I have compiled a listing of the 'best' normally aspirated engines in the world from the Civic Si to the Ferrari 550 Maranello. I'm currently seeking info on Lambo's V-12 to add to the list. I'm looking at bmep, torque/liter, bhp/liter, mean piston speed, bore:stroke ratio and compression ratio such that one can compare one engine to another even though one may be a four cylinder and the other a large displacement V-12.

BMEP: brake mean effective pressure, is the measure of the amount of work is done on the face of the piston throughout one complete cycle:

Civic Si: 171.8 psi
GS-R: 176 psi
ITR: 180.2 psi
S2K: 189.3 psi
GT-S: 183 psi
RSX Type R: 188 psi
E46 M3: 199.5 psi
911 GT3: 187.4 psi
360 Modena: 189.5 psi
BMW M5: 183.7 psi
550 Maranello: 189.2 psi

Bore:Stroke ratio of the Civic Si: 1.05, ITR: 0.93, S2K: 1.04, GT-S: 0.96, RS-X Type R: 1, M3: 0.96, 996 GT3: 1.35, 360 M: 1.08, M5: 1.06, 550 Maranello: 1.17.

Mean Piston Speed increases with decreasing bore:stroke ratio. While that is good at low to mid engine speeds [it facilitates turbulence which greatly assists combustion] at high speeds the stresses are quite large and premium components [$$$$$] must be used. These numbers are calculated at the redline rpm.

Si: 4042 fps, 8,000 rpm; GS-R: 4577 fps, 8,000 rpm; ITR: 4806 fps, 8,400 rpm; S2K: 4960 fps, 9,000 rpm; GT-S: 4350 fps, 7,800 rpm; RS-X Type R: 4740 fps, 8,400 rpm; M3: 4776 fps, 8,000 rpm; 996 GT3: 3787 fps, 7,800 rpm; 360 Modena: 4406 fps, 8500 rpm; M5: 4087 fps, 7000 rpm; 550 Maranello: 3691 fps, 7,500 rpm.

Torque/liter is perhaps a better measure than bhp/liter with bmep being perhaps better still. Si: 69.6, GS-R: 71.2, ITR: 72.9, S2K: 76.6, GT-S: 74, RS-X Type R: 76.1. But look at the Euros, M3: 80.7, 996 GT3: 75.8, 360 Modena: 76.7, M5: 183.7, 550 Maranello: 76.5

Then we have compression ratios. Si: 10.2, GS-R: 10.0, ITR: 10.6:1, S2K: 11.1:1, GT-S: 11.5:1, RS-X Type R: 11.5:1, M3: 11.5:1, 996 GT3: 11.7:1, 360 Modena: 11.0:1, M5: 11.0:1, 550 Maranello: 10.8:1

Interesting. If one puts them into chart form some trends become obvious. Higher compression ratios, higher peak engine operating speeds and higher torque/liter. Notice that the highest torque per liter also equates to the highest bmep, which is perhaps the best methodology of comparing engines.






[Edited by Chui on 06-02-2001 at 12:35 PM]

 

Interesting numbers and a lot of work on your part...thanks for posting that, Chui!

 

Fan-friggin'-tastic! Thank you.

 

Wow, I thought I was screwing off at work!

You put some time into that. Very cool.

 

Well, Chui does work for an auto company, so maybe he can write off the hours as a competitive analysis :-)

Lamborghini Diablo 6.0 VT

Displacement: 5997 cc
Cylinders: 12
Bore/Stroke: 87mmx84mm
Compression Ratio: 10.7:1
HP: 543 hp@7100 rpm
Torque: 457 lbs-ft@5800 rpm
Redline 7500 rpm

You can feed the numbers into your spread sheet and determine BMEP, piston speed, etc. And the M5 is 74.7 lbs-ft/liter, looks like you had a typo there (I wish I could buy a NA car with 183 lbs-ft/liter :-)

Also, I think you'll find that the prime determinant of BMEP (and therefore torque as you pointed out - which makes perfect sense since torque is _essentially_ force on the piston x the crank throw) is compression ratio. I'll bet you could get a pretty good best fit line on that with the data you have. Makes sense of course since that's the predominant factor in thermodynamic efficiency if you get sufficient airflow through the intake tract. In fact, if you weighted torque/liter by price, I'll bet the fit gets even tighter to compression ratio ;-).

UL

BTW, are they still using Internal Combustion Engines and Air Pollution by Obert in mechanical engineering programs these days?

 

[WARNING] not only is this post long, but it also uses some foul language. As if that was not enough, high school physics are also mentioned [/WARNING]


The way performance potential of motor engines is presented above does not tell the whole story.

For some reason, all that manufacturers, magazines, and people in general seem to care about, is maximum power and torque and little else. IMHO, maximum figures mean sh1t.

I said it before and I'll say it again: what is important in an engine is what's under the torque curve in its usable revolution range!


Let's take this assertion apart:


First, let's define what is "under the torque curve".

You plot a torque curve by drawing a line that represents available torque from 0 to the max rev range (redline). Now, torque is a moment of force (expressed as force times a distance) while rev is a frequency (expressed as time to the power of minus 1, or 1/t).

The area drawn by that curve (its integral) can thus be computed as force times a distance divided by time. Coincidentally (or certainly not!) this is the definition of Power. If you want to stick ISO units into this, you get Nm/s, or Watts.

So, power, not torque, is king!

But consider this: by calculating the integral of the torque curve you are not calculating maximum but average power.

This is a number that no one talks about, although this is the most important number. Go figure!


Second, let's discuss "usable rpm range".

This is rarely "0 to redline". It varies depending on application. And, as a result, so will the "area under the curve".

If your engine is used to power a racing car with close short gearing, then the usable range will be a couple of krpm on both sides of max torque (the "power band"). If you use it in a streetable car then it will be much wider. If you use it in a coffee grinder it will be close to zero!

So here I renew my plea for manufacturers to quote average power thrughout the rev range, and average power in the "power band", instead of quoting maximum torque and maximum power, which basically means faq all.


We would be much better informed as a result.


Coming back to the starting post, it would be interesting to calculate the specific average power for those engines. I'm sure you'd find a few surprises...

[Edited by Luis on 06-05-2001 at 11:23 AM]

 

Luis, I don't think anyone will disagree with you that horsepower and area under the curve are most important. But I don't expect manufacturers will ever do what you asked, simply because trying to establish standards for such an expression of output would be difficult at best. Its simply easier to compare peak power and peak torque. I wouldn't mind seeing accurate dyno charts from car companies though!

In terms of the rest of the discussion, torque/liter, which ties closely to BMEP is one of the best measures of combustion efficiency that I know for a NA engine (I really like to see the total pressure curve including before TDC, but that won't ever be given to us) HP/liter is a better measure of how well you can sustain that efficiency to high rpms. Which brings up a good point - if we look at hp and torque numbers that manufacturers give us, we get a pretty good idea of the overall shape of the torque curve in the critical rpm ranges.

For example:

E46 M3 - 333 hp@7900 rpm, 262 lbs-ft@4900 rpm.

We can either convert the peak torque number to a hp number, or vice-versa, doesn't matter. I like to look at torque curves, so, 333 hp@7900 rpm = 221 lbs-ft. So, from peak torque to peak power the E46 loses 15.6% of it's peak torque. However, it might be better to look at the differential as a function of overall spread (5.2%/1k rpm). Either way, that's pretty darn good.

Another example:

Honda S2000 - 240 hp@8300 rpm, 153 lbs-ft @7500 rpm

Torque at 8300 rpm = 151 lbs-ft. So we only lose about 1.5%, but the spread is much narrower (2%/1k rpm). Luckily we have lots of dyno sheets that show we're only down 2-3% even at 6000 rpm, so it's really better than it looks.

Example:

Ferrari 550 Maranello - 485 hp@7000, 419 lbs-ft@5000

Torque at 7000 rpm = 364 lbs-ft, or about a 13.2% loss(6.6%/1k rpm)

Example:

Ferrari 355 - 375 hp@8250, 266 lbs-ft@5800

Torque at 8250 rpm = 239 lbs-ft, or a 10.1% loss, (4%/1k rpm).

What Honda does better than anyone else is sustain torque output to very high rpms. But, while torque output at lower rpms is decent from a displacement standpoint, Honda has historically made some sacrifices compared to other engine builders (see E46 low end vs. S2000 low end) that typically show up in slightly lower BMEP numbers in return for sustainability. The introduction of iVTEC will mitigate that sacrifice, but not eliminate it, since you still have a fixed cylinder head, exhaust manifold and intake manifold (on the really high output engines) that are still optimized for high rpm operation. It is useful to note that the torque curve for the new RSX-S engine (K20C) is completely flat from 6000 to 7400 rpm, a larger spread than what we've seen in the past from Honda's 100+ hp/liter engines. iVTEC and a variable geometry intake manifold certainly play a role in this. The J32 in the CL-S, OTOH, is a bit non-traditional for Honda in terms of specific output and torque curve (similar BMEP, but trades high rpm output for a broad low rpm torque curve), but it has to pull a much heavier car than previous models.

UL

 

I'm not sure if I fully agree with you here. There are two simple measures that would be much more meaningful than what we get today.

One is overall average power thru the idle to redline range. Easy, I think no one would dispute this one.

The other is average power in the "power band" range. This last one may seem difficult, but it isn't really.

If you are seeking maximum acceleration there is always a given rpm at which you should upshift. That is the point where torque at the wheel falls below what you get after upshifting. I suppose this is also consensual. It can vary from gear to gear but in a given gear it is fixed. So, to find out the average power for this case just average the average power across all gears shifting at optimal points. Once you've got the torque curve and gearing, calculating this number is child's play.

But let's say for the sake of discussion that this last number is open to misrepresentation.

Then, I contend that max power and overall average power as defined on top is a much better view of engine capabilities than the current max torque crap. Which, without taking gearing into consideration is basically meaningless (just see the 320Nm of the latest incarnation of the volkswagen 1.9TDI. Better than an M3, ha?).

[Edited by Luis on 06-05-2001 at 04:37 PM]

 

O.k., lets take average power from idle to redline. I would argue that such a number is about as meaningless as giving peak power and peak torque. Below 3000 rpm the S2K gets pretty soft, but no one drives there, so it doesn't matter. To list average power would mislead people about performance (underestimate it).

Now, _if_ we were to agree upon that (and I understand we may not), then that would say we should choose the powerband argument. I like your suggestion of averaging across gears (I guess it would be evenly weighted for each gear), but then you would favor cars with tighter gearboxes and more gears allowing you to stay closer to the power peak. If you were looking for an accurate measure of accelerative ability, this might be a legitimate option and I give you credit for suggesting it. But I still don't think it will happen for a couple of reasons. One, the consumer is simple, the eval criteria is complex, so no go. Two, hp ratings are far more than measures of accelerative ability, it's about bragging rights and that's part of selling cars.

Maybe one of the car magazines should take your idea and start dynoing and rating cars this way. I'd certainly like to see it.

UL

 

[QUOTE]Originally posted by ultimate lurker
[B]O.k., lets take average power from idle to redline.