The Ram's Eye - A Driver's Blog: 2016


Thursday, 22 December 2016

Chevrolet 1LE & Grand Sport - How do they do it? Part 3

In Parts 1 and 2 (Links: +Chevrolet 1LE & Grand Sport - How do they do it? Part 1 & Part 2), I concluded that grip is where Chevys excel and decided to try and figure out how they do that by looking at test data from Car and Driver's Lightning Lap features. The first thing that stood out to me when the 5th generation Camaro 1LE came out was the wider tires compared to the Mustang Track Pack of the time and even the Boss 302. The tires on the ZL1 and Z/28 stood out as much.. only on those, they stood out compared to just about anything that isn't a supercar. So I decided to start looking there; tire sizes.

To evaluate tire sizes, I calculated a weight-to-tire-section ratio for each car. Similar to the idea of power to weight ratio, where the number tells you how much weight each hp is burdened with, this tells you how much weight each mm of tire section is burdened with, so to speak. For example, a BMW M235i weighs 3,490 lb, as tested during the LL feature. It has 225/40/18 front tires and 245/35/18 rear tires. The total available tire width footprint, pressure, tread, and tire deformation notwithstanding, is 940 mm (225 x 2 + 245 x 2), which gives it a weight-to-tire ratio of 3.71 lb/mm. Here is where the trust in measured data that I hopefully established in Part 2 comes into play. Without a monumental amount of work going into looking up tire pressure specs, tire wall and tread stiffness spec (IF manufacturers share in the first place), and calculating actual contact patch areas for each car, we can look at measured data (i.e. lap times) and see if we find a correlation strictly between tire size and lap times. Here's a plot of that weight to tire section ratio for each car vs that car's lap time.

You may have read or heard about how wider tires don't increase contact patch size and are only useful for managing thermal stresses and changing the shape of the contact patch. Whether it's due to the contact patch size increasing, better thermal management and patch shape, neither, or both is irrelevant here so I won't get into the theory. What's relevant is the real world result: cars tend to go faster on wider tires. It's unarguable that quicker cars have more tire width for every pound of weight. I couldn't find an advantage in specs in terms of power, weight, downforce, etc. but could tire size be a factor? Below is a table listing different weight to tire width ratios for Camaros and Vettes and a few other cars. If you arrange cars tested over the last three Lightning Lap features in terms of weight/tire mm, the Corvette GS is 4th best, Z06 is 8th best, and 5th gen Z/28 and 6th gen SS 1LE are 14th and 15th, and that's out of 50 cars (excluding special features like race and police cars).

Car lb/mm
Viper ACR 2.62
SRT Viper TA 2.64
911 GT3 RS 2.67
Corvette GS 2.78
Alfa Romeo 4C 2.81
Cayman GT4 2.83
Corvette Z06 2.86
Shelby GT350R 3.00
911 GT3 3.02
Ferrari 488GTB 3.12
McLaren 570S 3.13
Audi TTS 3.16
(5th gen) Camaro Z/28 3.17
Camaro SS 1LE 3.17

Clearly, Chevys measure up really well here. This is something you might have noticed in the past by looking at specs, and Chevy even bragged when it released the Z/28 and said it had the widest front tires (305) fit to a production car. You might have also found a lot of angry anti-Camaro comments on reviews that go along the lines of this: "throw wider tires on [insert car] to match the Camaro and see how it stacks up!" There is reason to the madness. Now, slapping a set of huge tires on a car not set up to handle the grip won't win races without a proper setup and supporting hardware, but the point is that you do need big tires. Chevy's clearly got that base covered. So, is that all, you just need wider tires?

Although they do have good tire-to-weight ratios, you can see from the table above and the one posted in Part 2 that they beat cars (in lateral g forces) with better numbers and, as mentioned in Part 2, they rank above nearly 200 cars in lateral g forces measured in the first corner of the track, which is especially impressive for the current Camaro SS 1LE, since it doesn't use the same type of aggressive tires like the Z/28 and top dog Vettes. If it's just down to this ratio, that shouldn't be the case, so it's likely not the only factor/advantage. What else could it be? Bear with me on this one.. I think it's wheel size. Now, everyone avoids big wheels like the plague. But what if the bigger wheels actually help? Let's first look at why big wheels are bad.

First is weight. Weight is the enemy of speed. Bigger wheels tend to be heavier, so that's more weight you have to accelerate, brake, and turn. Weight is also critical because wheels and tires are unsprung – meaning the car’s springs and entire suspension is downstream of it (relative to the road) and can’t directly respond to forces generated by the road on the wheels. Wheels actually have a suspension system, but it consists of just the tires and air filling them, and those components form the spring and damping properties. Since they aren't nearly as effective as a car's suspension system, especially on low profile tires, they transmit a lot of those forces from the road to the car instead of absorbing and dissipating the energy, and the heavier the wheels, the bigger those forces are. The suspension’s job of keeping tires in contact with the road, then, becomes harder with heavier wheels, resulting in relatively compromised grip and ride, all else being equal.

Then there's moment of inertia and rotational energy. I can't adequately cover these properties here but, putting it simply, moment of inertia is comparable to mass but for rotational motion as opposed to linear. Just as it gets harder to move something the heavier it is, it gets harder to rotate something the heavier it is, but also how far it is from the axis around which you are trying to rotate. If all else is equal, a bigger, heavier wheel results in a higher moment of inertia, making it harder to change its state/speed (i.e. accelerate and brake it). But worse yet, a wider diameter wheel pushes the weight of the wheel's rim/barrel and the tires further out, making that moment of inertia even higher. Moment of inertia, typically referred to as I, is defined as:

I = mass * r²,

where r is the "effective" radius - A distance from the axis of rotation that the entire mass can theoretically be concentrated at with the same result, similar to the idea of a centre of gravity. The heavier something is or the further it is from the axis, the higher the moment of inertia - I - and harder it is to turn. Then there is rotational (kinetic) energy, defined as:

E = 0.5 * I * ω²,

where ω is the rotational speed. That means that the higher the moment of inertia, rotational speed, or both, the more energy there is that you have to deal with. Pretty straight forward.

But what if you could minimize those disadvantages? Chevy has been using forged wheels on their high performance models for a while now. I can't find specs on the 6th gen SS 1LE wheels yet, which may be even lighter, but their 20" x 11" wheel on the 5th gen weighs approximately 28 lb. That's definitely not light, but if you do some research, you'll find that most good aftermarket cast aluminum 18" x 10" wheels weigh low-to-mid 20's lb, unless you get into the more expensive sub 20 lb options. And a cast aluminum would probably have a thicker rim/barrel since it's weaker than forged, so more of the weight is put further out away from the centre, which is worse than just adding weight. In fact, that 20" x 11" wheel is lighter than the 19 x 10" wheel Ford used on the back of the 2012-2013 Boss 302 Laguna Seca, which weighs approx 33 lb, although these were cast. 991 GT3 20" centre-lock wheels weigh 24 lb and 27 lb front and rear (source: Rennlist Forums: 991 GT3 Stock 20 Wheel Weights) and they have to deal with a good 300-400 lb lower curb weight. So, while no featherweight, the weight of those 20 inchers aren't nearly as bad as off-the-shelf 20" wheels that people upsize to for looks and are clearly well designed. 

Rotational energy is a double edged sword. On one hand, a bigger wheel, generally being heavier, negatively affects the weight element of the moment of inertia, making it more resistant to accelerating and braking (i.e. needing more power to do either). On the other hand, a wider diameter wheel and tire package has a longer circumference/perimeter, so one revolution covers a longer distance, which means it can spin slower than a smaller wheel and tire package while the car's speed is unchanged and, therefore, the speed element of the rotational energy goes down. To illustrate, if you compare the BMW M4, which has 255/35/19 front tires, and the M235i, which has 225/40/18 front tires, to the Camaro, you'll find that, because the Camaro's wheels and tires are bigger with a longer circumference, they spin less for the same speed. The Camaro's front tires have a circ. of 7.00 ft, the M4's are 6.81 ft, and the M235i's are 6.57 ft. That means that the M4's tires need to spin 2.79% faster and the M235i's 6.5% faster than the Camaro's to match its speed. And because rotational energy is a function of the square of the rotational speed, the M4 front wheels and tires would have 5.7% more rotational energy than the Camaro's and the M235i's would have 13.4% more energy while all three cars are going at the same speed, if moment of inertia (I) is the same in the E equation above for all three cars.

In other words, more braking power and acceleration (engine) power are needed to brake and accelerate the smaller M4 19" and M235i 18" wheels and tires, compared to the 20" Camaro wheels, assuming an equal moment of inertia (I). I expect the Camaro's wheels to have a higher (worse) moment of inertia, but it has to be at least that much worse for the Camaro's wheels and tires, in comparison, to need the same amount of rotational energy to accelerate and decelerate, let alone more energy. A similar story is true for the rear tires. The overall net affect is impossible to calculate without knowing the wheel's effective radius around its centre, but the point is that, once again, it isn't as bad as people think for well designed and constructed wheels.

So let's say you use light weight and well designed wheels to overcome most of the downside, what's up the side, just looks? I don't think so. Looking back to the data for answers, here are lap times vs (rear) wheel diameter.

Could it be? Do quicker cars tend to have bigger wheels? Only one car with a lap time under three minutes (3:00) uses 18" wheels - the Cadillac ATS-V - but everything else is 19" or larger. Ironically, the ATS-V rides on the same chassis that gave birth to the current Camaro.. and it has very slightly more power along with a fast shifting 8-speed auto vs the manual in the Camaro. Yet, it is noticeably slower, to the tune of a very significant five seconds (for the ATS-V sedan, the coupe is "only" 4.4 seconds slower than the SS 1LE). And, aside from the Viper, every car with a lap time under two-minute-fifty (2:50) uses 20" or larger rear wheels. Quicker cars tend to be expensive, special performance models, exotic, or any combination. You could argue that this makes them more likely to have larger wheels just for looks to match the "status". But there are two problems with that thinking. One is called the 991 GT3 RS and the other is called the Viper ACR.

These two cars are two of the most, if not THE most, hardcore production cars that are dedicated to the noble cause of speed and track performance. The 991 GT3 RS uses not 19" or even 20", but 21" rear wheels. What does that tell you? Keep in mind, that's the same car that, in pursuit of saving weight, does away with door handles and gives you something that James May described as "little bits of rag," in his review of the last generation Boxster Spyder on Top Gear. Do you think they would do that just for looks? The list of weight saving efforts on the 991 GT3 RS includes stuff like carbon-fiber panels for the engine cover, the front trunk, and fenders, a magnesium roof, lightweight lithium ion battery, removal of air conditioning, removal of audio system, centre locking wheels, and lightweight suspension components. It is hugely unreasonable to expect them to go through all of that and simply through big wheels on for looks. The GT3 (non RS) uses 20" wheels, not 21", by the way. The 911 R, the less hardcore, less capable, manual-transmission option that is not obsessed with lap times also uses 20" wheels. The story is similar for the Viper, where the less hardcore TA model uses 19" rear and 18" front wheels. The ACR uses 19" wheels front and back. You can draw your own conclusions.

Do you need more proof? Well, in a Car and Driver test of upsized wheels and tires (link: Effects of Upsized Wheels and Tires Tested), they found that 235/35/19 wheels generated 0.01 lat-g's less than the smaller 225/40/18 (0.88 g vs 0.89 g). Test tires were Goodyear Eagle GT so C&D asked Goodyear for their explanation and "they postulated that the added [tire] width may have given the outside tire more grip, which would increase body roll and could therefore decrease the load on the inside tire enough to lose 0.01 g on the skidpad." No mention of bigger wheels, more weight, etc. or even the suspension not being able to handle the added weight, despite the test car being a 2010 VW Golf with stock 15" wheels that, combined with 15" tires in stock size, weigh 14 lb LESS than the 19" ones.. EACH, meaning the suspension is guaranteed to not be designed to handle the added weight. The trouble was too much weight transfer. A little off topic, but the reason for that is the non linearity between the ability of a tire to generate grip and weight. In other words, two tires with 800 lb on them, each, will generate more grip overall than one tire with 1,200 lb and another with 400 lb, because the increase in friction forces at the loaded tire due to an additional 400 lb of vertical load is less than the drop at the unloaded tire due to losing 400 lb of vertical load.

Back to topic, if you dig a little deeper, the theoretical reason why bigger wheels help is their effect on tires. Bigger wheels better control tire flex under load. Limiting tire flex results in a stiffer tire. You'll hear people refer to sticky tires as "soft compound" sometimes, but there's a difference. You want a soft surface to conform to the road texture and shape but you want a stiff tire structure. Limiting tire flex effectively increases stiffness. And tire grip is directly proportional to its cornering stiffness. Increasing tire stiffness is why high performance cars have low profile tires. The stiffer tires also provide a lot of benefits like better response, less deformation and heat buildup, better stability at high lat-g loads, etc.

Chevy doesn't use the biggest wheels. That would be the rear wheels on the GT3 RS (21") and the front and rear wheels on the Audi RS7 (21"). But aside from those, it uses either the biggest in its class or tied for the biggest. The Camaro uses 20" front and rear wheels on both 1LE models, the V6 and the SS. The Corvette Grand Sport and Z06 use 19" front and 20" rear wheels. I suspect the reason why the Corvette doesn't use front 20" wheels is that it doesn't need to, because its front end isn't nearly as loaded as the Camaro, with a better rear weight bias combined with a more rearward engine placement. As a result, the tires have to deal with less load and can be downsized from the Camaro's. What Chevy is doing is more effectively using wheels and tires to maximize available grip that can be extracted from the tires. This thinking of maximizing available grip extends beyond wheels and tires. Going back to the weight transfer issue from one paragraph up, all manufacturers try to minimize weight transfer for better handling but Chevy goes a step further.

Since weight transfer minimizes available grip, you could throw the best wheels and tires available but if you transfer too much weight, you can't use them to their capacity. Plus, a lot of weight transfer means delayed responses, less stability and confidence, etc. An easy solution is to just increase roll stiffness through springs and dampers, but that also increase vertical stiffness and you may not want to do that. Roll bars are better in that regard, but you increasingly couple left and ride sides if you rely on them too much. The best solution is widening track - the distance between the centre lines of the two wheels and tires on one axle. Performance cars use a combination of all the above, but here's how Chevy's push one step further.

If you exclude light cars like the Fiesta ST and Miata, and exclude very front-end-light cars like mid-engine and rear-engine cars (including front-mid engine like the Corvette, Viper, and AMG GT), the Camaro 1LE has the widest front track for its weight of all cars tested over the last three years, with the exception of the spiritual successor to the BMW 2002 - the M2. Meanwhile, the 5th gen Z/28 had the widest track, period, of any car ever tested by C&D for Lightning Lap features over the same period, tying the Ferrari 488GTB for the honour. Does that matter? Once again, if you look at the data and plot front track widths, you'll find a very clear correlation between quicker lap times and wider tracks. The same is true for rear track.

So far, everything is done to maximize overall grip. The final piece of the puzzle is focusing on longitudinal grip and putting power down - the differential. Namely, the electronic limited slip differential. A differential that can make better use of available traction makes a massive difference in a car's ability to put power down and bringing down lap times. And, to quote Sir Jackie Stewart: "The exit of the corner is far more important than the entry of the corner, with regards to smoothness." You obviously have get the entire corner right to get the most out of a car, but corner exit is more important than corner entry as far as lap times. That's especially true for non-momentum cars like these. I have experienced first hand the improvement different types of limited slip diffs can make but, going by the numbers, a good comparison to demonstrate the difference was done by Car and Driver in 2015 (link: What's the Diff?), putting a Lexus RC-F to the test with the standard limited slip diff and the optional Torque Vectoring diff. The difference was 0.03 lat-g around a 300 ft skidpad (0.94 vs 0.91 g) and nearly half a second (0.4 s) on a minute-nineteen-second (1:19.1) course. That's on an otherwise identical car.

Now, Chevy doesn't use a torque vectoring differential, but a good, electronically controlled, variable locking limited slip differential should be able to provide the same traction benefits of a torque vectoring differential, just not the yaw control due to the steering effect from torque vectoring. The V6 1LE does away with an electronically controlled diff all together, like the one the V8 Camaros and Corvette use, but I suspect that there isn't much to be gained beyond a good mechanical LSD, considering the much lower power output of the V6, combined with the much, much lower low end torque compared to the V8's.

After I concluded that the above seem like the advantages, I started testing my conclusions by comparing those components in Chevys vs cars they beat to see if the advantages do hold up. And they seemed to. For the most part.. there two cars stood out; the 991 GT3 RS and the Cayman GT4. The above components or a combination of them (big wheels and tires, wide tracks, and good LSD's) point to a Chevy advantage for the vast majority of cars but not the Porsches. They do beat their closest Chevy competitors (if classed by specs) - namely the Corvette Grand Sport and the Camaro SS 1LE. But Just.

The GT3 RS is darn near 300 lb lighter than the Vette. It has 40 hp more and far better power to weight ratio (6.3 lb/hp vs 7.5 lb/ hp for the Vette). It has rear wheel steering. It has more downforce. It also uses a variable electronic locking differential, big, wide, lightweight centre locking wheels, and even has a slightly better lb/tire section ratio (2.7 vs 2.8 for the Vette) and upsized wheels (20" and 21" front and rear vs 19" and 20" for the Vette). How is it that all of this nets no more than one tenth - that's 0.1 sec - advantage, despite having Porsche's excellent PDK, which should alone save a multiple of 0.1 sec in total shift times compared to the Grand Sport's 7 speed manual? A similar story is true for the SS 1LE vs the Cayman GT4, although with a bigger time gap (0.8 sec), no auto transmission, but much better tire to weight ratio (2.82 lb/mm vs 3.17 lb/mm for the Camaro). Both Porsches should have a big traction advantage because of engine location. It seemed like it should be a bigger gap for both cars, especially the GT3 RS. That kept hanging over what I concluded, convincing me I must be wrong. But after going through the numbers (a few times), I finally found a consistent advantage - gross tire footprint.

The GT3 RS has a slightly better tire to weight ratio than the Vette as mentioned, but, if you sum up its total tire footprint at all four corners, it comes up 60 mm short of the Vette's - nearly one fifth of a foot narrower. The Cayman has an even bigger discrepancy, with a total tire foot print that's 100 mm narrower than the Camaros, darn near four inches or a whopping one third of a foot narrower. Does it matter that much? Going back to that table of lb/tire section ratio, showing cars that have better lb/tire ratio but don't beat the Camaro in lat-g forces measured in the first corner, you'll probably conclude a resounding yes. Every single car that has a better tire to weight ratio but lower grip (judged by lat-g) has less overall tire footprint, with the exception of the GT350R. But, assuming my earlier conclusions are true, that can be explained with the other factors since it doesn't use an electronic LSD like the Camaro SS 1LE and it has smaller wheels.

Obviously, suspension design and tuning is critical. If not done properly, everything falls apart. And it's critical to ensure the car is fun to drive, stable, predictable, etc. The C&D test I mentioned earlier of upsized wheels and tires is critical in remembering that you have to think of the complete package and the entire car. Putting my conclusions together, assuming they are true in the first place, and making modifications to a car on that basis without proper development, testing, and supporting upgrades is like pitting a Mustang GT and a GT500 against each other in a drag race and, when the GT500 wins, you make the conclusion that you need a supercharger, so you go out, buy one, and slap it on top of the engine in the GT and call it a day. Without supporting modifications and tuning.

The point here is that, assuming proper development from all manufacturers, that seems to be how Chevy carves an edge; wide tires, big but light and well designed wheels, wide suspension track, and good differentials. It seems that, when those are combined with a great chassis and a genuine focus on performance and handling, the results on track or a good back road are very impressive.

Wednesday, 21 December 2016

Chevrolet 1LE & Grand Sport - How do they do it? Part 3

GM, in general, is starting to build a very strong reputation for chassis engineering but Chevrolets, in particular, have very strong performance on track these days, not just good handling feel and fun to drive attitude. In Part 1 (link: Chevrolet 1LE & Grand Sport - How do they do it? Part 1), I looked at different aspects and concluded that Chevys appear to have the advantage in grip. If you are still unsure that grip is where those cars excel, perhaps this number will change your mind: 1.11. That's how much lateral forces, measured in g, the 2017 Camaro SS 1LE generated in Turn 1 of Virginia International Raceway (VIR) during Car and Driver's Lightening Lap 2016 feature. 1.11 g also happens to tie the 2014 Viper TA, the 2014 Ferrari F12 Berlinetta, and even the 2016 Ferrari 488GTB. It gets more interesting too..

Car Max Lat-g
2015 Chevy Corvette Z06 1.20
2017 Chevy Corvette Grand Sport 1.19
2009 Mosler MT900S 1.16
2015 Chevy Camaro Z/28 1.16
2015 Porsche 918 1.16
2015 Nissan GT-R NISMO 1.15
2016 Dodge Viper ACR 1.15
2016 Porsche 911 GT3 RS 1.14
2015 Lambo Huracán LP610-4 1.13
2014 Ferrari F12 1.11
2014 SRT Viper TA 1.11
2017 Chevy Camaro SS 1LE 1.11
2016 Ferrari 488GTB 1.11

If you pay attention to the order, you'll notice that the list isn't arranged in order of age or model year - i.e. starting with the Camaro, being 2017, then the 488GTB, being a 2016, then the Viper and F12, both 2014 cars - or vice versa. Despite all being listed at 1.11 g, the order goes F12, Viper, Camaro, and 488GTB; two 2014's, a 2017, and a 2016. Unless you want to believe it was random, that must mean that if you look at more decimal places, the Camaro beat the 488GTB and was beat by the F12 and Viper. That's a humble pony car beating a purpose built, mid-engine (new) Ferrari in grip.

It also places 12th out of every car ever tested in Lightning Lap features. The tally adds up to 201 cars and this Camaro beats 189 of them, including cars like the 458 Italia, 911 Turbos (pick a generation, it beat them all), 991 GT3 (non RS). Cars that beat the 2017 SS 1LE include stuff like the GT-R Nismo, Viper ACR, 911 GT3 RS, Porsche 918, you get the picture. Going further up the hall of fame, you find that three of the top 5 cars are Chevys, taking 1st (Z06), 2nd (Grand Sport), and 4th (fifth gen Camaro Z/28). The Mosler MT900S managed to just barely beat the Camaro (both are listed at 1.16 g, meaning they must be separated by a few 1/1000th's), but everything else is beat by the top dog Corvettes. I think I rest my case that Chevy knows grip. And it's easy to see why Chevy focused on grip.

If you can't grip the road properly, you can't put down power, you can't brake as aggressively, you can't carry speed through turns, etc. That's why everyone who's been around a track a few times will tell you that tires are one of, if not the most, crucial piece of the going-fast puzzle. If you have good tires, generally resulting in better grip, any individual under-performing aspect of a car doesn't necessarily have the same effect on others. For example, a low-powered car doesn't necessarily mean it's slow - it could have great brakes, great handling, great downforce, or any combination. Not having much power hurts acceleration, but the brakes can still do their job slowing the car down, suspension can do its job keeping tires in contact with the road, aero components can still generate downforce to increase grip at high speed, etc.

Even within one aspect such as handling, for example, you could have a car that understeers on entry - a bad handling characteristic - but it could very well be good at putting power down. You could still be quick if you slow it down, turn it, nail the apex, and hammer the throttle. Tires, on the other hand, can single handedly ruin all aspects of a car setup and prevent ALL of them, simultaneously, from performing properly if not chosen well and grip is compromised. Conversely, they can improve every single aspect of the car, if maximized. The question then becomes this: how do they generate more grip? It isn't compound because, while they do use good tires, they don't use anything more aggressive than what other manufacturers use (aside from manufacturer specific tuning).

To try and figure out where Chevy's stand out in any one area, or if they do at all, I looked at Lightning Lap numbers in more detail. I collected data from the last three Car and Driver's Lightning Lap features about each car, including lap time, front and rear wheel and tire sizes, front and rear track widths, power, torque, and weights. Then, I trended a bunch of different parameters about the cars vs lap times to see if I find any correlations pointing to a Chevy advantage. First, look at this graph of lap times at VIR during C&D Lightning Lap features vs power to weight ratios (expressed here in the inverse, lb/hp ratio) for the cars.

This isn't relevant to figuring out how those Chevys go quicker, but I just want to establish trust between you and data, if you're someone who isn't used to looking at empirical data, and making observations and conclusions, without knowing all factors. Generally speaking, cars with better weight to power ratios are faster. You probably already know that. But if you didn’t and you had no idea how power and weight affect a car, you’d look at that graph and say that as this weight-to-power ratio number goes down, lap times go down. Here's another graph.

This one is of lap times vs weight distribution over the front wheels wheels (i.e. the lower the number, the less weight there is on the front axle and tires as a fraction of curb weight; more rear weight bias). You could also look at the lap times vs weight distributions and say that front end heavy cars tend to be slower and as you move weight to the rear wheels, cars tend to be quicker. You could make this correlation, and the above between power-to-weight ratio and lap times, while all other factors are unknown - some of which are actually crucial to the going fast puzzle - just by looking at the test data. And you'd be right. With that in mind, can we use the data for more? Even without a deep dive into suspension geometry and roll centres, torsional stiffness, spring and damping rates, etc., can data point to a Chevy advantage, in the suspension or otherwise? Stay tuned for the conclusion tomorrow in Part 3 Chevrolet 1LE & Grand Sport - How do they do it? Part 3!

Saturday, 5 November 2016

2016 Focus RS vs 2016 Mustang Shelby GT350R - Track Video

While testing a 2016 Focus RS for the comparison test (link: Ford Focus RS vs Subaru WRX STI vs Mitsubishi Evo X MR), I caught up to a 2016 Mustang Shelby GT350R and had a friendly head-to-head battle. Both cars were completely stock. The video doesn't capture just how good that car sounds. We had a chat afterwords and the owner was very cool about it. His rear tires were starting to look old and he told me it felt a little less grippy than he was used to, so they could have been heat cycled out. Our track is also short and technical, so high hp cars don't get much room to stretch their legs, robbing them of some of the advantage they'd have at a power and/or longer track. The Focus had the optional Michelin Pilot Sport Cup 2 tires. Check out the video below for a couple of laps.

Wednesday, 2 November 2016

Ford Focus RS vs Subaru WRX STI vs Mitsubishi Evo X MR

All these cars have one common Achilles' heel. The engines sit entirely ahead of the front axles; a great family recipe for understeer. Then tell the front tires - already taxed from trying to keep that front engine sitting outside the wheelbase from going straight - to put some power down and you can only make matters worse. There are ways to mitigate the understeer with suspension tuning, of course, but the toughest part is power-on understeer. I don't want to get much into tires, but the thing to remember is that because tires have a certain "grip budget" - how much total grip they can hold/generate before they let go - when you get on the power in a car that sends power to the front wheels (FWD or AWD), you will rob some of the precious grip you were relying on to turn the car in order to put all or some power down. You'll run out of front lateral grip sooner than you would have otherwise, as a result. Worse yet, because of the unideal engine placement, you need every last bit of lateral grip in the front. So what you typically do to mitigate that understeer is, generally speaking, give the front tires far more grip than the rears.

You know what all good handling front end heavy, FWD cars have in common? They turn into tripods and kick up the inside rear wheel when turning. You need as much grip at the front as possible to keep that engine hanging in the front in check so you tune for weight transfer to the front, lightening up the rear end in the process. That's great in a FWD car but with AWD, you make the rears less effective - the same tires you're now trying to send power to.. after having just added a whole lot of effort, weight, complexity, and cost to the car in order to be able to utilize them. As a result, you can’t let these cars turn into tripods and pretend the inside wheel doesn't exist half way around a turn. You are now using the rear wheels for power so they are far more useful, for one. For another, you can afford to lose a little grip from the fronts to the rears, since they don’t have to transfer all the power anymore. You can never lose sight of the unideal engine placement, though, that demands a lot of grip in the front.

Why is all of this relevant? Because you need to spend just as much effort managing power as you do managing available grip by tuning the suspension. That's how each car here defines itself. Where their characters and attitudes come from. And because differentials are a big factor in this, I made another post recently on various types of differentials in an effort to make this post more focused on the cars but still discuss diffs in a little more detail and answer some questions about the diffs discussed here. Here's a link to that post: Limited Slip Differentials - The Basics. With that said, let's start with the oldest car here; the Evo.

Mitsubishi Evo X MR

How many comparisons have thrown this out-of-production car into the mix of new entries to the segment? This isn't only out-of-production, it is also very old, with a platform that's a full generation and redesign older than the other cars here. You can notice that in NVH, the way the car looks, the way the interior feels.. but not the way the car drives.

Compared to the STI, the Evo X feels sharper, more nimble, and more agile. This is surprising because the STI, at first, seems to have the advantage. The Evo is heavier, to begin with. It's an older, presumably less stiff chassis. It uses a conventional bevel gear centre differential, giving it a 50:50 torque distribution. The STI's centre differential is a planetary type, giving Subaru the flexibility to gear it for a 41:59 torque distribution front to rear. To make matters worse, the Evo can bias torque to the front wheels by varying lockup in a clutch pack sending power to the rear. The STI can't (without slip and centre diff lock). The option of front power bias without the possibility of rear bias, more weight, and possibly softer chassis sound like a few strikes against the Evo. But the Evo is happy to return punches all day. 

Without slip, any power you send to the rear wheels will cause understeer. An open diff sends virtually equal torque to both wheels so there's no steering moment and your tractive forces push the car where the rear wheels are pointing; straight. A mechanical LSD sends more to the inside wheel around a turn until there is slip, so you even have a negative steering moment. What do you do? You CREATE steering moment! Enter torque vectoring differentials. The power you send to the back in the Evo goes through a torque vectoring differential to then distribute that power side to side - forcing it to the outside wheel creates steering moment. At the front, the Evo uses a gear-type LSD, maximizing use of available traction so you can put more power down. More power down at the front (without slip) is good, because the tractive forces are pointed in the direction you want to go, along with the front wheels. Both the Evo and the STI use limited slip mechanisms to lock front and rear axles if there is slip and shuffle more torque fore and aft. However, the centre diff in the Evo is strictly an electronic LSD. When going around a turn without slip, it is completely uncoupled. No lock, no resistance to turning. 

The final peace of the puzzle is power distribution. Why only 50% to the rears and why allow front bias? Well, you need a lot of grip in the front to keep the engine in check. If the car is turning, you need to allow weight transfer to the front to increase grip there because of the unideal engine placement. The downside is that you rob the rear wheels of their grip so they can't put down as much power as you'd like. Now, front power bias here is very different from what you hear people complain about in a performance car. That kind of power bias, typical in the tried-and-true Haldex AWD system, hurts because it is the default. You always get front power bias until there is slip, at which point the front tires are already struggling and the best you can do is ease their pain a little by lightening their load. In the Evo, power goes from the engine to the transmission, then to a differential that evenly distributes power front and rear. Then there is a clutch pack transferring power to the rear axle, which you can progressively disengage to put more power to the front if conditions allow (i.e. whenever you don't need peak lateral grip). So with that in mind, you either allow that (good) front power bias to be able to utilize more power there where you've put more of the available grip due to weight transfer, or you prevent a lot of weight transfer to the front so you can bias power to the back without losing traction, at the expense of front end grip and more neutral handling. The Evo chooses the former.

All of this translates into a huge difference on track. It turns in with surprising precision. You can still find some understeer at the limit but it is easy to avoid and manage. You can get heavy on the power very early and trust it. You need very little corrections and just let the car drive the proper line as if Mitsubishi taught it during development. I may be exaggerating but you'd probably be, too, if you've driven a few different cars on track and felt the difference. Another plus is that, because the Evo is close to neutral, if the centre diff locks due to slip, it will help the car rotate. Because, if locked, the front and rear axles' speeds have to be closer than they would be, unlocked. That means the tires on one axle have to slip a little to more closely match the others. If you have more front end grip, the rears will slip first, helping the car rotate and reducing power-on understeer. But don't confuse that slip with lack of stability. The Evo wants to make a hero of you. It is so stable, so easy to control, that you'd be forgiven to think it can't go wrong. The whole car is focused on managing power as efficiently as possible - front to back and side to side - to optimize the turn. You can just feel the car working under power, managing the grip, managing the power distribution. And it feels so eager to do so.

Forget about lap times for a moment. They can easily be improved (to an extent) with some relatively minor tweaks. For example, when it first came out, this car did a lap of 3:13.3 at VIR for Car and Driver's LL 2008. That's only half a second ahead of the 3rd generation WRX STI, despite reviews generally agreeing that the Evo feels much sharper and track ready. It wasn't until they brought it back (in SE guise) for LL 2011 and did 3:10.5, putting it a little over 3 seconds ahead of the STI. And one way or another, it was still an Evo X. The times don't tell the whole story. Focusing on the times is missing the point. Despite the age, despite the crappy interior, the relative lack of refinement, the dire image of Mitsubishi, and even being out of production, this car feels just as sophisticated, track focused, and special as the other newer cars here. That's the point.

Subaru STI

I'm not sure if Subaru's reputation is what is influencing current design philosophy or current design philosophy is what's fueling the on-going reputation. But I imagine that at every design meeting during the STI's development, the head engineer always asked everyone involved: "so what have you done since the last meeting to make sure it better puts power down?" I can picture someone, at some point, suggested big rear anti roll bars during development for better turn in. He probably got fired. That's not to say the STI handles badly. It's still sharp with great turn in. It has a lot of grip and does not plow straight under power. That's just not its party piece. It's not its specialty. 

The STI claws the road under power. If it were an animal, it would be a big cat - a lion or a tiger - and every time you got on the power, it would crouch, wag its rear end sticking in the air, and pounce forward. Whereas the Evo is all about managing power, the STI is all about putting it down. Traction seems to be the priority. You need just a little more patience than the Evo before getting back on the power but then it will put power down with absolute tenacity.

It uses a gear type front diff like the Evo so there is no difference there. In the centre, though, it uses not one, but two types of limited slip mechanisms - a gear type and an electronically controlled clutch pack. The planetary gear centre diff utilizes helical gears to create thrust forces and provide some lockup under power. That means, as you roll into the power, the diff progressively locks. That’s great for traction as it locks before slip occurs (remember how the STI is all about traction?) but more lock means resistance to turning so it won’t be as agile. The second LSD is an electronically controlled clutch pack – similar to the Evo – to supplement the gear type for more locking if need be or if you want to manually select and lock the torque distribution instead of letting the computer do the work.

You can forget all of that torque vectoring non-sense in the Evo. That's what the STI would say if it could talk. In a slalom test by Edmunds back when the Evo X was introduced, they found that the torque vectoring diff can be caught off guard in short and quick transmissions like a tight slalom and wag the rear end a little. The STI wouldn't have any of that and would happily give up the effective steering from torque vectoring. The rear diff in the STI is another gear type differential, unlike torque vectoring in the Evo, which again locks under power. It won’t wait for slip or a computer to think and shuffle power. It doesn’t want to slip. That’s the STI’s mission.

Then you get to weight transfer, where the STI refuses to rob the rears of their traction, giving them better grip at the expense of the fronts. It's more stable and means you can put more power down without slip. But it's at the expense of some lateral grip at the front and that precise point-and-go attitude of the Evo. Corner speeds have to come down some, but you should be able to make up for it in traction in corner exit.

With that said, the STI has a certain charm to it - a certain mechanical feel that makes it more natural. It drives almost like a RWD car, but not the best and most sporty of the breed. It drives like a very tame one - one that is very capable but has safe understeer dialed in and massive amounts of traction. As is to be expected of the front weight bias, you get a good helping of limit understeer. If you go on the power, you can help the car rotate, but you do so by slip. Just as you would in a RWD car, especially if it's locked in 59% power going to the back. You can break traction and get it to rotate. It also uses brake-based lock, which does help, but isn't nearly as effective as a torque vectoring diff. The slip also feels abrupt compared to the Evo and the RS. That could have something to do with the modifications - this particular STI had camber (-2.5 degrees all around) and sticky BFGoodrich g-Force R1 track tires (which are different from the tires used by the owner for the lap time quoted at the end). The abruptness could easily be attributed to those near-slick track tires, but I would be surprised if, even when stock, the rear end would step out as smoothly as the Evo and or anywhere as gracefully as the RS.

The mechanical feel and nature does more to impress too, depending on how you look at it. The Evo will make a hero of you, but perhaps at the expense of doing more for you. You'll always wonder how much you could do without the car's help. In the STI, you feel more in control and will get it out feeling more accomplished, especially with the option of locking front to rear torque. It feels like the car just gave you all the traction in the world on a silver platter and what to do with it is up to you. The Evo has a wider range of torque distribution front to back and then again side to side on the rear axle. Aside from the torque vectoring diff simply sending torque to the outside wheel to help the car rotate, it's hard to say whether the Evo actually does that much more for you, as far as managing power, or the STI just hides it better. But it certainly feels like the STI is less intrusive or in control, which is a big plus in my book.

Ford Focus RS

If this car had any more hype and people talking about it, it'd have its own reality show. Drift mode, torque vectoring, optional Michelin Pilot Sport Cup 2 tires from the factory (standard in Canada), 350 hp, the lot. Plenty of buzzwords. Ford set it up for disappointment. It has to deliver on so many things, and deliver well, to meet expectations, let alone impress. It's bound to disappoint.. Or is it?

The RS takes the neutral balance of the Evo and turns it up a couple notches. It then takes the STI's rear power bias and turns that up a couple of notches as well. What's the problem in a front-end heavy FWD car, understeer? Ford says, with a smirk, let's fix that for you. When I said you could either allow front power and weight transfer bias or rear power and weight transfer bias, I actually left out secret option number 3. Give the front end a lot of grip, let the back lighten up, and still send plenty of power back there. Better still, allow the rear axle to use up to 100% of that power to one wheel to help the car steer. The RS has a very noble mission - make you forget it's a FWD-based hot hatch. And boy, does it ever try. You drive this car properly, and you may actually forget that you are driving a FWD-based hot hatch. The rotation under power is not only helpful, but very refreshing. And massively entertaining. Plus, you can't help but get all giddy when driving a humble Focus that power oversteers coming out of corners. The Evo makes use of an AWD system to beautifully manage power and maximize corner speed. The STI makes use of an AWD system to give you traction a 911 would envy. What Ford does, though, is use an AWD system to make the car turn. Speed seems like a byproduct since, you know, you do have to turn to get around a track. And the RS is very good at that.

The RS uses a torque vectoring rear differential like the Evo, although of completely different design that does not rely on a traditional differential at all. Instead, it uses two sets of hydraulically actuated wet clutch packs that individually control the amount of torque each rear wheel gets at all times. In the centre, the RS, once again, does away with a traditional differential and relies on those individual clutch packs to proportion power to the rear. If both clutches are disengaged, you get no transfer to the rear. If both clutches are equally locked - fully or partially - you get torque transfer in proportion to the amount of lock up, split equally between the two wheels. Or you could vary lockup between the two sides to individually send torque to either wheel. Using clutch packs to proportion power to the rear is very common in mainstream AWD vehicles, although is typically done via a clutch pack before the rear axle and a conventional rear differential. This is much less desirable than a differential in FWD-based performance cars, as discussed earlier. So how does the RS overcome this? It over speeds the rear wheels.

Similar to how the STI gears the centre diff to bias power to the back, the RS is geared at the power takeoff unit to drive the rear wheels faster, although much more aggressively at a ratio of 1.7 the speed of the fronts, thereby having a higher load and transferring more torque, resulting in rear bias. And, like the Evo, this also allows good front power bias if conditions allow by progressively reducing lockup on the clutch packs. At the front, the RS uses an open differential with brake-based lock, a disadvantage to the Evo and the STI. But, because of the aggressive front weight transfer, how much grip the the front axle has as a result, and how much power it sends to the back, I didn't run into the limitation of the front axle (i.e. excessive inside wheel slip). I have no doubt that you'd see an improvement with a true LSD at the front, but you'd probably struggle to get even near the same improvement that you'd see in a FWD car like the Focus ST.

Some people ran into overheating issues where the rear differential/drive unit (RDU) got disabled and the car basically became FWD. A friend of mine has an RS and ran into that problem the first day he was on the track. I never had a hiccup. That could be because of how little time I spent in the RS - a 15 minute stint, a 10 minute stint and a 5 minute stint. It could also be because of my driving style that we suspected strains the AWD system less (more on that in a moment). Or the fact that the car I drove was well broken in with over 12k kms on the clock, about 7.5k miles, because that same friend who ran into the issue the first time he was on track had no issues in subsequent times with more mileage. That lead us to thinking it might be an over protective feature during early break-in since the first time he was out, the car had just 1,800 kms on in, or just over 1,100 miles. I hope that we can say with more certainty with more track time next season whether or not that's an actual problem, but for now, it seems like a non issue.

So what does all of this mean on track? You'll have to recalibrate your turn in points and steering angles compared to a car with similar turn in response and handling balance because you simply don't need as much steering from the front wheels. You can feel the car working like the Evo but it doesn't seem "smart" in the sense of correcting your line. Where the Evo feels like it knows the right line and reads your mind, the RS doesn't. Yaw in the RS seems to be directly linked to your right foot - give more power and you get more rotation. If you are using a lot of steering and too much throttle, the car thinks you really want to turn and will put plenty of power outside to get the car turned hard, even if that means running out of the room on the inside and basically hitting an early apex. The Evo, I believe, knows exactly how hard the car should turn based on steering angle. If you're turning a certain amount and the yaw sensor says the car isn't turning enough, you have understeer, and it will shuffle power outside to turn the car. If yaw is too much, it will shuffle power inside to put the car back in line. The RS seems to have more faith in you, for better or for worse. The car gives you what you demand with your hands and right foot. Kind of like the way the STI gives you traction on a silver platter, only the RS gives you a hyper ability to turn

There are many excellent RWD cars that let you steer with your right foot. The difference is how much steering you can do and, let's not forget, that this is an AWD car. It still has a huge traction advantage. You can go flat out very early. In fact, in a lot of turns, I was flat out before apex because power no longer makes you go wide, it actually helps you turn. And you never have to worry about spinning out because it is massively easy to manage at the limit. It's fastest with some rear slip. The thing that you'll have to learn is trusting the car. Because it can hold and recover from huge yaw angles, all while still putting power down and gaining speed. You have to learn when to put your foot down and be able to put your foot down. Hard. That's the way you can maximize the AWD system. You need a lot of sisu or experience to do that, when your past experiences and your brain are telling you that you cannot do that here, otherwise you'll either spin out or plow straight.

It isn't without fault. The engine is still sitting far ahead and you can find limit understeer. But, compared to the other cars, it's like finding a lost sock in a dryer. The trick, in my opinion, is to not drive this car fast the way you do other cars. You don't maximize corner entry. In fact, you sacrifice corner entry just a little. The trouble is that, if you maximize corner entry, that means that you are coming in basically at the edge of grip. Right up to the limit. The car is as close to neutral as you'll probably get in an AWD hot hatch, but there's still some safe understeer left on the table. You'll find that and get frustrated. You can go on the power to correct, but because you're on the edge, you'll mostly correct by spinning, not by torque vectoring, because the tires are near the limit already in lateral grip and can't put much more power down without slipping. There's nothing wrong with that. That's what you do in a good RWD car to help the car rotate. But it's missing the point of this car and wasting all that went into the AWD system along with a good chunk of the torque vectoring benefits. You need to conserve some grip at the rear tires to use for putting power down because that power is going to help you turn.

The best part of all, I believe, is that you'll be just as fast if you suit your driving to the car. The same friend who bought an RS this summer complained about understeer in a few turns. He had better corner entry than me, with higher entry speed and later braking in nearly every corner on the track. His lap time? 0.28 s slower than mine. I could take much better advantage of torque vectoring because the rear tires weren't as burdened as they would have been with optimal entry. That means I can get the car rotated as I am on the power and gaining speed, giving me a better corner exit. I think throwing the car into a corner is good for fast turns. Conserve speed and momentum. In slow corners, though, you have to adjust your driving.

But no matter how you drive this car, the real treat is how close it is to absolute neutral handling. How easy it is to have fun with the power and how manageable and controllable it is when it lets go. It really does a very good impression of a true sports car, one that just happens to be a practical hatchback. And while it may fall short occasionally and let you know it isn't without compromise, it more than makes up for it by how much fun you'll have driving it.

VW Golf R

You might have noticed that the Golf R is curiously missing from the title. Or that the Evo X is in it, even though I previously posted and said that it is out (link: Mods and Update: Focus RS vs Golf R vs WRX STI vs Evo X). Initially, I was going to test all four cars (link to original post: Intro: Focus RS vs Golf R vs WRX STI vs Evo X). Some back and forth, scheduling conflicts, etc. meant that I could only get hot laps in the Focus RS, find out what the WRX STI and Evo X are like on track but no opportunity for a time, and no impressions at all in a Golf R. Such is the trouble without a big audience and manufacturer-provided cars for review.

I'm just as disappointed as you are. Due to the time of year, the season is coming to a close and I won't get another opportunity until the next season to test again so I thought I'd post what I have and hope for a better outcome next season

Lap Times 

Although I had no opportunity to do hot laps in the Evo and the STI, I got lap times and logs from the owners. First, here's a map of our local track, Atlantic Motorsport Park, to clarify a few turns. Namely, Turns 6, 8, and 10. All these turns don't need the use of brakes in entry or even backing off enough to scrub off speed. In fact, turn 6 is taken flat out from start to finish. As a result, they can't be marked well on the track logs so I hope the map below can clarify.

With that out of the way, here are the lap times, followed by track logs of the Evo vs the STI, Evo vs RS, and RS vs STI.

Best Lap
Evo X MR
- modifications:
  • Intercooler pipes
  • Cat-back exhaust
  • Cone air filter
  • Custom tune
  • --------------
  • Lowering springs
  • Rear anti-roll bar
  • --------------
  • 18x10 wheels
  • Firestone Firehawk Indy 500 tires sized 275/35/18
  • --------------
  • Carbotech track pads
  • Braided stainless steel brake lines
- modifications
  • -2.5 deg camber all around
  • Bridgestone Potenza RE-71R tires sized 265/40/18
Focus RS
- stock
  • Optional Michelin Sport Cup 2 tires (235/35/19)

STI vs Evo X: If they were stock and driven optimally by the same driver, you should expect higher corner speeds in the Evo but better exits in the STI. It's tough to say which car has the grip advantage as they stand, with modifications. The STI has stickier but slightly narrower rubber and very good camber for a street car but much narrower wheels. The Evo X has an AWD system more suited for maximizing corner speed, suspension upgrades, wider wheels and tires, but no camber and street tires. Their corner speeds are extremely close, suggesting it's wash if we assume comparable driver skills (remember, they're different drivers). The extra power in the Evo shows in steeper acceleration curves out of T2, T3, back straight, and T11 leading to the front straight and appears to have been the winning factor. Time wise, it looks like the Evo made all its lead on power, as they appear to be in a dead heat up to T5 leading to the back straight. The Evo should also have a slight advantage in shifting, being the MR with a dual clutch automated manual (it is quick).

Evo X & STI vs RS: Despite the stock suspension, much narrower wheels and tires, the RS appears to have a consistent advantage in corner speeds vs both cars through high speed turns - basically the second half of the track past turn 6 - presumably due to less understeer. The STI should have a huge grip advantage, with similarly sticky tires that are actually much wider and camber. The Evo is tougher to judge, since the RS has much better compound but both wheels and tires are a far narrower. I found that the Focus likes slightly slower-in-faster out approach in slow turns to utilize the torque vectoring, as I mentioned earlier, and this appears in the logs, where the Evo and STI have better corner entries into T2 and T3 but the RS seems to have better exits. 

The STI's modifications seem to provide it enough of a grip advantage to overcome the RS' handling advantage but without any more power, they're nearly tied. The Evo's suspension modifications seem to do the same, but the power advantage allows it to actually pull ahead slightly in just about every corner of the first half of the track. With that said, the RS was seriously held back by the 91 gas, IMO, since we don't have 93 locally.

It's clear, in my opinion, between T3 and T4 - just before braking point - the RS just loses steam and stops accelerating, then briefly gets back on. I could even feel that on the track. Between T4 and T5, the same thing happens. Then again between T5 and T6 and on the back straight. Finally, after exiting T11, the RS pulls power again before finishing on the front straight. How much is that worth? I went through the logs and adjusted the data as if the RS didn't lose power and here's the result:

Yep, just over a second. The lap time would have been 1:18.61, 1.01 s faster than otherwise. This is with absolutely no changes to the lap - same turn in, braking points, amount of braking, etc. - as you can see by the two laps being identical except for the sections where the RS seemed to have pulled power. How much is it making on 91? Don't know, but the Mustang EcoBoost is supposed to lose 35 hp when going from 93 to 87, 11.3% of peak. If we assume the same in the RS, it would be making 310 hp on 87 and somewhere between 310 and 350 on 91, which should actually be optimistic because the RS makes more boost so it should be more sensitive to octane than the Mustang. To calculate acceleration assuming no loss of power due to 91 gas, I used Car and Driver's test data of the RS, trended them, and used that data to calculate adjusted speeds had the car not pulled power. With the corrections, the lap times would look more like this:

Best Lap
Evo X MR      1:18.06
WRX STI1:19.72
Focus RS 1:18.61

Now, you can't claim the 1:18.61 time as the Focus time. Because it isn't. But I wanted to give it a more fair representation by correcting the time, in case you're curious how much better it could be on better gas. Another disadvantage for the Focus is seat time. Where best laps for the Evo and the STI were set by the owners, with hours of seat time in the cars, I had just under half an hour in the RS overall. The owners of both cars have had these cars for as long as I've known them, about two seasons. But even within one day, in a car you're already very familiar with, you can expect to get better as the day goes along (unless heat becomes a factor). Case in point; the 1:18 time in the Evo was preceded by slower lap times, where the first session was all 1:20's and slower, the second was 1:19's, and the third had the 1:18 lap. If we assume a similar drop in the RS with more seat time (~ 2 seconds), along with 93 gas or octane boost (another ~ 1 second), the RS would have the best lap by a substantial margin, with a best lap in the 1:16 range.

3rd Place: Subaru WRX STI: It was really tough between this and the Evo. Assuming all are stock, I really think the STI would be the easiest to drive fast for an average track guy who goes to a couple track days and HPDS's a year. It has the most traction so it puts power down really well and, because it would make most of the time in corner exit, you can still get a great lap even if you don't nail the braking and corner entry. The RS and Evo need you to work more on those areas. It is also the most natural feeling in terms of handling and it's the one I would put my money on when the white stuff starts falling because of the better stability and traction. But on track, compared to the other two, it just feels like it's missing an edge.

2nd Place: Mitsubishi Evo X MR: I still cannot get over how this car drives on track despite its age. It is hugely impressive. It had the best raw time, albeit helped a lot by the modifications. It feels so precise, yet so stable, and manages power really well. For someone who doesn't care about the added practicality of the hatch, the refinement, or the subjective fun to drive factor of the RS, the Evo would probably be a more appealing car.

1st Place: Ford Focus RS: If one of these cars would serve as a daily driver, you'd never regret picking this car just because of how much more refined it is, plus the practicality of a hatchback. And then you'd take it to a track and find out what a blast it is. It is more neutral than any other production hot hatch dares to be. It will powerslide and dance, at full throttle, with just the right amount of yaw in corner exit, in a way a FWD-based car has no right to. And, if you consider the modifications of the other cars, the lack of 93 gas and octane boost during the test, and the very limited seat time, you'll find the lap times to be very impressive. The potential, assuming better gas, is even more so and, with more seat time, it could be even better. Did I mention it's easily the most fun?

I've put a lot of time researching the AWD systems of these cars, especially the Evo X and the STI, since it seems no two people quite agree on exactly how they work. I cannot say with 100% certainty that I interpreted the countless articles, diagrams, and drivetrain sections properly, so take that for what it's worth. There are a lot of conflicting opinions out there and you may have done your own research and come up with your own understanding that's different. If that's the case, feel free to comment or message me and include links/source and I'll be happy to update if there's an error. 

Like us on Facebook! If you like this post, please remember to like us on Facebook through the link on the sidebar on the right!

Photography by Kevin Doubleday and Albert Hofman

Monday, 24 October 2016

2017 Camaro ZL1 Beats Previous Generation’s Nürburgring Lap Time

The new 2017 +Chevrolet Camaro ZL1, expected in showrooms by the end of this year, just beat the benchmark set by the last generation ZL1. With a lap time of 7:29.60, it is 11.67 faster faster than the last generation and even beat the last generation Z/28's time of 7:37.9 - which was done on Pirelli P Zero Trofeo R tires, far grippier than the Eagle F1 Supercar used on the ZL1. The car used is unchanged from the one you'll be able to buy, aside from the installation of data acquisition equipment, a roll hoop, and Sparco racing seats with six-point harnesses. Otherwise, the car was production stock and included the following:

  • 6.2-liter supercharged LT4 V-8 making 650 horsepower and 650 lb-ft of torque
  • All-new 10R90 10-speed automatic transmission (set to Track mode to enable Performance Algorithm Shift calibration, providing optimal gear selection without the need to manually select gears)
  • FE4 Suspension with Magnetic Ride Control
  • Performance Traction Management
  • Forged 20-inch wheels with Goodyear Eagle F1 Supercar 3 tires
  • Brembo brakes with front 15.4-inch rotors and six-piston calipers and rear 14.4-inch rotors and four-piston calipers
  • Lift-reducing front fascia elements with cooling ducts and Chevrolet “flowtie”
  • Full underbody shielding
  • ZL1-specific rear spoiler and diffuser
  • 11 heat exchanger

And best of all, with the exception of the 10 speed auto, all the above is included as standard, with no requirement to purchase additional performance packages. On a track like the 'Ring, the auto would have been worth a few seconds, I figure, with the number of shifts required so if you get it with the manual, you might close the gap compared to the last gen Z/28, but the time is still phenomenal and runs with plenty of very capable cars. A lot of times aren't official on Wikipedia or Fastest Laps aren't manufacturer official so they're hard to compare, but a notable victim is an AMG-confirmed lap time of lap time of 7:30.0 for a 2013 SLS AMG GT.  Plus, beating the mighty 5th gen Z/28 by over 8 seconds is quite a feat, since we've seen that Z/28 beat cars like the GT-R, 991 Turbo S, and Mercedes-AMG GT S when tested by Motor Trend or Car and Driver. This should be one heck of a tough car to beat on track. I can't wait to see how it does in a comparison test. Until then, here's a video of the lap:

Saturday, 22 October 2016

Limited Slip Differentials - The Basics

I'm finishing up a comparison post (link to introduction: Intro: Focus RS vs Golf R vs WRX STI vs Evo X) and, throughout the post, I realized that I have to go off topic a lot to talk about how each type of differential changes the way the car drives. As a result, I thought I'd write a separate post to go into more detail before I post the comparison to keep it more focused on the cars and avoid veering off topic too much.

By saying "Limited Slip Differentials" in the title, I am including torque vectoring diffs because, although current conventional terminology treats them differently, a torque vectoring differential is, in essence, a very sophisticated limited slip diff (LSD) that can be manipulated to actively help the car handle better. And while none of the cars in the comparison use open (without help from the brakes) or non-gear mechanical LSD’s, I’ll briefly discuss them so that the post is more inclusive. I’ll only focus on using power to help the handling or how a diff can handicap that, since the reason I started to write this post is to demonstrate how the differentials help each car. I won’t talk about other techniques that could help you manage a car’s weaknesses, such as changing turn in points, apexes, trailbraking, etc.

So how do traditional LSD's and torque vectoring diffs (TVD) help the car? Let's first start with open differentials.

Open Differentials

These are the most common differentials and they are the best at being differentials. The differential's job is to allow two wheels on the same axle to spin at different speeds so a car could smoothly go around a corner since each wheel has to travel a different path and, therefore, at a different speed - hence the name - to reach the end point of the turn at the same time. The video below has been used countless times to demonstrate how a diff works and, although made by GM almost 80 years ago, is still one of the best videos I've found that explains very simply and visually how a differential works (fast forward to about 2:00 in).

I don't want to get into the internals and workings of a differential, but I wanted to share that video because understanding the basics will help with understanding the impacts of various types of differentials. As you can see in the video, an open diff allows one wheel to spin endlessly, even if the other is completely stationary. The demonstration at 5:30 into the video shows that. If one wheel has a lot of traction, it's harder to spin, much like being held still in the video relative to the other one, (the road is "holding" the tire, in effect). If the other has little traction for some reason, the diff will spin it, since it is easier to spin. The diff transfers virtually equal amounts of torque to both wheels so the wheel with little traction will dictate how much torque the wheel with a lot of traction gets because if you give more torque than the low traction wheel can hold, it will spin, reducing your traction even more as well as lateral grip.

This is a double whammy if you have uneven available grip between two wheels on the same axle. When you have one wheel that has relatively little torque carrying capacity, but no way to unevenly distribute torque, you can more easily overpower it. Moreover, the wheel with a lot of traction and, therefore, good torque carrying/transfer capacity is underutilized. The result is limiting how much power you can use to move (accelerate) and increased likelihood of reducing your grip by spinning the low-grip wheel, which still contributes to the car's overall lateral and forward grip available. If that happens at the rear axle (RWD), that spinning low-traction wheel means less grip at the rear end and more likely to oversteer. On the front axle, it's understeer. This is assuming that, in either scenario, you're applying power.

How does this work on track? When you're going around a turn, the inside wheel is unloaded because weight is transferred to the outside wheel, which means the inside wheel has less grip. That means it can transfer less torque than the outside and if you exceed that, it will spin. If it starts to spin (excessively), it will have even less grip. Less grip means you'll be able to use even less power and your corner speed has to come down since one of the tires now has less grip. In short, an open diff works really well at allowing different speeds between the two wheels but under-utilizes available traction so it limits how much power you can put down and makes it easier to spin under power.

So what does that mean if you're pushing the car? When approaching a turn, as you start to turn in, you come off the power. Typically.. The reason I say typically is that with some of the other differentials, you can actually start using more than maintenance throttle as soon as you come off the brakes, but more on that later. For an open diff, you are off, aside from maintenance throttle. If you can add power and gain speed between turn in and apex (where you start to unwind), you lost more speed than you needed on the brakes. Assuming optimal entry, you shouldn’t be able to add speed without understeer, oversteer, or a neutral drift, depending on the balance of the car. The diff can’t help you here. Worse yet, on a FWD car, you can’t use the power to help the car rotate. Unlike RWD, where you could judiciously overpower the rear wheels, inducing slip and rotating the car that way, if you overpower the driven wheels in a FWD car, there’s no way to go but straight. And this is very easy to do in an open diff while going around a turn, with the inside wheel being unloaded.

Need more bad news? The vast majority of FWD cars have a transverse engine layout, placing the engine far outside the wheelbase of the car. The Dodge and Chrysler Intrepid come to mind as exceptions, with longitudinal engine, FWD layout. Audi A4s, too, if you don't get the AWD option. But even those still put the engine basically entirely in front of the front axle. This generates very nasty forces and moments that do their best at pulling the car straight when you want to turn. Then, of course, you have typical OEM suspension tuning that favours the front end letting go before the rear end for safe limit-understeer. The result is frustration and anger, perhaps some cursing, and eventually vowing against open differentials on the track and maybe even FWD all together (which can actually be made to work very well on a track).

Limited Slip Differentials

There are many types of limited slip differentials and, like I mentioned, I won't get into how they operate, just how they affect the car. I'm referring strictly to mechanical, non-gear type limited slip differentials here. These differentials are typically open differentials at heart with modifications or additions. Those modifications are designed to resist a speed variance across the differential. The result is a limit to how much faster a wheel can spin relative to the other, overcoming the limitations I mentioned for an open diff. This is achieved by locking the two axles together (to an extent). That extent depends on the design and spec of the differential - typically referred to in a percentage (%) number and occasionally as a Torque Bias Ratio (TBR). That % number is the difference in torque (in % of total) the diff can provide between the two axles. TBR is the ratio between the torque sent to the outside wheel to the inside wheel that the diff can deliver. The higher either number, the better the diff will be at putting power down as it allows more lock up. But higher isn't always better.

Limiting slip of a low traction wheel is great, as it can be the difference between accelerating and backing off the power when exiting a turn on a track. Trouble is, when under power, a limited slip diff of this type can't differentiate between turning and a slipping wheel. If you're going around a turn and starting to feed in power, the outside wheel is spinning faster than the inside wheel, which is normal. But the diff will start to lock up, in response to the speed differential, thereby transfering torque to the inside wheel. That means the unloaded wheel gets more torque, the opposite of what you want, and generates a steering moment in the opposite direction of the turn. Moreover, by locking up, there is resistance to the wheels spinning at different speeds, which is resistance to turning (i.e. understeer) since that requires each wheel travels a different arc at a different speed around a turn.

So how do these help, considering all that? You can go faster by using more power earlier in corner exit and, due to limiting inside wheel spin, you won't lose traction as easily which means you can better maintain your available grip. The downside is understeer on a RWD car. This is introduced by three factors:

- Locking up to any degree provides less speed differentiation than no lock up at all, which we've established is required for the car to turn.

- Putting more power down means more weight transfer to the rear end, which results in less grip at the front end; more understeer.

- You can maintain your grip for longer due to no inside wheel spin. If the rear axle can hold on for longer, you'll increase understeer.

- Torque transfer to inside rear wheel in a turn prior to it slipping generates negative steering moment (in opposite direction to the turn), resulting in understeer.

With that said, a car without a LSD will be slower than one with because, even if your corner speeds come down a little, you can get back on the power much sooner and more aggressively and that's where you make most of your time. Moreover, you can tune the suspension and chassis to reduce understeer so you typically only notice the understeer on a car that had a LSD added but is otherwise unchanged. And most good summer/track tires generate their highest grip with a very small amount of slip, meaning that if you're aggressive with the throttle, enough to just barely overpower the inside wheel where the diff is working as intended, that very small amount of slip is not hurting you and now the inside wheel is beginning to slip, causing lockup and torque transfer to the outside. As a result, you'll find that most good handling RWD cars actually have LSD's, such as modern Camaros and Mustangs, Corvettes, BMW's, Subaru BRZ/Toyota 86, etc. 

It gets even better on a FWD car, since you only have the first two factors against you. The other two are actually helping you; more grip at the front is less understeer and torque delivered by either front tire generates positive steering moment.  That means a LSD typically curbs understeer on a FWD car, even with all else being the same. The one caveat is that the axle locking can make it difficult to steer, if aggressive.

Torque sensing or gear type differentials

Torsen and Quaife are probably the most common of these types of differentials. They are very similar in function to more common LSDs discussed above, except rely on gears configured in a way to bind and provide locking. A gear type LSD is torque sensitive, hence the name Torsen (Tor for Torque and sen for Sensing). Torsen has two major designs T1 (first gen) and T2 (shown above). T2's are very similar to Quaife design (main picture above introduction). They all operate based on friction between the gears and the differential casing. Due to the inherent angle of the teeth on a helical gear, transmitting torque from gear to gear also generates thrust forces. These forces pushes the "pinion" helical gears against the differential case, providing lockup, instead of using a clutch pack to lock the axles to the case, for instance. The great thing about them is that they transfer torque even before slip occurs, since they progressively lock up as gear thrust forces increase and these forces are proportional to torque transferred by the diff and independent of speed differential across axles. In other words, as you apply more power, the diff progressively locks up and its capacity to carry torque increases, regardless of whether one of the wheels has begun to slip or not.

Both Torsen generations, T1 and T2, use the same basic principle but T1's are very rarely used now in new applications, if at all, and they rely on a different design that increases lock up. They utilize two different types of gears (helical and worm). Inherent to the design and arrangement of gears, the gears will progressively bind as speed difference between the wheels increases when there is excessive or uneven slip. Due to this nature, T1 LSD's typically have very high TBR's and provide a lot of lock-up.

The downside to gear type differentials is that they typically can't take as much abuse. They don't like to be launched hard and high hp, high grip cars seem to have issues with them since they operate on the principal of binding gears and diff cases. With that said, they are low maintenance and last longer in more forgiving cars. That's not necessarily slow, pedestrian cars - the list of high hp, high performance cars that includes them as OEM diffs includes 5th gen Camaro Z/28, the '12-13 Mustang Boss 302's, and the current Shelby GT350's, all of which utilize the T2 generation.

So how do these differ from the more conventional LSD's in operation? The main difference is how it locks. As discussed earlier, they lock because of the helical gears generating thrust forces that push the gears against the diff case, effectively binding and locking it up. Since the force generated is proportional to the force (torque) being transferred by the gears, lock up is proportional to input power (i.e. how much power you're applying). If you're off the gas, it's basically an open differential. As you roll into the power, it progressively and smoothly locks up. The benefit to that is, because lock up is smoother and the diff is more open off power, you can typically get away with higher torque bias ratios than non gear LSD's at maximum lockup without seeing as much of the side effects of higher locking. The higher TBR allows better traction performance.

Moreover, the fact that lock up is proportional to input power means the diff locks up as you send more power, without the need for slip. Non-gear LSD's need slip to work as intended. If you are going around a turn with no inside wheel slip at all, the outside wheel is traveling faster and the traditional LSD is locking up, slowing it down and speeding up the inside wheel, therefore, transferring more torque to the inside wheel. As you increase power, the inside wheel begins to slip and only as its speed passes the outside does the diff begin to slow it down. In other words, the inside wheel must slip first and then be limited. The diff will go through the sequence of little lock up (outside wheel faster than inside), then no lock up (inside wheel beginning to get over powered, spin, and accelerate to match outside, which from the diffs perspective is like the car going straight), then lock up again as the inside wheel speed starts to exceed outside. In a Torsen, the inside wheel is limited before it slips, since lock up happens before slip. Subtle differences, but can change how the car feels, plus the higher TBR means cars can better put power down, accelerate faster out of turns, and generally perform better on track.

Brake-based Differential Lock

A car that uses brake-based limit slip action utilizes an open differential just like described above but attempts to solve the shortcomings by applying the brakes to individual wheels. If you go on the power and one wheel spins, the car realizes that and applies the brakes at the spinning wheel. From the differential's perspective, that wheel now is harder to turn and more torque will get transferred to it. Fortunately, just as much torque will get transferred to the wheel with grip, giving better traction performance.

In high performance driving, this solves the two shortcomings of an open differential, loss of grip due to a spinning wheel and under utilizing good grip at the loaded, outside tire. As we've established, an open diff transfers equal torque to both wheels. In order to distribute torque where you want it (unevenly), the brakes are engaged to slow down the one wheel spinning excessively. It is artificially creating resistance at the low traction wheel (i.e. the brakes "grip" the wheel instead of the road through the tire). This increases the torque holding capacity of that wheel, and the diff as a whole, thereby allowing the diff to transfer more torque overall, half of which goes to the outside wheel where it can all be used. The downside is wasting some power simply spinning the low traction wheel against the brakes. The second problem is, as a result of trying to power the low traction wheel against the brakes, the brakes can over heat and prematurely wear.

What's it like to drive? The tech is extremely flexible because it provides complete uncoupling and independence between the two wheels when no lock up is needed and infinitely variable and adjustable bias when you do. You have non of the shortcomings of mechanical LSD's. But you'll hear a lot of owners and reviewers complain about their effectiveness, or lack thereof. The problem is the application, not the tech. In a Focus ST or a Golf GTI (non PP), you don't have liberally sized brakes, brake cooling, brake system capacity, etc. In reality, an optimized brake based set up can work very well. McLaren uses them, for example. You won't hear too many people complain about their performance.

When you already have massive braking thermal capacity, cooling air flow, braking power, etc, you could rely on this system and avoid a similarly flexible torque vectoring system. That would not only save cost and complexity since all you're adding is the programming to control the brakes the car already has, it also saves weight since a torque vectoring differential can be heavy. The one Lexus uses on the RC F and GS F, for instance, adds almost 70 lbs compared to the standard Torsen differential offered, itself a heavier system than open differentials. But brake-based lockup does have issues, otherwise, on non optimal cars - think non mid-engine, cost constrained, limited in space, or just about every other car that us mere mortals can buy.. And you waste the engine's power spinning the inside wheel.

To put that into perspective, my car has a Torsen diff with a bias ratio of 2.7 - meaning it can allow the outside wheel to get 2.7 times the amount of torque the inside wheel has while remaining locked up (which is equivalent to a 46% clutch type LSD, if you're curios). My car has 380 lb-ft of torque. That means, in an ideal traction scenario, at peak torque and lockup, going WOT, the engine is sending 380 lb-ft of torque to the diff and the diff is transferring all of it - 103 lb-ft will go to the inside wheel and 277 lb-ft will go to the outside - a difference of 174 lb-ft. To achieve the same bias with brake lock, each axle has to get the same amount of torque - the difference is that some will be used to spin the brakes. How much? 50% of the difference goes to the brakes or 87 lb-ft.  That translates to 75 hp (at 4,500 rpm where peak torque occurs), in effect turning my car down from a 444 hp car to a 369 hp car. Bad.. very bad. Moreover, those 75 hp have to be transferred into heat by the brakes and once those brakes overheat, they back off and you approach an open differential. This extent of power loss would be rare and if it were to happen, wouldn't last long, but it demonstrates how bad it can be. In a mid-engine car like a McLaren, where you have gobs of traction due to rear weight and optimal suspension tuning, you may not need as aggressive a torque bias and you have brakes the size of the moon that can handle the heat. But in every day cars, it doesn't work very well. Not yet anyway.


eLSD is typically used as short for electronic limited slip differentials. They basically combine the benefits of all the above without any of the downsides. It can take more abuse than a gear type. It can lock up smoothly like a gear type. It doesn't have the same resistance to speed differentiation (unless activated) so it can have a higher lock-up with no downside. It is electronically controlled so it can selectively lock and unlock as needed based on real time calculations and inputs from various sensors. It doesn't have to worry about brakes overheating. Its only downside, really, is complexity.

Unlike a mechanical LSD, it can distinguish between a faster spinning outside wheel as you turn (with no slip) and a faster spinning inside wheel due to slip. It won't lockup unless slip is happening, eliminating the negative moment due to torque sent to the inside rear wheel before slip occurs - what a mechanical LSD will do. It can give you high lock when you need, say, exiting a slow narrow turn, and low or no lock in corner entry, reducing understeer you'd get with a high lock. In summary, it gives you higher traction performance while reducing understeer when you don't need lockup.

Torque Vectoring Differential

A torque vectoring differential is very similar to an eLSD. The main difference is that eLSD's only transfer torque from slower to faster. An eLSD is a basically a typical LSD, say a clutch-type, where the clutches, and therefore amount of lockup, are electronically controlled instead of passively based on the difference in speed between the two wheels on an axle. That means that they can control when and how much to lock up but after that, the same principles apply and torque is biased from the faster spinning wheel to the slower wheel. However, a torque vectoring diff can transfer torque either way and can do so independently of a speed differential. The benefit is better control (from the car) and wider range of adjustability, allowing the car to correct and/or improve more frequently. Moreover, a torque vectoring differential typically doesn't transfer torque by locking up. Instead, it utilizes actuators (clutches and/or motors) and gear sets to overdrive or under-drive each wheel independently.

What does this mean? It means you have no resistance to speed differentiation (i.e. understeer) but as much torque transfer to either wheel (can be up to 100% of torque sent to the diff) as you want. The downside of lockup (understeer) is eliminated, a huge plus to start with. Then you get the best torque bias (basically infinite if designed to allow 100%) you could get, allowing you to utilize every last bit of traction available, and to top it off, you can have individual wheel torque control to help the car handle. For example, if the back end is coming out, you could vector torque to the inside to bring it in. This has the same effect as a stability control system applying individual braking to a wheel or more to bring the car back in shape but braking means scrubbing off speed. Torque vectoring doesn't. You could also transfer torque to the outside wheel to help the car rotate if the car is understeering, without forcing lock up. It's a technically wonderful system. The range of adjustments is huge. Opportunities to improve the handling are plentiful. It can make a huge difference in how the car handles. The only downsides, really, are even more complexity and a less "mechanical" feel.


Which one is best? Well, brake-based systems are hugely variable, but because of the heat issue and power wasted, I'd only take one if I had to choose between one and an open diff. Otherwise, my preference is the gear type LSDs, even though they aren't as capable as electronic controlled ones. They provide the best performance without going to an electronically controlled system, in my opinion. Compared to electronically controlled ones, they're more natural feeling, simpler, and cheaper to maintain. They demand more of you, expect less of the car. But the same can be said for better tires, better shocks, stiffer chassis, etc. so I can see arguing both ways. I would imagine that someone learning on and sticking to torque vectoring tech on track would see it as natural and consider it a baseline, anything else is a compromise. That wasn't the case for me, so they will always seem like the deviation from norm - the cool tech that manipulates the car to your advantage, not the bare essentials.

With that said, in a FWD (never seen one) or FWD-based AWD car, I would absolutely and unquestionably give in and take torque vectoring and/or eLSD's. In a heart beat. The drivetrain layout is working very hard against you to prevent the car from doing what you want it to do. The tech does its best to mitigate the limitations. The way I see it: it almost undoes the crime that is FWD-based, transverse engine layout as opposed to improve the text book longitudinal front engine, RWD layout.