sillydriver

I love mysteries and controversies, so I felt compelled to do an objective quantitative analysis of the infamous Car and Driver SL600 review. My starting point was blueSL’s clever calculation of the amount of power that actually went into the kinetic energy of the car. The kinetic energy is mass times velocity squared over two, while the average power needed to produce that is just the kinetic energy divided by the time it took to put it there, i.e. the time it takes to accelerate to the final velocity. Once you calculate that power figure you can relate it to the rated horsepower of the engine. Once you analyze this relationship for a large number of road tests you can see where the SL600 test fits in. The advantage off looking at the data this way is that the calculation of the average power needed to produce the car’s kinetic energy is exact: it’s not a matter of guesswork or applying heuristic fudge factors, it’s right from an elementary physics text.

sillydriver

Traction

One of the key variables in an acceleration test is traction. Zero to sixty times vary wildly because of driver skill, wheel spin and road surface conditions. To reduce these variables I decided to look at acceleration north of 60 MPH. All US road tests include zero to sixty times, quarter mile times and quarter mile trap speeds. It’s possible to calculate the amount of power needed to take the kinetic energy of the car from sixty up to its trap speed (120 in the CD test) over the time interval between the zero to sixty time and the quarter mile time (which is 11.9 – 3.6 = 8.3 seconds in the SL600 test). The tires should be fully hooked up during that interval so traction would not be an issue. Another advantage is that the engine is working consistently at high revs at speeds over sixty as the car runs through the gears. At very low speeds without wheel spin the engine would be working at low RPMs in first gear, producing much less than its rated power. Looking at speeds over sixty avoids this problem.

sillydriver

Data

I pulled together a bunch or road test data from Road and Track magazine. I have more back issues of R&T than Car and Driver, and R&T explicitly gives the test weight of the car which is critical to the calculation. I compiled data on 22 cars ranging from the Beetle Turbo S to the Ferrari Enzo. In order to check whether Car and Driver had results significantly different from Road and Track, I pulled out four reviews from the few issues of Car and Driver I had around the house and compared them with full tests of the same models at R&T. The four cars were the Gallardo, the GT2, the Viper SRT-10 and the Cadillac CTS-V. On average the zero to sixty times were one tenth of a second faster in CD than RT, as were the quarter mile times. In other words, the times from sixty MPH to the end of the quarter mile were the same, which is what you’d expect since CD’s more severe launches aren’t a factor above sixty.

sillydriver

Initial Findings

There was a remarkably good relationship between rated engine power and the power going into kinetic energy in these tests, specifically 58% +/- 3% of rated engine power ended up in the energy of motion of these cars. The balance would be lost in driveline friction, air resistance, tire flex, the angular momentum of the wheels, etc. In addition, the engines only produce peak power at one point in their RPM band. The cars within half a percent of 58% include the Enzo, the Ford GT, Maserati Spyder, New Pontiac GTO and Impreza WRX. Quite a varied group, demonstrating the stability of the relationship.

What about the SL600? The car put 74% of its rated 493 HP into kinetic energy above 60 MPH, an impossible result five standard deviations above normal. Very fishy, but explainable as we will see later on.

sillydriver

Torque

I ran regressions to look for factors that explained the 3% deviation around the average 58% efficiency for converting rated engine power into kinetic energy. The level of engine power itself didn’t matter, weight didn’t matter much and surprisingly, the kind of transmission didn’t matter. The one thing that did matter was whether the engine’s torque curve was fat or flat. The average power delivered by an engine as it sweeps through the gears is the time integral of the engine’s torque, which is related to the area under the torque curve. An engine with a humped torque curve, where maximum torque is much greater than the torque delivered at the point of maximum power output, is actually delivering more power on average. You can calculate this excess of maximum torque over torque at maximum power. For example, the torque curves of the turbocharged Beetle Turbo S and Bentley continental GT are electronically regulated to be dead flat so the excess is zero. On the other hand the torque curve of the Mitsubishi Lancer Evo is hugely humped with maximum torque 25% higher than torque at maximum horsepower (which is calculated by taking horsepower, dividing by RPM and multiplying by 5252). The Mercedes ttV12 also has a humped torque curve, the second highest excess being the Maybach 57’s at 22%. When you do a regression of kinetic energy efficiency against torque excess the result is 0.3 times the excess plus 54%. Hence the predicted efficiency of converting rated power into kinetic energy for the Maybach is 61%, which is above the average 58%. In fact the Maybach test result implies 64% efficiency (due to a 13.4 second quarter for a 6300 pound car!) so the calculation implies that its engine is slightly underrated. Specifically, the Maybach engine would have to have 570 peak horsepower, not 548, to product the 61% efficiency that’s predicted from its torque curve.

sillydriver

The SL600

So what about the CD test? The car’s rated 493 HP implies an efficiency of 74% which is absurd. If you put the Maybach engine’s rated 543 HP into the equation the required efficiency is 67% which is still impossible. However, if you use the 570 HP implied by R&T’s Maybach test you get 64%, which is still high but only one standard deviation more than what you’d expect given the torque curve.

Now here’s the trick. Take the specs of the Stage 2 ECU conversion from Renntech’s website: 585 HP and 690 lb-ft. The calculated efficiency from the CD test becomes 61%. The predicted value given the engine’s torque curve is also 61%.

I think CD was lent a car modified in just this way. The other key piece of evidence is the comment in the review about the “three-mode manumatic function” which the stock SL600 doesn’t have but the Renntech engine computer does. The review also talks about “quick throttle response” which I have heard is lacking in the stock car.

Finally, the timeslips posted on the Renntech website for 625 HP Stage 3 modified cars show 11.6 quarter mile times at 121 MPH. If we assume the zero to sixty time drops a tenth from the CD review, the Stage 3 conversions efficiency also calculates out to 61%, exactly what you’d expect. I used to think the HP figures claimed by Renntech an other tuners for the ttV12 were absolutely ridiculous, but the data seems to bear them out remarkably well.

My 600 is supposed to show up Tuesday – it’s off to the tuners for me after the break-in is done!

Trimmer2

Very interesting reading and seemingly quite concise. It's refreshing to read facts from people who seem to have done their homework as opposed to the uninformed emotional opinions that seem to fly around here.

shoes

Great analysis although none of us are probably smart enough to argue with you.

I have been staring at the C&D numbers from the SL600 since the magazine came out in complete disbelief and shock. It was only two weeks ago that I drove an SL600 and found it lethargic and unwilling.

Actually, I use a much less scientific testing methodology. You see, my dealer keeps most of his cars in a garage about a block and a half away from the showroom. The only opportunity the salesmen have to drive the cars is when they run down to this garage. Invariably, they drive the cars as fast as they can between these two locations and they give me a report of their perception of how fast the cars are. The same guy always brings me cars and he always drives like he is in a funny car drag race. His report after bringing me the SL600 was that it did not seem as fast as the SL500. When I drove the two cars back to back, I agreed with him, that the new 7 speed in the SL500 made the car feel faster. Sure, once the transmission hooked up in the SL600, it was a beast, but it did not seem as willing to dance as either the SL500 or, better still, the SL55.

I hope you will give us your observations once your SL600 arrives since you owned an SL55 and should be a good judge.

sillydriver

Thanks for the kind words, both of you. Trimmer2, now that you've had them for a while, any trouble or issues with the mods? I plan on going ahead but I don't want to break anything. Not sure whether I should get stage 3 or stage 2.

RU_MATRX

This is great information and coming from an engineering degree years ago, I understand the basics of what you are stating with time-integration for the area under the torque curve and KE=mv (where's that hat?) squared. I know that most of us were expressing opinions based on real-life experiences and comparative analysis on a much lower layperson's level. Your precise calculations along with standard deviations would be what one finds in a scientific automotive journal: The best way to analyze measured/claimed figures.

What one sees in real life is best explained through a scientific method.

Thanks again for your time and efforts in explaining what the numbers really mean in quantitative terms!

Trimmer2

To tell you the truth, I don't remember how the car felt before the mods. My neighbor just took delivery of a SL600 a few weeks ago so I might take his for a spin to remind myself before he goes in for some mods.

I haven't had any problems with my mods from RennTech. Although, I'm what some call a "spirited" driver I have no intention of taking the car to a track to push it 100%+. Day to day driving is just plain a blast. It took a few weeks to get my eyes out far enough in front of the car to get comfortable with the speed. Most of my first time riders frequently speak of feeling their stomach get light as in a carnival ride...and all ride with a huge smile on their face!

RennTech recomends getting the transmission computer (ECU 3) because the stock computer has a hard time interpreting the additional torque and tends to shift up and down trying to find the right gear. The SL65 computer from RennTech is very smooth and holds gears longer and downshifts noticably and automatically. It also adds the "manual mode" which holds gears up through redline and allows the rev limiter to do its work. In the Sport mode, it will automatically shift up wether you use the shifter to shift up or not when approaching redline.

In fact, my dealer has sent a number of SL600's up to RennTech to get mods after seeing/driving my car and have sold them off the floor for a nice premium within a couple of days of their return.

I wouldn't hesitate doing the upgrade. For me, it was a great bang for the buck! Next mod might have to be more rubber on the road with bigger/wider wheels...

sillydriver

Sounds great. If dealers are putting the modified ECU into the SL600s they sell, then the theory that Car and Driver ended up with a modified car becomes all the more plausible.

sillydriver

AMG493, I’m not going to try zero to sixty until after the break-in but the throttle response and lazy shifting makes the car feel slower than the 55 at low speeds. It only begins to feel faster floored in third gear, i.e. north of 70. I stand by the analysis I posted in this forum several days ago: CD tested a modified car.

AREITU, my track driving experiences (SL55 around Lime Rock, Viper on an autocross course) have taught me that driver skill is the key commodity. I don’t have enough of it and I won’t lend my baby to someone who does. I probably will eventually go with the mild modification idea.

I still love the car – I commuted in with it this morning. I’ve got to admit Shoes was right in his post a few days ago: the car feels “unwilling.” I had to floor it once on the way in just to make sure there was a huge engine under the hood. You feel launched from a catapult when you do that but there’s no trace of that ferocity when driving normally. Frankly, it’s an easy car to drive slowly, unlike the 55 which always seemed to be screaming “get out of my way or I’ll kill you” at the other drivers. Or at least that’s what I heard it scream.