# Why do most radial engines use an odd number of cylinders?

To help suppress vibration, most 'V' and 'horizontal' engines use an even number of cylinders as closely opposed as possible. Radial engines are well known for excessive vibration. For some reason they nearly always use a staggered odd number of cylinders that would seem to promote vibration, why?

• V engines have even number of cylinders by definition. Jan 4 '18 at 16:04
• @Agent_L, you can get a V5 engine. I've no idea if they're used in aircraft or not but VW use them in cars. Jan 4 '18 at 17:06
• @Holloway VW calls it "VR engine", not "V engine". It's pretty much a staggered inline (R stands for reihen, meaning inline here), not having much in common with V engines except superficially similar cylinder layout. Jan 5 '18 at 8:51
• @Agent_L / Holloway - Its always interesting to learn the exceptions to the rule; just when you think you pretty well know a subject completely. Thanks for the additional info. Jan 5 '18 at 9:09
• By definition, V engines have cylinders arranged in a V. The V designation has absolutely nothing to do with the number of cylinders. Dec 18 '19 at 2:48

An odd number of cylinders is required by the combination of the single-crank radial design, the four-stroke (Otto) work cycle, and the desire to keep the power strokes evenly spaced in time.

To keep the design simple and lightweight, a single-bank radial airplane engine has one crank, which means that the pistons must reach the top of their travel in rotation order. But the four-stroke cycle requires that a piston must reach the top of its travel twice for each power stroke. The only way to promote evenly timed power strokes is to fire every other cylinder in rotation order.

With an even number of cylinders this would require a hesitation or skip in the firing sequence on every rotation as the engine switched between the odd and even cylinders. With an odd number of cylinders the timing is quite naturally smooth. For example, the firing order of an eight-cylinder radial would be

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
*   *   *   *     *   *   *   * *   *   *   *     *   *   *   *


while the firing order of a nine-cylinder radial is

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
*   *   *   *   *   *   *   *   *   *   *   *   *   *


If you could watch a radial airplane engine in slow motion you would see that when a cylinder is in its compression stroke, the cylinders on either side of it are in their exhaust strokes, and when a cylinder is beginning its power stroke, the cylinders on either side of it are near to beginning their intake strokes.

Two-stroke radial engines do not need to have an odd number of cylinders.

• +1, couldn't explain it any better. For an example of two-stroke radials with even cylinder count the Zoche diesels come to mind. And to study the motion of a radial in slow motion I recommend the CAD models in the enginehistory.org site. Caution: The last link is to an animated GIF which takes some time to load. Jan 3 '18 at 22:29
• @jwzumwalt: A single-bank single-crank radial can never fire two cylinders simultaneously. Jan 4 '18 at 4:42
• @Grimm The Opiner: Infinity is not possible. As the number of cylinders approaches infinity, their firing sequence approaches simultaneity. - Interestingly, in addition the size of the aircraft approaches the size of the universe, obviating the need to take off. Jan 4 '18 at 18:09
• @Grimm The Opiner: You may call it a radial. I call my little Briggs&Stratton lawnmower a "one cylinder radial" but that don't make it a simple lightweight airplane engine. Jan 4 '18 at 18:09
• @A.I.Breveleri Well, whatever kind of twin a BMW motorcycle is then. In bike-land they're generally referred to as boxers because the pistons punch "one-two". I love the single cylinder radial! Not least because it's the only radial engine wherein all cylinders fire at once - made possible because it has a crank per cylinder. Jan 5 '18 at 8:20

Surely a 1, 3, 2, 4 firing order (just like an inline-4) seems possible in theory, but [one of] the issues is the ring camshaft.

(...) unless there is an odd number of cylinders, the ring cam around the nose of the engine would be unable to provide the inlet valve open - exhaust valve open sequence required by the four-stroke cycle.

Notice the two rings, each with two opposite steps (lobes).

As one step pushes a rod, the opposite step does the same two cylinders down with a delay. If they were an even number, the ring cam would be operating two opposite cylinders simultaneously due to the shared crankshaft mounting.

In other words, the ring cam would be letting in the air/fuel mixture with one of the pistons (in a pair) going up, not down.

• You could solve the camshaft problem with a geared camshaft, but of course, a 1, 3, 2, 4 firing order would not have an even spacing in time, so the engine would run very rough.Either way, an odd number of cylinders is the way to go, +1. Jan 3 '18 at 17:11
• A single-crank radial with an even cylinder count would be easily serviced by two pairs of ring cams, one for the odd cylinders and one for the evens. Similarly, it would have two sets of magnetos, unless it was modern enough to have electronic ignition modules. Jan 4 '18 at 4:36
• I don’t think the cam has anything to do with it. The cam will simply rotate at half the speed of the crank (or 1/4 the speed with two sets of lobes as in the linked animation) The valves will always be in time with the cylinders. The problem is in evenly spacing the firing order. You simply can’t space them evenly with an even number of cylinders. Jan 8 '18 at 19:24
• Interestingly, cam plates on radials turn at a speed relative to the crank = 1/no. lobes x 2. On a 9 cyl radial with 4 lobes, the cam plate turns against crank rotation, while a 9 cyl with 5 lobes turns in the same direction as the crank. Dec 18 '19 at 2:21

It is possible to have a smooth firing order for an even number of cylinders so long as the number of cylinders is not a power of two. Simplest example is therefore a six-cylinder radial with a firing order of 1-2-5-6-3-4. Power pulses come alternately every 60 and 180 degrees of crankshaft rotation (work it out). Think of it as two superimposed three-cylinder engines sharing the same crankshaft and rotated 60 degrees from each other. A ten-cylinder similarly would have a firing order of 1-2-5-6-9-10-3-4-7-8, with power pulses alternating between every 36 and 108 degrees This would actually be quite smooth as the power pulses are not instantaneous and sequentially firing cylinders would have considerable overlap in their power strokes, especially with a 10 cylinder design.

A V-engine analog would be a 12 cylinder with a bank angle of other then 60 degrees. Since a straight six has perfect primary and secondary balance, so does a V-12 at any bank angle. Only bank angles that are multiples of 60 give all equally spaced power pulses, but with other bank angles, the engine is still perfectly balanced and smooth, the power pulses just come at gaps that alternate back and forth between two values just like a six, ten or twelve cylinder radial would.

Such a six cylinder radial would have an interesting sound at idle, no doubt. It would have a boomboom-boomboom-boomboom rumble, similar to but faster than the "burble" of a cross-plane V-8, but there would be no uneven "hiccup" as the gaps between both sequential cylinders and cylinder pairs are evenly spaced although different. At speed, I suspect a 10 cylinder would not sound any different than a 5 cylinder, the sound of sequentially-firing cylinder pairs blurring together.

Engines with a power-of-two number of cylinders truly can't be built to run smooth, as they can't be timed as a set of superimposed engines with an odd number of cylinders since powers-of-two have no odd divisors. For them, there will always be a gap in sequential power stokes once every two crankshaft rotations that doesn't match any other gap, a "hiccup" that could probably cause a lot of vibration.

• +1 Interesting theory -- timing an even-numbered engine as a set of superimposed odd-numbered engines. This would probably be practical for 6, 10, 14, or even 18 cylinders. A 12-cylinder, however, must be timed as four 3-cylinders, which makes the groups of power strokes 4 adjacent cylinders long, with only 3 groups per rotation. This would probably be noticeably rough. - The problem of not enough groups also appears at 20 (four 5-cyl) cylinders, 24 (eight 3-cyl) cylinders, and other impractically large sizes. Dec 19 '21 at 20:17

Look at valve timing figures for a average four stroke and bearing loads this will give a clue as to the firing order and to make sure that only one cylinder would fire at a time and avoid bearing overload. As to vibration you have to have equal reciprocating forces in opposite directions but in the radial this cannot be achieved without some sort of balancing shaft this would take valuable power and increase weight obviously

first of all not all V engines have even number of cylinders. Honda RC211V uses a v5 (yepp, 3 cyl on one bank and 2 on the other) engine. i just do not have enough rep yet to answer there. now, odd number of cylinders in a radial engine is required for 4 stroke engines only.it was well described why. but 2 stroke ones may not have odd numbers. allso if double ended pistons are used (rare engine concept) then it would be possible to fire 2 pistons at the same time.

• Honda has a penchant for odd-number V-engines. The NS500 race bike had a V3 engine (2 up, 1 laid down) and also the MVX250 road bike which had the orientation reversed (1 up, 2 laid down). Dec 18 '19 at 4:07

A even number of cylinders in one bank of a radial engine just does not work, purely for the fact that each cylinder has to fire once in every two rotations of the engine, as their is only one crank it is impossible to achieve a firing sequence that would work.

The firing sequence on a nine cylinder engine is 1 3 5 7 9 2 4 6 8, this provides a smooth production of power