Engine Theory: Exhaust

Last month we chronicled how air and fuel get into your engine’s combustion chamber. This month we’ll see how that air escapes back to the atmosphere via the exhaust. Unlike the seemingly laid-back intake side of engine breathing, there’s impressive energy involved in the exhaust system. That’s both boon and bust when selecting an exhaust system.

Like everything else, not only is exhaust system design fraught with compromises, but it’s also a gray area of responsibility. Engine companies may or may not offer an exhaust system—Rotax does, Lycoming does not—and airframe manufacturers consider exhaust more an engine consideration than something they’re equipped to handle. Even worse, many times no one is seriously thinking about the exhaust until well into the aircraft’s design or construction. This is especially true with homebuilders, where exhaust is often an afterthought to a builder eager to finish his project and go flying.

Thankfully, exhaust system appreciation has risen in the last decade or so. More specialist exhaust manufacturers are available to service the homebuilt market so quality tubing, bends, flanges, tabs, spring mounts, and mufflers, not to mention entire systems, are waiting on the shelf.

Big Energy, Big Waste

Burn gasoline in a laboratory and it’s easy to determine how much energy is released—a lot! Unfortunately, in a piston engine, if you’re lucky, possibly a third of that energy is harvested for productive work, while a second third is radiated as waste heat by the cooling system, and roughly another third (or much more) is expelled as hot waste gas through the exhaust. Clearly there’s potential for turning something we’re throwing overboard to good use. And just as obviously turbosupercharging is aviation’s exhaust energy recovery system of choice, which we’ll discuss in a separate article.

But exhaust gas is also hot, dirty, poisonous stuff. Sure, the main byproducts of burning gasoline are CO2 and water (as a vapor), but the reality is, piston engine exhaust reeks of trace elements, many of them harmful. As a consequence it’s often far easier to simply get rid of the stuff as fast and inexpensively as possible and call it good. And that’s mainly what we’ve been doing in Experimental aviation for decades.

Furthermore, we pilots have an extra incentive to be rid of the exhaust: backpressure, the Great Satan of air-cooled engines. Cylinder head temperatures soar quickly when a restrictive exhaust system is asked to process large volumes of hot gas (takeoff and climb), and those hot heads lead to hot oil and high heat rejection under the cowling. The result? More draggy cooling airflow through the engine compartment, aggressive heat cycling of the cylinder heads, and a reduced detonation margin.

Water-cooled engines handle backpressure heat less strenuously because water’s higher density doesn’t allow ballooning of CHTs and can more easily, and evenly, absorb heat from localized hot spots, such as the cylinder head around the exhaust port.

No less importantly, backpressure is a hindrance to engine breathing. If the exhaust doesn’t exit promptly and completely, there is less room in the combustion chamber for the incoming fuel/air charge. Power and efficiency suffer.

Exhaust System Tuning

As with intake systems, exhaust system physics include the expected mass flow of exhaust gas, plus the more abstract wave tuning.

Mass flow is mainly intuitive—hot gas at high velocity rushes past the exhaust valve into that cylinder’s primary exhaust pipe where it cools (but still stays burning hot as the EGT instrument shows) and slows while rushing for the pipe’s exit to the atmosphere. Pipe diameter and shape play the expected roles in either smoothly guiding the slug of exhaust gas down the pipe, or, if the system has crimps, sharp bends, poorly formed intersections, or other airflow speed bumps, the system introduces flow-killing turbulence and backpressure. Perfectly practical aviation exhaust systems are designed using nothing more than good old mass flow thinking. Keep the pipe sizes reasonable for the gas volume, make a minimum number of bends, make no sharp bends, keep it simple, and you’re good to go.

Far less intuitively, pressure waves form inside the exhaust primary pipes. If the primaries are joined to form an exhaust manifold, the waves and gas flow from one cylinder can influence other cylinders positively or negatively. Wave tuning is truly arcane stuff; engineers write thick books full of math on the subject, and the physics of it responds to a bewildering array of engine parameters and environmental conditions. That’s all fun for the propeller heads, but what matters if you’re trying to put some nice pipes on an RV is primary pipe diameter and length. Diameter is important for mass flow, sure, but also to pressure wave speed. Too large a primary pipe results in too slow wave (and gas) speeds.

Pipe length is the most important factor in wave tuning. Curiously, pressure waves run down, then reverse and run back up the primary pipes. Get the primary pipe length synched with the oscillating wave speed, and you’ll have a pressure wave just starting to race down the primary pipe when the exhaust valve opens. That helps scavenge the exhaust from the cylinder and better engine breathing results. Get the pipe length wrong and a pressure wave is racing toward the cylinder when the exhaust valve opens; gas flow is stymied, breathing suffers, and cylinder-head temps rise due to backpressure.

How do you determine the correct pipe length? Consult an exhaust specialist offering design software services. Burns Stainless, for example, offers X-Design, which is their proprietary software. For about $80 they’ll plug all your numbers (engine displacement, valve size, compression ratio, port dimensions, detailed camshaft specifications—they’ll want a bunch of numbers and the more you can provide the better) into their computer and report back the workable pipe diameters and lengths.

Because wave tuning resonates in frequencies there are first, second, third, fourth, and many more orders of waves to work with. Typically a 360/540 Lycoming cylinder first order pipe length is about 78 inches, so the second order length of 34 inches is far more practical. Third order would be half of 34, or 17 inches, which might also work, but will likely be tough to package.

Bad pipe lengths peak halfway between the good pipe lengths. For our 360/540 cylinder you’d definitely want to avoid 51-inch and 25.5-inch primary pipe lengths.

Naturally, all primary pipe lengths need to be the same length if you’re making a tuned system. This is your goal, as you want all the cylinders to run as similarly as their neighbors, mainly for aggressive lean-of-peak operation, but also for peak power. Burns Stainless considers +/- -inch as close for a racy exhaust (you’ll have fun developing that pipe layout), but you can likely be +/- 2 inches and not be able to measure the difference on a daily-flier engine.

Mind you, all of this tuning is appropriate only if designing a tuned exhaust system as we’ll detail shortly. If you are not using a tuned design, then practically speaking, pipe length doesn’t matter. Also, all the usual gray and gold general aviation engines run camshafts and rpm designed for maximum torque and fuel efficiency—they never rev high enough to reach their horsepower peak. So tune for torque; you’ll never see the power peak anyway.

Popular Exhaust Designs

Stub Stack—Several aviation exhaust designs have remained popular for decades. Simplest is the abbreviated stub or short or open stack. From the “get rid of it” school, stub stacks are very lightweight, inexpensive, pose few packaging challenges, and essentially no backpressure. They’re also louder than an indoor firefight, popping and barking annoyingly. They also offer no carburetor or cabin heat, nor tuning possibilities. Just the same, given the torque-centric camshaft and ignition tuning of traditional aviation engines, stub stacks will make all the horsepower a typical Lycoming is going to make, and not lose too much torque.

These days stub stacks are often run in dyno cells because they’re easy to bolt on and cause no backpressure-induced CHT worries. They were popular on some antique radials and met glory on the big V-12s of WW-II fame. There they sound great, thanks to the full chorus of cylinders.

Rolls Royce recognized the tremendous gas flow roaring from their Merlin’s stub stacks added measureable thrust if the pipes were aimed aft and given somewhat restricted, high-velocity exits. Called ejector exhausts, Rolls Royce calculated they were worth about 150 hp of propulsion. The physics remains valid, but even with what we now consider a “big” 300-hp 6-cylinder, the exhaust mass is puny compared to a 1700-cubic-inch supercharged V-12’s. A jet a LyContinental really isn’t. Furthermore, the worthwhile efficacy of ejector exhausts at altitude are at cross-purposes to sea-level horsepower (they are an airflow restriction given thick intake air). A variable exhaust nozzle might help a cross-country cruiser up in Class A airspace, but seems more weight and complexity than it’s worth for lower altitudes.

Manifold Design—Easily the most popular exhaust design in general aviation is the “manifold” type, sometimes called a single side stack. Simplicity is its virtue; by running all primary pipes on one side of the engine into a single exhaust pipe, things are kept lightweight, physically robust, easily and tightly packaged, even affordable, and definitely less raucous than stub stacks.

No tuning is possible with manifold systems so primary pipe length is not a factor. It’s as simple as running the first cylinder’s pipe into the second cylinder’s pipe—and continuing into the third cylinder’s on a 6-cylinder engine. Because gas volume rises as more cylinders are added, pipe diameter should be increased as each successive pipe joins in.

For daily fliers manifold exhaust systems are the clear front runners. They aren’t the most efficient engineering solution, but for standard duty they’re where the cost, power, noise, cabin or carb heat, and packaging curves intersect.

Individual Pipe—Lengthen the stub stack to about a yard and noise subsides and torque rises. The Swearingen SX-300 is just one Experimental employing such a system, that is four or six individual pipes thirty-something inches long. Essentially this is a set of headers, but without the collector, and such systems respond to wave tuning, so you want to get the pipe length correct. Without the collector the tuning effects are peakier (work in a narrower rpm range) and not quite as powerful, so the theoretical power and torque curves won’t be quite as broad as they could be, and the peak numbers should suffer slightly as well. Given the narrow operating range of our engines these are minor losses, and for the experimenter looking for superior exhaust efficiency without the overreaching cost of headers, individual pipes are a smart and often overlooked choice.

All told, such a system is a major improvement over stub stacks. Noise is reduced, carb and cabin heat can be collected, cost is reasonable, and the pipes are easy enough to package and fabricate.

Headers—Take the individual pipe design, bring the exit end of the pipes together in a collector and you have a set of headers, or 4-into-1s as the O-320 and 360 crowd call them. This is the ultimate in exhaust plumbing on a naturally aspirated engine and the most tunable design.

The big attraction is bundling the primary pipes in the collector. Again, timing is the key concept; joining the primary pipes together with the engine’s firing order in mind means a mass of exhaust gas racing out of one primary pipe puts negative pressure in the collector. The next cylinder’s slug of exhaust gas thus meets less resistance and therefore speeds up. The higher gas velocity in the primary pipe helps better evacuate the cylinder and, similarly to wave tuning, better engine breathing results.

Just as primary pipe diameter and length are important in this tuning, the collector is also tunable, although typically not quite as critically as the primary pipes. Again, there are many variables, and an exhaust specialist and his design software are the answer.

Headers are a fact of life in tuned-up car engines and, as a practical observation, once specific output approaches 1.2 hp per cubic inch in a naturally aspirated engine, some sort of header, even a compromised production piece, is a necessity.

So, if headers are so neat, why doesn’t every Piper Archer have them? There are many good reasons, but the engineering answer is stock airplane engines don’t get close to 1.2 hp per cubic inch, so the gain from headers is minimal. Headers gain importance as exhaust gas volume rises from rpm, supercharging, or hot camming and compression with their attendant greater fuel burn.

And there are hassles with headers. Costs skyrocket if headers must be custom built (available off-the-shelf header systems such as for RV-10s are more cost effective). All those exact-length primaries are tough to package, leading to prolonged design time, and the many linear feet of primary pipes contribute to under-cowling heat and require plenty of fabrication time. They also take up space, compromising service access.

Benefits are extra torque at high manifold pressures, a practical elimination of hot CHTs because backpressure is nil or even negative, excellent parity among cylinder EGTs and CHTs everything else being equal, and done right, a far more melodious and partially lessened exhaust note.

Headers are a folly for pragmatic run-of-the-mill aircraft and the racer’s, aerobat’s, cross-country speedster’s, and backcountry STOL artist’s must-have. They’re also an excellent special touch for the experimenter wanting the fabrication challenge or looking to optimize each cylinder for aggressive lean-of-peak fuel economy.

Crossover Exhaust—Lycoming 4-cylinders employ an unusual firing order that sees both cylinders on one side of the engine firing closely together followed by the two cylinders on the opposing side doing the same. The result is exhaust gas from one cylinder partially chokes gas flow from its adjacent cylinder. This is a meaningful effect on manifold systems and a prime generator of backpressure.

Obviously stub stacks or headers would eliminate this firing order concern, but not without introducing cost, noise, and other issues we’ve already discussed. Therefore it’s a good step forward to run an exhaust pipe from one cylinder to the other side of the engine simply to even the timing of exhaust gas pulses flowing through the pipes. This is an oddly tuned system—pipe lengths vary wildly among cylinders—and scavenging effects can be pronounced or nulled depending on the combined pipe length. (For tuning purposes the entire length of joined pipe, even those running backwards of gas mass flow, are in play).

Crossover exhausts are a good choice on 4-cylinder Lycomings because cost, noise, packaging, carb and cabin heat considerations are all reasonably handled. And if you’re thinking of lengthening the two short primary pipes in a crossover system until they equal the long primary pipes, you’re thinking correctly—and leading yourself to a set of headers, which is what you just imagined.

Exhaust System Pecking Order

There are many variables in selecting an exhaust system, but to simplify the choices, here is how we’d generally rank exhaust system designs for typical applications.

Getting Practical

Nut and bolt details of exhaust system design are fodder for separate articles, but a few points regarding materials and stress relief are important enough to mention here. The takeaway on materials is 321 stainless steel is the aircraft standard because it is the least expensive material (and we’re definitely speaking relatively here!) that works. You’ll find plenty of lesser stainless steels available, such as 401, but 321 has the mechanical strength and corrosion resistance to withstand aviation use where the engine runs at high output almost all the time. Don’t even think of mild steel as it will wear quickly. At the other end of the spectrum, super nickel alloys such as Inconel are too expensive and more difficult to fabricate. Their superior heat resistance is not needed in a naturally aspirated system anyway.

As for stress relief, aviation exhausts run at sustained high temperatures and thus grow considerably due to heat expansion. Plus, air-cooled aviation engines expand like balloons to begin with, and shake like wet dogs when starting and shutting down, so the exhaust must be flexibly mounted in addition to being free to slip between pipe sections. Ball joints are typically used to provide the needed suppleness, but slip joints work as well.

Next time we’ll examine ignition and engine management systems.

Further Reading

An excellent summary of exhaust testing and tuning for 4-cylinder aircraft engines can be found on the Experimental Aircraft Association’s Member Hangar web site.