The (mostly) correct way to measure specific power

"An RB26 can get 380HP per liter if tuned!" "Nascar engines take four times the displacement to get the same power as F1 engines!" "All big, fat american V8s are inefficient" "There's no replacement for displacement!" Please stop. Let me explain why you are wrong.

A recent comment (not a bad one! A good comment!) made me think about how we cro-magnons beat and wail about our favorite engines and why its the best. I'm guilty of it too, often using the 4v DOHC 160HP 2-liter Nissan S20 engine, from 1969, as an amazing example of ahead-of-its-time tech. But I'm wrong too.

Left: The average Honda or Mazda enthusiast during a HP/liter argument. Large bone club and posse of barbarians not pictured.

Engine efficiency has nothing to do with RPM, Displacement, or boost. It has to do with all of them. So if you compare two engines, you have to compare all of that. I'm going to show you how to make an accurate, apples-to-apples comparison between engines. This only works if the use the same fuel, however, and you only care about peak power. (if you can find numbers for low throttle driving, you can compare by fuel economy)

The only accurate specific power measurement is an amount of air taken in an amount of time, providing a proportional HP level. If you need more air to make the same peak HP in over the same amount of time with the same fuel, then you are losing efficiency somewhere. That is why air flow (l/s, CFM, whatever you measure it by) is how most racing parts (such as heads, exhaust, intake, etc.) are measured.

So here's how you can accurately measure specific power, to end internet debates once and for all:

1. Start with the displacement. This is the amount of area swept by the cylinder per cycle. If it is a 2 stroke or a rotary, you double it, because for the same as a 4-stroke cycle, a 2-stroke can run twice.

2. Multiply by the number of atmospheres that the engine takes in. This is boost, measured in Bar. Since N/A engines are in 1 bar already (14.5 psi at sea level) any additional boost is added onto it. So an engine at 15 psi runs at 1 bar boost, or 2 bars of air pressure. So multiply the displacement by 2 if that is the case.

3. Multiply that by the peak RPM that the engine is getting this air. We know how much air per cycle is flowing, now we want to know how much per minute. A more accurate way (instead of doubling displacement) to measure rotaries or two-strokes is to double their RPM (as it should be called cycles per minute), but either way works.

4. Divide the resulting number by the HP produced at that RPM.

The lower the number, the less liters (or ci if you use that) air is needed per minute for each unit of HP produced over that minute.

For our basis of comparison, let's use the Ford 5.0 Coyote:

5.0l liters, 1 atmosphere of pressure, 412HP @ 6500 RPM
5 x 1 x 6500 = 32500 / 412 = 78.8 liters per minute per HP.

So let's run some examples:

Nissan RB26DETT: 2.6liters x (1+.7 bar) x 6800 = 30056 / 276 hp = 128 per minute per HP

WOAH WOAH WOAH. What the hell happened? Is the RB26 really that much less efficient than this big, heavy, American V8?

No. It's because the published numbers, as many people know, about the RB26DETT are false. It produced about 330 HP, not 276, at 10 psi, and was designed to produce 450HP at 15psi with the stock, restrictive exhaust and intake system modified. So let's do that again:

2.6 x 1.7 x 6800 = 30056 / 330 = 91 per HP
2.6 x 2 x 6800 = 35360 / 450 = 78 per HP

So as you can see, the race-trim RB26DETT, 20 years ago, was about as efficient as the new Ford 5.0. This shows that while our power production isn't any better (both are DOHC 4v) our reliability is way up and our costs way down compared to the race engine.

What about some of those theoretical, 1000HP builds? Well, let's go with the JUN GT-R, which I believe had 1000HP at about 2 bar of boost and 9000rpm. (they would rev to 10,000, but the peak power came on earlier)

2.6 x 3 x 9000 = 70200 / 1000 = 70.2

That's a pretty damn efficient engine for the power output, but what about the new Alpha Omega GT-R from AMS? 1700 HP is a lot...

3.8 x 3.8 (40psi) x 8000 / 1700 = 115520 / 1700 = 67.95

Despite being almost 15 years newer, with what they have to work with, the record-breaking Alpha Omega only manages 3 HP per liter of air per minute better than the JUN R32. This makes sense, since there is virtually no new technology that one utilizes over the other. Small marginal improvements across the board, but nothing big.

Best of all, this counts for fuel economy as well, if you calculate the HP, airflow, and RPM at varying throttle levels. You could find out which engine gets the best fuel economy by measuring how HP is necessary to move the car as fast as you need, and then try to make that HP with the least possible airflow. The reason hybrids are able to get so good fuel economy is that the loss of engine efficiency is made up for in gains with electric motors.

Here are some more numbers I worked out, for your enjoyment. See if you can find the lowest HP/liter/minute engine out there:

NOTE: High numbers means either a really inefficient exhaust/intake (so the engine is starved for air) or lots of friction and resistance. Low numbers mean the opposite. Engines that use E85, Nitromethane, and Biofuel are proportionally lower, and (depending on energy-per-volume and air/fuel ratio, are going to show up as amazing)

Nissan S20: 2 x 1 x 8000 = 16000 / 160 = 100 per HP
1930s Deusenberg Model J: 6.8 x 1 x 3500 (est.) = 23800 / 265 = 90 per HP
1930s Deusenberg Model SSJ: 6.8 x 2.5 x 3500 (est.) = 47600 / 400 = 119 per HP
Nissan L28: 2.8 x 1 x 5600 / 150 = 15680 / 150 = 104 per HP
Nissan L28ET: 2.8 x 1.5 x 5600 = 130.6 per HP
Ford Flathead V8: 3.9 x 1 x 2000 = 7800 / 110 = 70.9
413 Wedge: 6.8 x 1 x 4800 = 32640 / 390 = 83.7 per HP (it was probably higher, SAE Net VS Gross)
2008 GT-R: 3.8 x 1.7bar x 6400rpm = 41344 / 485 = 85 per HP
2013 GT-R: 3.8 x 2bar x 6400rpm = 48640 / 550 = 88.4 per HP (sacrificing efficiency for more power!)

2013 Subaru BRZ: 2 x 1 x 7000 = 14000 / 200 = 70 per HP (!!)
1960's Mustang 315 Racecar: 4.7 x 1 x 6000rpm = 28200 / 315 = 89 per HP
Porsche 917 Can Am engine: 5.4 x 3.7(!!!) x 7800 = 155844 / 1580 = 98 per HP
Honda S2000: 2 x 1 x 9800 = 19600 / 247 = 79 per HP
EK Type R: 1.6 x 1 x 8200 = 13120 / 185 = 70 per HP
Top fuel Dragracer: 8.2 x 4.5 x 8400 = 309960 / 8000 = 38 per HP (Nitromethane, baby!)
Ferrari 458 V8: 4.5 x 1 x 9000 = 40500 / 562 = 72.64 per HP
McLaren F1: 6.1 x 1 x 7400 = 45140 / 618 = 73 per HP
Bugatti Veyron: 8 x 2.1 x 6000 / 1001 = 100800 / 1001 = 100.7 per HP
Bugatti Veyron SS: 8 x 2.25 x 6400 = 115200 / 1200 = 96 per HP
Hennessey Venom GT: 7 x 2.3 x 6600 = 106260 / 1244 = 85.4 per HP
Koenigsegg Agera (gas): 5 x 1.3 x 7500 = 51750 / 960 = 55 per HP
Koenigsegg Agera R (biofuel): 5 x 1.4 x 7500 / 1140 = 47 per HP
1950's Offenhauser 4-cylinder: 4.1 x 1 x 6600 = 2760/ 420 = 64.4 per HP
1963+ Indycar Offenhauser: 2.65 x 4 x 9000 = 95400 / 770 = 123.8 per HP
1961 Ferrari F1 engine: 1.5 x 1 x 9500 = 14250 / 190 = 75 per HP
1967 Ferrari 312/67 F1: 3 x 1 x 10000 = 30000 / 390 = 77 per HP
1970 Lotus 72 (Cosworth DFV): 3 x 1 x 11000 = 33000 / 520 = 63 per HP
1984 Toleman TG184 (Senna): 1.45 x 3 (est.) x 12000 = 52200 / 750 =70 per HP
1984 Toleman TG184 (Senna) no restrictor: 1.45 x 5 x 12000 = 90000 / 1300 =67 per HP
1988 MP4/4 (Senna w/Honda engine): 1.5 x 3.5 x 12000 = 63000 / 850 = 74 per HP
1990 MP4/5 (Senna): 3.5 x 1 x 13000 = 45500 / 650 = 70 per HP
2002 BMW P82 (F1): 3 x 1 x 19200 = 57600 / 900 = 64 per HP
2006 Toyota F1 Engine: 2.4 x 1 x 19000 = 45600 / 750 = 60 per HP
2007-2013 Renault RS27: 2.4 x 1 x 18000 = 432 / 750 = 57.6 per HP
2014 F1 engine: 1.6 x 2.5 x 15000 = 78750 / 600 = 131 per HP
My Dad's old 413 Wedge: 7.5 x 1 x 8500 = 63750 / 850 = 75 per HP (nitromethane)

Some Rotaries:

13B REW: 1.3 x 2 x 1.7 x 6000 rpm (est.) = 26520 / 300 = 94.7 per HP
13B Renesis: 1.3 x 2 x 1 x 9000 = 23400 / 250 = 93 per HP
2.6 x 2 x 1.7 x 10500 = 92820/ 930 = 99 per HP

Note: F1 numbers a little sketchy, I had to use a variety of sources and the peak RPMs may not off by 1000 either way, so take them with a +/- 3 liters per minute per HP grain of salt. Efficiency has been about the same since the 70s, with fluctuation based mostly on new fuel and reliability requirements.

How come the Flathead V8 is more efficient than a V-TEC engine, but Nissan's Flathead is horribly inefficient? Reciprocating weight and friction. At 2000rpms, the Flathead (also lacking a lot of peripherals) had so much less friction than a OHC engine. A vast majority of engine power (Koenigsegg estimates about 30%) is lost turning camshafts, and that gets increased exponentially if you have more of them. The Flathead only has to turn a single cam. A DOHC design only becomes more efficient when you add valves or increase the RPMs, since the small friction of the added parts on a OHV engine is much less until higher RPMs.

So wait, then why the heck is the Agera R so absurdly more efficient than every other production engine in existence!? The Agera runs on E85, and the Agera R runs on Biofuel. Both have a lot lower energy density, but a better air/fuel ratio. This means they can make more power out of less air, but have to use lots more fuel to do that. Same goes for Nitromethane. The power gains from E85 (~15-20%) are the same as the specific power drop (80% of 70 = gasp, 56!) on the Agera. Nitromethane (30-40%) is also the same drop with a top fuel dragster.

The Agera R, similarly, has so many parts made of carbon fiber, that it weighs considerably less than most engines. The Veyron Engine has a ton of added weight, because it was designed from two V8s, and the doubled parts double the weight. It has 2 more turbos to push as well.

What about my Dad's old 413 wedge (bored to 450-460ci)? While he ran Nitromethane, without boost and with all that heavy weight, he only came in about 13 liters /minute/HP lower than stock. That's how much efficiency is lost to peak at 850HP out of such an old, heavy engine.

And the F1 engines? Oh boy. Here comes the hate. In 2014 engines increased in weight by almost 80% (95kg to 145kg) and have a 30% tighter fuel usage requirement. The only way to meet those efficiency requirements is to vastly underpower the engine, letting it use very little fuel with less throttle than possible. This makes its peak power efficiency really, really bad. Granted, if you consider the KERS system, it is reduced to 100 liters per minute per HP.

Now we can stop arguing about which engines have the best specific power, and be aware that things like low-throttle fuel economy (vs peak throttle fuel economy), reliability, boost pressure, and peak RPM have more to do with efficiency then straight up displacement.

At the end of the day, yes, there is a "replacement" for displacement. It's call air mass, and the more of it you have, the better. Cylinder displacement, RPM, and Boost all increase it proportionally.