Pelican's Perch #78: Props Driving Engines

  • E-Mail this Article
  • View Printable Article
  • Text size:

    • A
    • A
    • A

If your CFI (or mechanic) instructed you never to let the propeller drive the engine, or never to run the engine 'oversquare,' it's time for a reality check. AVweb's John Deakin addresses a few more of these engine myths that started back in the days of radial-engines.

Pelican's Perch

I think I'm seeing the rapid growth of yet another OWT (Old Wives' Tale). Some pilots are going to extraordinary lengths to prevent this badly misunderstood phenomenon, and a few are creating far more danger for themselves and their airplanes as they try to avoid a problem that probably doesn't exist.

I believe the Commemorative Air Force (formerly "Confederate Air Force," and yes, I like the change, thank you very much) lost the world's only flying example of a Heinkel 1-11 -- a WWII German medium bomber -- because a pair of highly experienced pilots, good guys, good pilots, probably bought into this foolish OWT. Many, perhaps even most, of the pilots flying these precious old relics are also making the same mistakes, placing themselves, their passengers, and the airplane in harm's way, when it's not necessary.

Like many OWTs, there is indeed a germ of truth behind it all. But it has become so overdone that it's ridiculous. Follow along as I try to explain.

Just what to we mean by "the prop driving the engine"?

Picture an airplane in steady cruising flight. The engine develops the power by combustion, pushing on a piston (a reciprocating motion), which is transmitted to the crankshaft by the connecting rods (converting reciprocating motion to rotary motion), then to the prop shaft, and to the propeller, which converts the torque to thrust.

Can We Stop The Prop?

Now, assume for the sake of this exercise that we simply shut the engine down with the mixture. The combustion stops, and now the prop becomes like a child's pinwheel in the window, spinning in the breeze. In many airplanes, the RPM won't change at all; the constant-speed prop (controlled by the governor) simply twists to a flatter pitch to maintain the previously set RPM. Many pilots, taught by 300-hour wonders, just assume the prop will stop if the engine quits, but it's actually very difficult to make one stop. On my Bonanza, I have to use full flaps, throttle fully closed (making the pistons suck harder against the resistance of the closed throttle plate), and fly right at or in the stall (well under 50 knots) to do it. The prop will take many seconds to slow down. With patience, it will finally stop against a compression stroke. After it stops, it takes about 90 knots to start the prop turning again with airflow alone. When it does, it will tick past one compression stroke, stop again, squeeze past another, stop again, and finally will make it through that next one. From there, it spins right up to whatever RPM the airspeed will produce. Pushing the mixture back in will cause it to start right up again, of course.

During the time the mixture is off as above, the prop is driving the engine. Roughly speaking it takes about 30 HP to turn the IO-550, so that's what the prop has to produce to turn it.

Why Is This "Bad"?

To get to the real roots of this, we need to go back nearly 60 years, to the heyday of the big radials. Regular readers will perhaps remember that I've said many times that the principles, the metallurgy and the combustion event are identical in all internal-combustion, spark-fired, fixed-timing, gasoline-powered engines, from the Briggs & Stratton lawnmowers to the giant R-4360s in the B-36.

However, there is one crucial difference between radials and "flat" engines, and it lies at the heart of the OWT that is the subject of this column. Most readers will be at least faintly familiar with the layout of the usual flat-six Continental or Lycoming engine: six cylinders, three on each side (giving rise to the term "horizontally opposed"), with each cylinder having its own connecting rod, and each connecting rod having its own crankshaft journal and bearing. Note that this means each crankshaft journal will take the heavy load of a power stroke (up to 18,000 pounds) only once every other revolution, because it is a four-cycle engine. Keep that thought: a hard push, but only once every other revolution.

For comparison, take what became the most widely used, big radial: the superb Pratt & Whitney R-2800. There were something approaching 130 models of this engine made, producing anywhere from about 1800 to 2500 horsepower (double that at the Reno air races). They pulled a huge variety of aircraft through the air, from the Grumman F8F Bearcat to the Douglas DC-6. This 18-cylinder engine has two rows of cylinders, nine in each row.

This engine works great and lasts a long time. But with all nine cylinders arranged in a circle around the crankshaft, the designers couldn't have a journal for each connecting rod.

(30 Kb)
One cylinder row in the P&W R-2800 (click photos for larger versions)

They very cleverly solved this by making one massive rod that runs from one piston/cylinder to the sole crankshaft journal for one row. The "big end" is attached to the crankshaft, and the "little end" is attached to the piston, just as in our flat engines. But around the circumference of the "big end" there are 8 evenly spaced holes, and each of these will have an "articulating rod" bolted to it, with the other end attached to the piston. Since this is also a four-stroke engine, each cylinder will produce a power stroke every other revolution. With nine cylinders, that means 9 power strokes for every two turns, or 4.5 power strokes per turn.

Put another way, that radial crank journal takes nine times the number of power strokes as the journal on a flat engine in any given time at the same RPM.

Radial engines fire every other cylinder in turn, and are numbered clockwise as seen from the rear. The top center cylinder is generally #1, so let's start with that, and treat this as a one-row engine for the moment. As the piston rises on the compression stroke on #1, it slows and then stops momentarily at TDC and then starts down slowly, accelerating. The bulk of the "push" of about 18,000 pounds (at full power) comes after top dead center, ideally peaking at about 16 degrees after TDC. That puts all that "push" on just one spot on the crankshaft journal.

Forty degrees of crankshaft rotation later, cylinder #2 goes over TDC and starts down, but it's on its intake stroke, so the force is negligible by comparison. Forty degrees after that (80 degrees after #1 did its thing), #3 reaches peak combustion pressure, putting that same 18,000 pounds of force on exactly the same spot on the crankshaft journal where #1 pushed! That one area of about 40 degrees of the crankshaft journal is taking that same enormous pressure nine times every two revolutions. Every power stroke of the engine for its whole life will hammer on that one spot on the crankshaft. (This is a key concept -- don't skip this!)

By comparison, the "flat" engine crankshaft journal only takes that load once every two revolutions. Big difference!

The Key Is Lubrication

In both types of engines, no metal-to-metal contact can ever be allowed to occur -- especially at higher power settings -- or major damage will take place. There is a "bearing insert" between the connecting rod and the crankshaft journal: usually some sort of mix of special alloys, often lead and indium. There will be oil passages inside the crankshaft from end to end that provide oil for all the crankshaft main bearings (those supporting the crankshaft in the case) and the connecting (or master) rod bearings on the crankshaft throws (called "journals"). There will be a large hole strategically located so as to provide exactly the right amount of oil between the bearing surfaces of each journal and its rod, so that no metal ever makes contact with metal. That hole is placed to provide lubrication to the surfaces that get squeezed together by the power stroke -- always on that same spot on the crankshaft. (Forgive me for sounding repetitive, but this is the central concept here.)

(53 Kb)
P&W R-2800 Crankshaft

The concept is crucial for those operating radial engines. A power stroke occurs and is transmitted to the bearing surface on the crankshaft. Eighty degrees later (with a nine-cylinder row), another power stroke occurs from a cylinder 80 degrees around the circle. That same spot on the crankshaft gets hammered again, every 80 degrees. No significant force is ever applied at any other point. Lubrication at that precise location is all-important: the single most-critical point in the whole (radial) engine.

As a side note, an absolutely wonderful (and rather large) book has been written about the development of this crankshaft. It reads like a novel.


(49 Kb)
Oil Gallery Around Crankshaft

Here's my idea of what the cross-section of a crank journal looks like at the oil gallery. Maximum-pressure point from combustion is at the top of the picture, and this is where the film of oil is thinnest. At the bottom there is a much larger space, because we can't make mechanical devices to perfect tolerances. This also explains why there cannot be a lubrication hole there, because all the oil would simply squirt out, making it impossible to maintain oil pressure.


Here are two close-ups of this freshly overhauled crankshaft on the test bench at Precision Engines in Paine, Wash.

(46 Kb) (44 Kb)
Crankshaft Lube Hole Crankshaft No Lube Hole

Note that in the left picture, the lubrication oil hole is showing, with the crankshaft just past top dead center as viewed by the reader. That's the key to this whole (hole?) issue. A very significant flow of oil comes out that hole, squirting oil in between the surfaces. The film of oil squeezes down to a thin film under the heaviest pressure.

The right picture is of the back side of that same journal: obviously, no hole, no lubrication. Under normal loads, the "backside" bearing surface and the journal surface never come close, so no oil is needed. Sure, some probably gets smeared around in there, for oil gets all over everything; but not enough to lubricate, and not enough to cool.

Now, The Crunch

Now, picture what happens when the prop is driving the engine. The loading reverses, and now the crankshaft is driving the pistons, with the crankshaft loads on the side with no hole. Some of this is probably tolerable, but it's very dependent on how much push is involved, and how long it lasts. A few seconds, no problem. A few minutes, who knows? If it becomes "too much," the bearing heats up, and the oil temperature starts rising rapidly. The bearing becomes red hot, melts, metal-to-metal contact gets worse, the end of the connecting rod and the crankshaft journal become white-hot, and something fails. It's usually a catastrophic event, a "master rod bearing failure," or "throwing a master rod."

The old engine designers were smart. They designed these engines with healthy margins for adverse conditions and ham-fisted pilots. But there are some things that can make life worse for them. For example, if the oil gets too hot, loses viscosity, and becomes less effective, or if the bearings don't get enough oil, these bearing surfaces may get hot. If the RPM is high, they are closer to the design limits. If the "push force" is "too much" from a "too high" power setting, the limits are closer. Start adding these factors together, and it's hard on the engine. On the other hand, even at full takeoff power, redline MP and RPM, and oil temperature in the normal range (or even up to redline), everything is just fine, well within limits. The whole engine is massively overbuilt to take this for thousands of hours, with occasional periods of abuse, perhaps even momentary heavy abuse.

The History

The legend is that one airline was having a series of master rod bearing failures in their fleet (my faint memory says "DC-6") while other operators of the same equipment were not having such problems. The only obvious difference was in the descent techniques. One was maintaining cruise altitude until very late (keeping high true airspeed), then doing a high-speed, low-power descent with high RPM. This is very effective at getting down in a hurry, and a standard technique for an emergency descent. But it does result in the prop driving the engine very badly, and hammering on the "back side" of that master rod bearing.

There is more to it. The equations for lift are ultimately driven by true airspeed (TAS), not indicated (IAS). In the same way, the aerodynamics of propellers are related to TAS (of the airplane and the prop blades). So, several things came together for this early DC-6 operator. For the first time, pilots with something of a cowboy attitude had a pressurized airplane that could descend like a manhole cover without bursting passengers' ears. For the first time, they were starting descents at high altitude, where TAS was much higher than they were used to. It was a lot of fun to honk the throttles back, shove the RPM up, and stick the nose down.

(Yes, WWII aircraft did all this, but no one was tracking overhaul times, engines were "free" and plentiful, and no one was interested in economic matters.)

Once the engineers got wind of the problem, the solution was easy. Ease the RPM back along with the throttles on descent, keep a bit of power on to keep the engine loaded a little (engine driving the prop) and start down sooner, making a shallower descent. Passengers liked it better, too. Problem solved, and master rod bearing failures dropped to the same level as everywhere else.

A Question

The big question is, "Just where is that magic point where the engine stops driving the prop, and the prop starts driving the engine?"

Aye, now there's a question! There are lots of little rules of thumb, such as, "Maintain at least 100 BMEP," or, "Keep the MP equal to (or 5" more than) the RPM (in hundreds)." Fact is, none of these mean a lot.

The best way to get a sense of this is to fly a DC-3 with the cockpit window open, and monkey with the power settings while listening to the airflow through (or around) the prop just outside the window. It's very distinctive, but unfortunately, very few will get this chance. I can tell you that at normal pattern speeds (90 knots), the transition point occurs when the MP is down around 16" at 2050 RPM. At cruise speeds (130 or so), the old rule of thumb "one inch per one hundred RPM" is approximately the transition point. In a descent with cruise power (160?), it is around 25". At least one old operator in bygone days had a pretty good rule: "Five inches over RPM at high speed, one inch per one hundred at normal speeds, and in the pattern, we don't care." Personally, I think that's overly conservative, but at least it has a minimal impact on operations. That bearing can certainly take some load, but how much is an unknown.

Frankly, I'm unconvinced that it's even an issue at all on the old bird: The airspeeds are just too low, and the RPM is too low to do damage. Remember, the problem has only been "proven" on airplanes like the DC-6, descending at 300 knots (TAS) and better, for long periods of time, with the MP way back.

There are even reports of old-time operators who practiced emergency descents for real: full RPM, throttles closed, and stuff the nose down to high speeds, even redline speeds. Several old-timers have told me they used to do that, with no known harmful effect, and good records of few master rod bearing failures.

Several radial operators today take this to insane lengths, to the point where safety is compromised. They want 1" MP per 100 RPM minimum at all times, even crossing the runway threshold, if they can. C'mon folks, you gotta land sometime, and somewhere you gotta cut the power! They fly around with gear down and partial flaps when level, and full flaps when descending for the landing, and they make very low, shallow approaches so they can keep the power "square or better." Some even deliberately go below the glide slope (electronic or visual) to accomplish this, which is completely illegal and often dangerous. Airline pilots with a lifetime of flying jets fall into this trap very easily, because it fits with "The Jet Transport Stabilized Approach." High drag and high thrust is indeed a better way to fly jet transports, but it is not a good way to fly the old recips. There are huge differences.

This is probably the syndrome that brought down the Heinkel. I had personally observed them making flat approaches, with gear and flaps down. I've had several emails indicating that others had seen that, and overheard one of the pilots make the comment about keeping the engines at higher power, so as not to allow the prop to drive the engine. Under normal conditions, that's not good in that airplane, as it suffers from very poor performance even on two engines. Far better to make a much steeper descent, with less drag, so that an engine failure is of little consequence. The accident occurred at a density altitude of about 10,000 feet -- even more reason to keep the drag down, and the descent angle up.

Flat Sixes

You may be able to see where I'm going here. If the "problem" was only on the old DC-6 at high speed, high RPM, and very low MP, for long periods at a time, and many times over the life of the engine (even if a shortened life), and if it's not really a problem on the DC-3 at the low speeds and modest RPMs, then how can it be a problem on a "flat six," with that crankshaft journal taking only 1/9 the number of power strokes that a big radial journal does?

Some Hard Evidence

For data, look at Bob Hoover's incredible routine. He goes from full, wide-open power to feather in a few seconds, then restarts and goes to full-bore again. He comes down out of some of those maneuvers with full RPM and closed throttles at very high (over redline) speeds. He has told me personally that he routinely goes to TBO on his engines. He's not the only one: All the air show performers do this sort of thing routinely, with some pretty high-performance engines. Some of them go to TBO and beyond, a few overhaul their engines every season. None, to my knowledge, has ever noticed a problem with connecting rod bearings.

So much for shock cooling and "under-square" operations. As we say in the seminars, "It's not how hard you run your engine, it's how you run your engine hard!"

For completeness, there are a few other issues that always come up in this argument:

Ring Flutter

During most of the time the engine runs, the cylinder rings are forced "down" against their lands, which also pushes them against the cylinder walls. If the RPM is very high, and the MP is very low, there is a large, negative pressure created in the combustion chamber during the intake stroke, due to the closed throttle plate and the piston trying hard to suck air in. This may lift the ring off its land during that stroke. The next stroke is the compression stroke, and while the pressure will be greatly reduced because not much air got in, it's still enough to push the ring back down again. This repetition may well cause the rings to "flutter," beating up and down within the land, and this may well cause damage.

Reciprocating Loads

Radials (and to some lesser extent, flat engines) are constantly throwing large weights in and out. Under power, these forces are pretty well balanced. As power is reduced (and RPM increased), these forces become less balanced, and create stresses within the engine. These may cause vibration and fatigue, but there isn't much data on this.

Counter-Balancing and "De-Tuning"

Crankshafts live a very hard life. Each power stroke twists the crank one way, and a few degrees later, that force drops to zero or less, and then there's another power stroke. Simply by having a crank journal, unbalanced forces are created, so there are always "counter weights" directly opposite the journal to balance this. But the twisting forces impart "torsional vibration," and the harmonics from these can be devastating. So holes are bored in the counterweights, and shock-mounted weights are installed. By careful design and incredibly complex calculations (done originally on slide rules!), these will move around in their mounts, and "tune" the crank's twisting motions. By running at low power (and high RPM), or by changing RPM very quickly, these forces are thrown out of whack, and may actually create adverse twisting moments of their own.

Gearing

On engines with geared props (all the big radials, and some "flats"), the loads on the gearing are carefully designed to take the normal loads of the engine driving the prop. Ever play with gears? Anyone but me remember "Erector Sets"? If you turn a little gear by hand, driving a big gear, the little gear turns quickly and is easy to move, but the big gear turns slowly. Compare this to an engine turning a prop at some reduced speed. But if you turn it around, and turn the big gear by hand, it's much harder; the forces on the teeth are greater to produce the same RPM. This is analogous to the prop driving the engine through a reduction gear. Are the gear teeth designed to take this "reverse force?" Probably. How much? I don't know. Is there a lubrication issue?

Thrust Bearing

In both radials and "flats," the prop is literally trying to rip the prop shaft out by the roots. This force is absorbed by a "thrust bearing" in the nose of the engine. It's usually a conical bearing surface, with all the lubrication issues of all bearings. Allow the prop to drive the engine, and that force reverses, possibly creating a lubrication problem on the back side of that assembly. I really don't know if there's any issue at all here, but it sounds interesting.

In Conclusion ...

... Don't worry so much! Operate any of these engines in a reasonably gentle, normal, non-abusive manner; keep RPM commensurate with power, control the temperatures (CHT), and you'll do just fine. Don't hesitate to pull that RPM well back at low power, it won't hurt a thing, and may well reduce stress and strain on the engine.

Be careful up there!