competition.

So, I'm looking for some resources (online or otherwise) which might point me in the right direction for muffler design. I've heard that most of the stuff you can find via Google basically tells you to build a design, see if it works, and then rebuild it if it doesn't, which obviously doesn't help me much. So, I was hoping maybe some of the engineers on here might be able to point me to some textbooks, SAE technical papers, etc. that I could use as a starting point.

If that doesn't work, I was thinking of just taking a look at cutaway pics of several mufflers which I know flow decently well (Borlas, etc) and try to "re-create" those designs. I'm in mechanical engineering, but I haven't taken any fluid dynamics classes or anything of that sort yet, so I don't have much hard science knowledge to work with.

Anyway, sorry that was so long, thanks for any help!

 

Thoughts:

Minimum restriction is Job #1. A straight-thru pipe would be best, but too noisy of course. Many free-flowing mufflers just use perforated cores with sound absorbing packing around them.

As you have Db limits, perhaps you should analyze the frequency spectrum of your engine exhaust and decide which frequencies you want to minimize. Look at Corsa's reflective sound cancellation technique:

http://www.corsaperf.com/arsctech.htm

I have seen this used with tubes attached at right angles to the exhaust pipe rather than how Corsa does it in their mufflers. Each stub can be tuned to cancel a certain frequency. You might need three or more to shape your noise curve.

A simple way to tune it is to make the development stubs long and straight, and insert a movable plug on a rod, run the engine under load and slide the plug in/out until you cancel the offending frequency. By using 3 or 4 or more tubes, you can tune the sound to almost anything you want. When you get there, you could fabricate a nice "bunch of snakes" clustered around your primary exhaust pipe to look cool. Obviously you want to flow the final result on a flowbench vs. the straight pipe to see if you are efficient.

For the Helmholtz theory guys, they can do all the calculations for lengths...just so they end up with what you get experimentally. This whole process would make a nice write up. So would the methods you used to get the thinwall bends in Inconel or whatever you choose. Think slicing a hollow bagel to make nice tight bends.

An added benefit is that you can still have your collector the correct length for best torque curve shape, or whatever you are going for as well as get a legal and "cool" sound.

My $.02

 

We make more power on the dyno with Flowmaster mufflers baffled design...scavanges better than a straight through muffler.

 

Thanks for the replies, this is all good info. I took a look at the rules, and the max noise level is 110dBs. This is measured 0.5m from the end of the tailpipe at a 45-degree angle with the horizontal. The test is done in neutral at whichever engine RPM corresponds to a piston speed of 3000ft/s. At first glance, 110dBs seems pretty generous for a 4-banger motorcycle engine. Am I wrong, or do I not need to focus a ton on noise reduction?

Also, Machinist, who do you mean by "we"? We as in Camaros in general?

 

Wasn't there a big noise problem in 2005 competition forcing some last minute cobbling? I think you need to focus at least 1/2 ton on noise and lack of backpressure. 3000 ft/min sounds more like piston speed than 3000 ft/sec, right? P/S = stroke * rpm /6. For a 2 inch stroke, that's 9000 rpm, which should make considerable noise about 20 inches from the tailpipe.

FWIW, many folks find baffled Flowmasters restrict power more than stuff like Magnaflow or Dynomax welded, etc which are straight thru designs. Look at what guys/gals in the Engine Masters Competition use.

Are you guys new to the FSAE competition?

 

No, our team is not new to the competition (this year was our third entry in four years), but I am. And yeah, 3000ft/min is what I meant, you can chalk that one up to not thinking while typing

Anyway, thanks for the ideas, I'm definitely interested in what I can do with a design similar to the Corsa you linked to above. If I can cobble something together in SolidWorks that looks decent (I'm new to that too), I might post some drawings of it up here.

 

I envision 3-5 tuning tubes about 1 inch diameter on a 1-3/4 or so collector diameter. The tuning tubes bend tightly from their junction with the collector like fingers sticking out of a hand, and point upstream perhaps even spiraling around the collector. They could all spiral to the bottom to keep weight low. To me that would be cool looking. You could even machine fancy end caps and Tig them on.

This technology was developed in the mid 1960s at a major OEM, but wasn't used for a full exhaust system. in production. It was just used as a resonator at the end of the tailpipe in a few models.

I think it was Japanese entry in the 2005 contest that had quite an exhaust system. It might have won some award.

Are you using E85? The fuel mixup in 05 was a shame.

 

There is way too much to say here so I'll try to condense it a little.

The .5 m @45 degrees is a standard location and 110 dB is a lot, but sound pressure levels drop more (non-linearly) as distance increases. You're right there close to the pipe so sound pressure will be high. Is it flat, A-weighted, or C-weighted?

Engine type is critical. It will determine what frequencies make the largest overall contribution. Is it a motorcycle engine as you mentioned? Inline or V? Motorcycle engines won't require nearly the tube sizes to minimize back-pressure that larger car/truck engines would.

In 4-cylinder engines the 2nd order sound usually dominates, with 4th and 6th orders contributing to overall. At 1000 RPM, 2nd order is 33 1/3 hz for a 4-cylinder engine. A V-type engine may introduce (from pipe configuration) various 1/2 orders that are difficult to tune. The best way to get rid of these orders is to make sure the y-pipes have equal-length legs to the junction.

There is a reason you won't find too much on the web about muffler design. It's because there's a lot more to the designs than at first meets the eye and because companies still do muffler development to sell them. No one posts muffler internals because they don't want to give proprietary information away. A stick diagram is a far cry from a producible, functional muffler design. Even Corsa's and Flowmaster's cutaways don't give away the farm. That said, here are some pointers:

Two major things make exhaust quieter, tuned volume and back-pressure. The more volume you tune with, the quieter it can be given equal back-pressure. Conversely, for a given sound level, the higher the volume, the lower you can make the back-pressure.

The Helmholtz equation works. If you want to tune to a frequency you can use a Helmholtz chamber. Chamber volume, tuning tube diameter, and tuning tube length drive the functionality. More volume, better tuning. Larger throat diameter, broader-band attenuation. Since the speed of sound changes with temperature, you have to compensate some lengths to take into account the hot gasses.

OldSS perfectly described a quarter-wave tuner. You can attenuate a very narrow frequency with a capped tube the length of a quarter the wavelength. The adjustable quarter-wave tuner was called (and this was before my time so I'm not certain of the spelling) a kwinkie tube. We don't use them much any more because we can measure sound and rpms simultaneously and get order information (and thus individual frequencies) on a waterfall graph.

As OldSS said (he knows his stuff!), straight-through designs are lowest in back-pressure. However they tune very little in the lower frequency spectrum. Packing materials typically absorb sound above ~600 hz and almost nothing below. A tuned muffler will usually have a three-pass design or some type of flow reversing and one or more separate tuning elements inside the shells.

There is an enormous amount of work that goes into a modern exhaust system. They have to live through some of the most corrosive environments around, at high heat, and tread a fine line between cost, weight, sound quality, back-pressure, and federal noise pass-by requirements.