Why It’s Almost Impossible to Rev to 21,000 RPM

With an engine revving at 20,000 RPM this piston has a mass of 2.5 tonnes. It accelerates from 0-60 miles per hour in 0.003 of a second. And it pulls 10,000 G.

And this is what 20,000 RPM sounds like - it’s Mark Webber driving a 2006 Williams with a Cosworth CA2006 and it’s the highest revving F1 engine ever.

It’s been almost 20 years since this engine was released, so why haven’t F1 engines - or any engines for that matter - gone above this magical 20,000 rpm?

Well, it’s more difficult than you might think, so today I’m going to explain why it’s almost impossible to engines to rev over 21,000 RPM.

More from Driver61:
What If F1 ENGINES Had No Rules?

What If Formula 1 Had No Rules?

The Porsche Faster Than F1

How This Car Does 0-100 In 0.9sec

👉 Follow Driver61 on:

👉 Follow Scott on:

🛞 My Sim Kit:

#F1Engine #Piston #Revs

[bythelee]

NONE of the thermal or stress issues are a deal breaker. Having worked in F1, analysing stresses in pistons and conrods and crankshafts and valves, I'm quite sure everything could easily run much faster - 30, 000rpm would be no problem, particularly if exotic materials are allowed (like aluminium - beryllium). Turbochargers in humble road cars can hit 100k rpm while red hot...
The killer factor is time. Higher RPM means less time per cycle, but the physics of fuel combustion are "fixed", in that flame fronts take finite time to propagate. The flame-shooting injection is a great idea - an advance on Alfa Romeo's "Twinspark" idea that used two spark plugs to start two flame fronts in the combustion chamber.

The second problem is keeping everything synchronised. While the components can be designed to survive the loads and stresses, keeping the valve timing synchronised with the piston motion is harder than it might seem. (Quite apart from the problem of "when to start combustion" to get an optimal power output from the cylinder (fire too early, combustion works against you. Fire too late, most of the power goes out the exhaust. Ignition advance is a huge field all to itself. But I digress.)
Because, nothing is rigid, and everything is flexible. And twistable.
V12s and V16s have much longer crankshafts and camshafts, and will twist along their length, resulting in some very serious timing errors in the cylinders remote from where the shafts are linked together. Various efforts to deal with this include using link gears at both ends of the engine, or putting them in the middle. Which brings their own problems with "forced" positional control, that can induce their own breaking loads. This can all be engineered around, not least via novel valve technologies.
But to even be aware that everything is flexible, and avoid the typical "rigid body" thinking that seemed to afflict everyone else working in the field, is a challenge in itself.


For those that doubt if I know what I'm talking about, take a closer look at the bending crankshaft animation at 8:03. That is NOT showing "stress" (so, we don't even have to debate whether it is Maximum Principle Stress, or Von Mises stress, or some value of fatigue alternating stress - ALL of which MIGHT be the one dominating "failure".)

It is showing a colour contour map of DISPLACEMENT. Because, the entire "red" part of the crank does NOT have similar stress - that would be concentrated at the fillet between crank web and big end... certainly not distributed equally over the whole section.
Worse than that, the deformation is created by moving the centre main bearing while holding the mains at each end - a totally false load case that simply never happens in reality. At least, not to the extent it is shown here. Some movement is possible because these bearings are oil film (journal) bearings, and the oil film thickness allows a small amount of movement. But this animation is not intended to show that, since displacement at the centre WOULD cause movement in the end bearings too - the crankshaft stiffness forces that.


One of the reasons I gave up doing what I did, was the disparaging remarks about being "the pretty pictures department" by folk that did not bother to look at my track record of "ZERO in-service failure" of any part I ever worked on. Sigh.
But sadly, even 20 years later in 2024, it seems not only has nothing changed, but the "pretty picture" is not even a valid picture at all, but something created purely to look pretty. Sigh.


TMI but for those still interested, "zero in-service failure" means inifinite life while in use in the engine. NOT "unbreakable parts" (you just have to subject them to loads they never see in service, to get them to break).
This included redesigning finger followers that broke in 10 minutes, and conrods that snapped in an hour. All of these became "infinite life" parts that could last forever, just by fettling design details. And, making them lighter for better performance at the same time. For me, durability usually meant removing redundant material, and sometimes redistributing material, to better disperse loads (and hence stresses) throughout the structure.
I never had to fix anything by adding any material to it. Which is contrary to intuition, and is the "go-to" response of about 99% of others doing this kind of work. "Beef up the weak bit" is NOT a good thing to hear.

Just like the crankshaft that drew envy because it NEVER broke, not only lasting an entire race season, but also having the smallest main bearings of any crankshaft on the racetrack. (and smaller mains meant less parasitic torque was drained from the power output, giving more BHP at the flywheel...)
But smaller bearings mean a huge increase in stress concentration, making it much more difficult to endure the loads any crankshaft has to survive.
Anybody else that even tried to match our bearing size, never finished one race.

When you get a handle on engineering, and materials, and stresses and strains, and loads, and resonance, and vibration, and heat transfer and temperatures, and... then designing engine parts for conditions thought impossible, becomes quite possible.

There certainly are limits. Yield strengths of materials, for one. And when you factor in that the strength depends on temperature, too... But rather than get sidetracked further on "strain rate sensivity" and why F1 pistons survive stresses more than four times higher than their yield strength at operational temperature, just know that we are still quite far from reaching those boundaries with designs at 20k rpm.

I mean, 99% of the piston is not even close to failure. It is the 1% detail problem where stresses concentrate that start the cracks that lead to failure.
Same with that crankshaft. The challenge is fettling the details WHERE THEY MATTER.


Sorry for the essay. Thanks for reading.

[lenmetallica]

The fact that this video is still live after all the inaccuracies and wrong information speaks louder than the mistakes themselves.

[TheNecromancer]

The main reasons are actually the speed of the flamewall starts to be too slow, the combination of high compression and ultrashort stroke reaches mechanical limits, and the extremely narrow gap between head and piston at TDC prevents a clean burn. The mechanical load is a problem that is solvable.

[katchF]

Mass doesn't change with velocity, the piston doesn't magically become 2.5 tons. If you're talking engineering you absolutely have to be precise with your terms, man

[Prelude]

The video never actually answered the question. It pointed out the difficulties of running at 20, 000 rpm, but did not actually tell us why it could not go higher.

[dr.hugog.hackenbush]

The only problem is that they DID 20K rpm...
It CAN be done. Honda had a roadrace bike that would rev to 22K rom in the '60s.

[jimpartridge]

I remember being under the grandstand during USGP practice when Williams was very nearly 21, 000 rpm. What a thrilling sound. Menacing, spine chilling, and the Mercedes McLaren engine still sounded more violent than the rest. Miss those V10s

[benjaminshropshire]

An interesting side note: it's not uncommon for the massive diesel engines on cargo ships to be 2-cycle, but that's at least in part because they run slow enough to be able to effectively replace the exhaust with fresh air via blowers (which they have the room for) and that also allows them to not run oil mixed with the fuel (which might not be as big a deal given the fuel they run isn't that much different in weight than motor oil).

[thekinginyellow]

0:10 Mass of... Beg pardon? You need to look up the definition of mass. I'll give you a hint, acceleration doesn't change it.

[dirtygarageguy]

Is this a challenge on how to get as many details wrong as possible in 13 minutes?

[TheHypnotstCollector]

F1 engines were touching 22, 000rpm in c.2006. Saw it on TV. The tach touched 22K from time to time,

[TheSnivilous]

15 seconds in and you're killing me. Mass isn't changing, force is changing. There's no extra "stuff" just because it's accelerating so quickly.

[mitchellsteindler]

You made it out to seem that there was a reason that specifically 20, 000 rpm is the limit.

[syntropy]

Frame front velocity of the fuel is what determines rpm limit. Different fuels have different flame front velocities, and thus different rpm limits for the same engine.

[pauldonnelly]

No mention of the Honda air cooled, conventional valve springs, sixes of their racing bikes in the 60's, that routinely reved to 23k rpm? Relevant as they won at least 2 world championships and remained reliable. Also no mention of the, again Honda, CVICC system of the 70's? Strange because it is exactly the system you describe as new in 2015...

[pjlangford]

My little Honda CBR250RR will rev to 21, 000, it's stopped making power at 19, 000 but I've seen it as high as 23, 000 on the track. And that's 1990 technology and it's still going strong.

[spacecadet]

To quote von Braun "I have become very careful about using the word 'Impossible'."

[JohnJohn-zhov]

A couple of things that I noticed in the video.
Engine friction wasn't mentioned but is a key reason for not increasing engine speed further.
The NA F1 engines used port-fuel injection, direct injection came along in 2014.
Current F1 engines do not use TJI as described in the video, they use a passive pre-chamber as only one injector per cylinder is allowed.
The stoichiometric air-fuel ratio of gasoline isn't 14.7:1, it depends the formulation of the particular fuel.

[grantfuller]

The more power strokes you can squeeze into a minute, ( rpm ) the more power you make -But - the volumetric efficiency gets worse as you try to speed things up ( that whole getting the air and fuel into the cylinder and mixed, thing ) . There’s a point at which at which the gains by higher rpm get overtaken by the loss in volumetric efficiency . Had a tutor who said “ at 20, 000 rpm pistons don’t go up and down - they just vibrate “ 😂

[Impeller_GR]

6:25 Massive blunder here, con-rod length does not influence stroke and therefore the bore/stroke ratio and engine „squareness“, only the crank can influence stroke. Rod length to stroke ratio is a whole other matter.