Torque vs Horsepower: Auto Racing Power Showdown

[fedorfan]

Ive got a question, which is more powerful and better to have in auto racing, torque or horsepower? I've gotten mixed answers and wanted to ask some people who know their stuff.

 

[Moridin]

Torque is force times distance, while horsepower is a unit of power. Which would you rather have in which types of races?

 

[fedorfan]

Which one will make you go faster, a lot of torque or a lot of hp?

 

[chemisttree]

Enzo Ferrari was once quoted as saying "Horsepower sells cars, torque wins races". Obviously Mr. Ferrari believed that torque was most important of the two provided that both were present in sufficient amounts. With lots of turns, acceleration, and braking, torque wins hands down IF there is sufficient horsepower. Without the horsepower, diesel trucks would win auto races. Without torque, jet-powered cars would win all the races. By the way, I believe a jet powered vehicle holds the land speed record... of course it just went straight and probably didn't handle well in the turns!

 

[brewnog]

To ask which is more powerful is like asking "which is faster, speed or weight?".

The answer to which is more advantageous in racing is not a clear one. Moridin alluded to the fact that different types of racing will require different balances of torque and horsepower. To complicate matters, different engine configurations will develop different torque and horsepower characteristics across a range of engine speeds.

So an engine setup which develops lots of low-down torque may be preferable for a race with lots of tight corners, acceleration, and gear changes, and an engine with a lot of high-end power may be more suited to a race on a high-speed ring type track.

The answer is not a clear one.

 

[cesiumfrog]

Got to say, torque seems like an odd criteria for an engine (regardless of race style), since a car always has a gearbox.. but despite this it tends to always be listed among the "vital stats" of any supercar.. Does the torque just provide insight into the engine characteristics at various rpm, or am I misunderstanding something else?

 

[russ_watters]

Torque is reported at the engine, not at the wheels, so the gearbox is irrelevant to the reported value. Sure, you can always lower the gear ratio to increase the torque at the wheels, but if you redline at 2mph, you haven't helped yourself any unless your goal is to power the space shuttle mobile launch platform...

 

[cesiumfrog]

Sure, but in the ideal world (where no energy is lost in the gearbox), if two engines (with different torque) output equal horsepower, than both should accelerate at the exact same rate (even if one does have to shift gears at 2kph) and reach the same top speed (if both cars have the same aerodynamics). Right?

 

[turbo]

Sure, but in the ideal world (where no energy is lost in the gearbox), if two engines (with different torque) output equal horsepower, than both should accelerate at the exact same rate (even if one does have to shift gears at 2kph) and reach the same top speed (if both cars have the same aerodynamics). Right?

No, the engine that is able to to deliver more torque will be able to kick your butt to the finish line faster.

 

[russ_watters]

Sure, but in the ideal world (where no energy is lost in the gearbox), if two engines (with different torque) output equal horsepower, than both should accelerate at the exact same rate (even if one does have to shift gears at 2kph) and reach the same top speed (if both cars have the same aerodynamics). Right?

No. Every time you shift gears the torque delivered to the wheels drops, so the acceleration drops. Those two cars will only have the same acceleration for about a tenth of a second.

And the most important gearbox loss here (which I didn't mention before) isn't about friction, it is about inertia - since being in a lower gear means the rpm increases faster for the same acceleration, there is more dynamic loss in the drivetrain (due to rotational inertia), so even if the torque at constant rpm is the same, the acceleration would be much, much lower. Even in the ideal case, you still have to account for it because it has a huge impact.

I think my mom's '68 camaro was a 3-speed. Your first gear is like 2nd or 3rd gear today. The torque is so high on those old v-8s, they would still tear off the line.

 

[cesiumfrog]

Even in the ideal case, you still have to account for it because it has a huge impact.

What you're saying seems to be that I'm correct (noting the ideal situation I specified, think massless continuous variable transmission); the only difference is the losses that occur between the engine and the wheels (obviously from real life experience, time lost on the clutch early-on is a major factor).

This is interesting because it seems to demonstrate horsepower governs maximum speed (circular track), but torque influences acceleration (drag racing).

Despite my experience with and without lightened clutch-plate flywheels, I'm not convinced a slightly-torquier engine spins much of the drive train significantly slower; it seems (rather than the rotational inertia of the drive train) the biggest problem (in the real world) must be the number of gear-changes required (and their timing, ie. how fast before the first one). I get the impression that torquier engines will tend to remain useful over a much wider range of rev's. Is it that all engines redline in around the same place, so torque is a direct measure of the range of useful revs (ie. how many gearchanges before a particular speed)? Or would the difference still be pronouced if the engine with more torque also redlined sooner and so changed gears at all the same speeds (since obviously low-rev power importantly decreases clutch-riding waste, especially at low speeds)?

 

[joema]

This is a frequently-debated topic in automotive performance circles.

Given a conventional geared transmission, supposedly best momentary acceleration is at peak wheel torque. This comes from F=MA, arranged as A=F/M. Acceleration is highest when force (torque) is highest.

Based on this viewpoint, optimal average acceleration would happen if an infinitely-variable, lossless continuously variable transmission kept the engine at torque peak throughout the acceleration run.

A contrary viewpoint is best average or sustained acceleration is obtained by operating the engine at peak power. You want the highest power-to-weight ratio. From this viewpoint, a hypothetical perfect CVT would keep the engine at peak power rpm, not peak torque.

Of course there's no such thing as a perfect CVT, but considering that helps determine what's theoretically possible, hence the underlying physics.

I don't know the definitive answer, just mentioning the two viewpoints.

 

[russ_watters]

What you're saying seems to be that I'm correct (noting the ideal situation I specified, think massless continuous variable transmission); the only difference is the losses that occur between the engine and the wheels (obviously from real life experience, time lost on the clutch early-on is a major factor).

Uh, ok... I guess if your incorrect scenario were correct, you'd be correct.

But your incorrect scenario is incorrect for more than one reason.

-It is incorrect because it doesn't consider the mass of the drivetrain and leaving that mass off is just as big of a sin as leaving the mass of the car out. An "ideal situation" never leaves out such an important factor.

-It is also incorrect because while torque at the wheels is the torque at the engine multiplied by the gear ratio, speed is the rpm times the gear ratio, times the circumference of the wheels. Since the shift point (whether you use a cvt or a standard transmission is irrelevant) is at the redline of the engine, in order to get the same acceleration out of an engine with half the torque, you'd need to be able to rev it to twice the rpm.

Apply some quick math:

Engine a:
Torque: X
Max RPM: 5000
Gear Ratio: 4:1
Wheel Circumference: 4 feet

Enging b:
Torque: 1/2X
Gear Ratio: 8:1
Max RPM: 5000
Wheel Circumference: 4 feet

The torque at the wheels for each is X*4 = 4x, 1/2X*8=4X.

But the speed of the car with engine A is 5000/4*4=5,000fpm when it needs to shift,

While the speed of the car with engine B is 5000/8*4=2,500fpm when it needs to shift.

They reach 2,500 fpm at the same time, but after that, the car will accelerate much faster with engine A.

This is interesting because it seems to demonstrate horsepower governs maximum speed (circular track), but torque influences acceleration (drag racing).

That much is true.

 

[fedorfan]

Ive gotten torque=acceleration and hp=top speed. Well, I want to ask something, if you have a lot of torque all over the powerband especially up on the top end does that mean while one car is topped out on the top end you`re still accelerating?

 

[Stingray]

This is a frequently-debated topic in automotive performance circles.

That's certainly true. It comes up so much in some other forums I frequent that I've thought about writing up a canned response. But I haven't done that yet, so here's another try:

People often confuse power and torque because car enthusiasts tend to (unknowingly) use these words for different concepts. This is a physics site, so I'm going to go ahead and use the definitions from physics.

The full-throttle behavior of an engine can be approximately modeled as a device which has some function \tau(\omega) associated with it. This fixes the torque it can produce as a function of engine speed (rpm). This function is not at all constant, although engineers often strive to make it as flat as possible.

Regardless, given the torque function, there is an associated power P(\omega) = \omega \tau(\omega). So if the torque is known at all speeds, the power is known at all speeds (and vice versa). You can't have one without the other.

Despite this, it is common practice for engines to be advertised only in terms of their peak torque and peak power. The engine speeds where those conditions may be found are also usually given. The peak power is very important for reasons I'll get to later, but the peak torque is essentially useless all by itself. The reason is that the gearbox can multiply the torque to (essentially) any amount whatsoever at an appropriate speed. But an ideal gearbox cannot change the power.

Staying with the ideal case, the maximum forward force that a car can produce is entirely determined by the power its engine is producing and the car's overall speed. So fixing speed, maximum acceleration is always reached by maximizing the engine's power output. It is the job of the transmission (and driver) to use the gearbox to keep the revs as close to the engine's power peak as possible if full acceleration is desired.

Modern transmissions have many closely-spaced ratios, so except at very low speeds (at the bottom of 1st gear), an engine may be kept close to its power peak for as long as desired. That means that a well-designed car that is driven well may produce a force F \sim P_{\rm{peak}}/v. This depends only on the peak power (and velocity), and explains why the power-to-weight ratio is such a good predictor of acceleration performance.

Having said that, the torque peak is not completely irrelevant. Its position relative to the power peak is usually a good indicator of the size of the car's "powerband." Essentially, how high do you have to rev it in a given gear before the engine really gets going? Having a wide powerband is extremely important in everyday (or moderately aggressive) driving where you're not going to redline in every gear. It makes the car feel much more powerful even if the maximum performance is the same. Of course, a wide powerband is also useful if your have a poor transmission or don't want to shift as much.

Russ, differences in drivetrain inertia between reasonable designs are not usually not a huge effect. They're certainly significant, but I don't think I'd include them given the approximations already inherent in this sort of discussion.

 

[JeffKoch]

Ive gotten torque=acceleration and hp=top speed. Well, I want to ask something, if you have a lot of torque all over the powerband especially up on the top end does that mean while one car is topped out on the top end you`re still accelerating?

Power is proportional to torque at any RPM, so the original question doesn't actually make sense. I think what you mean is, is it better to have a broad power band ("torquey") or a narrow one ("peaky"), assuming the peaky motor puts out more maximum power. If your shifts took zero time, if you had as many gears as you wanted, and if could keep the engine operating at the same high rpm all the time, then a narrow peaky 2-stroke-style power band would be quicker if it puts out more peak power. Factor in limited human abilities, and a broad power band becomes more useful to more people even if peak power is down - so more people would go faster.

 

[5.0stang]

Power to weight comes in effect.

You have to ask, at what rpm are you going to be "moving" at?

Even in a drag race, (let's say a 6,000 rpm limit) you are below 3,000 rpm (intial launch) for a short period of time (probably 5% for example). The upper rpm horsepower is crucial here. You can have "gobs" of torque below 3,000 rpm...but if you have trouble breathing up top - no horsepower output...the car with less torque and more horsepower (equivalent vice versa situation) is going to take you. The 95% of the time, he is in his "sweet spot."

I like to say torque gets you moving, horsepower keeps you moving.

The fastest production cars, EVO's, McLaren F1's, all have more peak (higher revving) horsepower than torque. They post the quickest acceleration numbers to 60 and the fastest top speeds.

As far as a street car goes, I pick horsepower.

Diesel engines (production trucks) have lots of torque but lower powerbands and lots of weight.

 

[TGarzarella]

Stingray and 5.0stang got it right.

TQ and HP are proportional to each other, so it makes no sense to speak of them as if they're independent. The more appropriate question, as was pointed out, is to ask what is the optimal distribution or powerband (TQ vs RPM, HP vs RPM) for auto racing.

I think the answer will really depend on the type of car and the type of auto racing. In 1/4 mile drag racing, for example, low-rpm horsepower/torque will be useless (unless you're launching from idle speed) since the engine will be spinning above 4000-5000 rpms for most of the race. Some AWD race cars can get away with powerbands that have all "top end" TQ/HP and nothing down low, because they can keep their engine speed up in the high rpms from the launch to the finish line.

In road racing, where there are a lot of turns, some low end power/torque is important when coming out of a turn. The gearing of the car also comes into play because that determines the how much power/torque that you have at a given speed.

So this becomes a complicated issue with no single convenient answer. This is why car racers spend so much time experimenting, modifying things, etc. trying to find their optimal set-up.

 

[joema]

...TQ and HP are proportional to each other, so it makes no sense to speak of them as if they're independent...

The problem is on many engines the hp rpm peak is above the torque rpm peak.

This raises the question, is optimal acceleration at hp peak or torque peak. Also we must define "acceleration". Do we mean momentary acceleration, as from 60-61 mph. Or do we mean sustained average acceleration, as in lowest time from 0-60 mph or 1/4 mile? Also transmission type is a factor: traditional geared transmissions vs CVT.

I agree from a F=MA standpoint, it appears best acceleration is at torque peak. However power peak is by definition producing more power, hence better power-to-weight ratio. From that standpoint it seems best sustained acceleration with a perfect CVT would be at power peak, not torque peak.

Real world transmissions tend to obscure the underlying issue. However over time geared transmissions are more closely approximating a CVT. E.g, the new Lexus LS 460 has an 8-speed (!) automatic.

 

[Stingray]

The problem is on many engines the hp rpm peak is above the torque rpm peak.

This raises the question, is optimal acceleration at hp peak or torque peak. Also we must define "acceleration". Do we mean momentary acceleration, as from 60-61 mph. Or do we mean sustained average acceleration, as in lowest time from 0-60 mph or 1/4 mile? Also transmission type is a factor: traditional geared transmissions vs CVT.

Just to clarify, all engines have their hp peak after their torque peak (at least if there's only peak for each curve). It's an amusing little math exercise to prove that.

Anyway, the answer to your second question is as follows. In a given gear, the instantaneous acceleration is highest at the torque peak. At a given speed, a vehicle allowed to select any gear ratio will have the highest instantaneous acceleration at the power peak. If that sounds contradictory, try reading it a couple of times. The first condition has you choose a gear and vary speed. In the second, you fix speed and vary the gear ratio.

So say you were at the torque peak in some gear. Even though you're accelerating as hard that gear will ever allow you to, you might be able to downshift and accelerate even more. A perfect CVT optimizing straightline performance would keep you at the power peak at all times.

Also, if you always maximize instantaneous acceleration, you'll also maximize average acceleration. So there's no need to worry about that (at least at the level of this discussion).

 

[fedorfan]

So hypothetically, if you had an engine that had unlimited revs or unlimited gears would it have infinite top speed?

 

[turbo]

Even in a drag race, (let's say a 6,000 rpm limit) you are below 3,000 rpm (intial launch) for a short period of time (probably 5% for example). The upper rpm horsepower is crucial here. You can have "gobs" of torque below 3,000 rpm...but if you have trouble breathing up top - no horsepower output...the car with less torque and more horsepower (equivalent vice versa situation) is going to take you. The 95% of the time, he is in his "sweet spot."

You might want to visit a few drag tracks and buy a pit pass so you can rub elbows with the racers. One of the most incredible performers in the NE in the early 80's was a Canadian driving a Camaro called Mouse Trick. He ran pro-built small-block Chevy engines and he said that he was dumping the clutch at 10,000 rpm off the line. I don't know if that was true, but I have no reason to doubt him. He had reliability problems because of this "pushing the envelope", and I recall him and his crew at Pembroke, NB one year scouring the camped-out drag crews for clutch parts, so that he could advance to the Sunday eliminations. My buddy was running a 340 Duster (eventually the US champion in his class for a couple of years) and he was dumping the clutch at 6000 rpm - only after a lot of efforts lightning the rods, grinding off piston skirts, etc. 340 Mopar parts are a lot more massive than small-block Chevy parts and he had a lot of work to do to overcome that differential.

 

[brewnog]

So hypothetically, if you had an engine that had unlimited revs or unlimited gears would it have infinite top speed?

No, absolutely not. The top speed will be limited by:

power = force * velocity

Where power is limited by the engine's ability to get fuel and air into the combustion chambers and produce work as a result, and force is the total of the aerodynamic and resistive forces acting against the vehicle's motion.

 

[fedorfan]

What, why doesn't infinite revs=infinite speed? I am talking like there's no wind or air or anything and no universal speed limit like the speed of light.

 

[cesiumfrog]

What, why doesn't infinite revs=infinite speed? I am talking like there's no wind or air or anything and no universal speed limit like the speed of light.

But without that friction (and with limitless gearing) every engine could have "infinite speed", trivially (regardless of power etc).

 

[fedorfan]

Thanks you that's what I was wondering.

 

[brewnog]

It's a senseless question. If, mechanically, you could spin an engine to 'infinite' rpms, you still wouldn't be able to get enough fuel or air in it to create 'infinite' power. In fact, that amount of air doesn't even exist. It's not even worth thinking about it.

 

[5.0stang]

You might want to visit a few drag tracks and buy a pit pass so you can rub elbows with the racers. One of the most incredible performers in the NE in the early 80's was a Canadian driving a Camaro called Mouse Trick. He ran pro-built small-block Chevy engines and he said that he was dumping the clutch at 10,000 rpm off the line. I don't know if that was true, but I have no reason to doubt him. He had reliability problems because of this "pushing the envelope", and I recall him and his crew at Pembroke, NB one year scouring the camped-out drag crews for clutch parts, so that he could advance to the Sunday eliminations. My buddy was running a 340 Duster (eventually the US champion in his class for a couple of years) and he was dumping the clutch at 6000 rpm - only after a lot of efforts lightning the rods, grinding off piston skirts, etc. 340 Mopar parts are a lot more massive than small-block Chevy parts and he had a lot of work to do to overcome that differential.

It appears you misunderstood what I was getting at.

I did not say that all cars have a limit of 6,000 rpm.

I did not say that all cars launch at max rpm.

I was referring to street cars (limited traction - no slicks/suspension). They simply can not launch at 6,000 rpm (again, just an example) to get optimum time and mph. They have to launch a little soft. Say...2,400 rpm pedaling it.

Then again, even after the soft launch the rpm picks up quickly and it is in the upper rpms that keep the car accelerating.

Torque gets it moving, Horsepower keeps the gains going.

That is how I look at it.

So again, I do not know what you were getting at. I'm quite familiar with the track

 

[turbo]

So again, I do not know what you were getting at. I'm quite familiar with the track

My point is that if you go to the track and ask the successful drag racers how they're tuning their cars, you will find that they want their torque (powerband) optimized for the higher rpms and they want the powerband wide enough to cover their shift points. They launch pretty hot and flog 'em down the stretch and they want to continue to accelerate strongly in each gear, so the torque has to be available over a range of RPMs that exceeds the RPM drop caused by shifting to each higher gear. Peak horsepower is a "nice to know" number, but to continuously accelerate, you need to deliver that power effectively over a usable range of RPMs. That's the importance of the torque vs RPM curve on a dynamometer plot. If the absolute value of your car's torque is relatively high and the high portion of the curve is wide enough to cover your shift points (gearbox dependent), you will outperform a competitor with higher absolute torque values if his curve does not adequately span the RPM differential at his shift points. His acceleration will not be consistent because his torque curve is narrow and peaky and since acceleration adds cumulatively to velocity, your car, with a slightly lower but broader torque vs RPM curve will win out.

 

[fedorfan]

Not infinite power, I am saying if you have endless gears or an endless revs then your car wouldn`t stop accelerating. Not acknowledging wind resistance, gravity resistance, and whatever else restricting it. Its just a stupid little meaningless waste of 2 minutes and feel free not to answer it. Thank you if you do or did.