The Mark Ortiz Automotive
Reproduction for sale subject to restrictions. Please inquire for details.
Mark Ortiz Automotive is a chassis consulting service primarily serving oval track and road racers. This newsletter is a free service intended to benefit racers and enthusiasts by offering useful insights into chassis engineering and answers to questions. Readers may mail questions to: 155 Wankel Dr., Kannapolis, NC 28083-8200; submit questions by phone at 704-933-8876; or submit questions by e-mail to: firstname.lastname@example.org. Readers are invited to subscribe to this newsletter by e-mail. Just e-mail me and request to be added to the list.
BRACKETING THE SETTING
What do you think of the idea of “bracketing the setting” – changing a chassis setting by a really big amount, probably an excessive amount, and then backing up?
It depends on the type of setting, and also how confident you are of the qualitative effect of the change. It can be a real problem if the effect of the change reverses when you go past a certain point, especially if you don’t know that.
Wing angle would be a setting where we pretty much know the effect qualitatively, and just want to find the optimum. We know that increasing attack angle will increase both downforce and drag, at least as long as we stay short of stall. We also know that when we do hit stall, we will see a big increase in drag and probably a reduction in downforce, and we know that increasing attack angle will lower the stall speed.
So we know in general terms how a big change will affect the car. We know that if we increase the rear wing attack angle, that will add understeer. We can make a big change, and if the car has gone from too loose to too tight at high speed, we can go back part way. We can ask the driver, and/or look at the data, and decide whether to go back a third of the way, or half, or two thirds. We may very well be able to zero in on the optimum more quickly this way than by trying to sneak up on it.
Tire pressure is different. There’s a pressure where lateral grip or force capability is maximized, and grip falls off both above and below the optimum. The same holds true for longitudinal grip but the optimum pressure is generally lower.
Suppose optimum pressure is 38psi and we’re at 34. If we try going all the way to 44 and the car is worse, do we conclude that we should go softer than 34 instead? That wouldn’t be correct.
If we are reasonably close to a good setting, and we aren’t even sure which side of optimum we’re on, or indeed whether we’re at optimum, then small changes make sense. If we make a change that’s
just big enough to produce a measurable or noticeable effect, and the car gets better, we can try another similar one and see if the trend continues. If the car gets worse, we try a similar change the other way. If the car gets worse both ways, we know we’re close to optimum.
One advantage of making incremental changes is that it gives us a better picture of the “shape” of the car’s response. We can see when we’re approaching optimum. We can eventually get a feel for how the car responds when it’s close to or far from optimum.
One disadvantage of making a lot of small changes is that over a large number of runs, just running the car may change it, or conditions may change. We may go slower no matter what we do, just because we’re using up the tires or because the weather, the track temperature, the amount of rubber or oil on the track, or something else changes. The usual way to avoid being misled by such things is to start at baseline setting and return to that for the last run of the session. That will give us a reasonable read on the effect of changes in conditions or changes due to just running the car.
This is not totally foolproof, because more than one factor can change at once, occasionally with mutually cancelling effect, but it’s good practice. It’s a bit like balancing the books in accounting. If they balance, that’s good but it’s still possible there’s an error; if they don’t, you know you need to correct something. If you go back to baseline and the results repeat, you don’t know for sure that nothing changed in an uncontrolled manner, but if the results don’t repeat, you know something other than your setting did change.
DRIVING ALL FOUR WHEELS, FOR A HIGH SPEED PAVEMENT VEHICLE
I recently learned of the CERV II experimental vehicle that GM created back in the ‘60’s. It was a rear-mid-engine sports racing car that had all four wheels driven, using two 2-speed transaxles, both with torque converters. The front transaxle was driven from the front of the engine. The engine was on center and there was no transfer case and no drive shaft alongside the engine.
What do you think of this idea? What are your thoughts on how to do all wheel drive for a powerful car that runs at high speed on pavement?
This car still exists. It came up for auction in 2013. A good description of it, and some of its history, with pictures, is still available on-line:
Zora Arkus-Duntov was heavily involved. According to the linked article, the idea was to produce the car with and without drive to the front wheels. The transmissions were manual 2-speeds, but with torque converters, as used in the first Chaparral 2. Indeed the two cars show similar thinking in a number of respects, although they share no parts except possibly the rear transaxle. Both have fiberglass monocoque construction. The very first Chaparral 2 had somewhat similar front end
styling, and similar individual exhaust stacks (only straight ones, not megaphones). The front end was redesigned early in the Chaparral’s development, due to problems with lift.
The front transaxle had a smaller torque converter than the rear one: ten inch rather than eleven. The car was tested extensively, using a variety of front and rear gear ratios. Duntov was reportedly trying to get 35% of the torque to the front wheels at low speeds and 40% at high speeds. Using a smaller torque converter at the front would have produced a higher stall speed for the front, and this would have caused the front to get a higher percentage of the power at high speed. However, it wouldn’t be road speed that mattered for this, but rather engine speed, since the torque converters drive the transmission input shafts.
I don’t know how the car scaled, except that according to the article the total was close to 1400 pounds. However, just looking at the layout of the car, front/rear weight distribution would have to be similar to others of comparable layout. The engine was an experimental small-block with aluminum block and heads and titanium rods, but the rest of the car was also unusually light. That would mean the two-wheel-drive version would have had around 60% static rear and the four-wheel-drive one perhaps 58%. Wheels and tires appear to be equal size front and rear.
Even on early ‘60’s tires, to avoid having excessive oversteer in a car that tail-heavy without larger tires at the rear, the front has to have sufficient roll resistance to almost completely unload the inside front wheel in limit cornering. This would create a problem getting any power down with the inside front wheel while cornering. Solving this would require tire sizes more nearly proportional to weight distribution, or else a front differential capable of generating considerable locking torque at very low throughput torque, which tends to create undesirable forces in the steering.
I would consider 35% of the power to the front to be reasonable for a car with 50% static rear weight, not a tail-heavy rear-engine car. For best results in low-grip conditions, we want torque distribution similar to static weight distribution. However, for high speed work on pavement, the percentage of power going to the rear needs to be at least 15 percentage points greater than the static rear weight percentage. This is partly to allow the driver to control the car with the throttle when cornering in a manner somewhat like a rear-drive car, and partly to allow for rearward load transfer under power.
If a car’s c.g. height is 15% of its wheelbase, 15% of its weight transfers rearward for each g of forward acceleration. With modern tires and all wheels driven, accelerations of well over 1g are possible, even without downforce. If we do have downforce, generally we want the percentage of downforce at the rear to be greater than the static rear weight percentage. Even if the car has two times its own weight in downforce, and enough power to use the tires to the full, the rearward load transfer under power will still be something like 15% of the tire loading from gravity plus aero.
Reasonable people may differ as to whether we’d like the front tires to break loose first under power, or the rears, but there is a strong case for not losing the ones we steer with first. In any case, we’d
like to be making nearly full use of all the tires, and we’d like to at least know which end we’re more likely to lose when we do hit wheelspin.
We also want the front wheels to help propel the car all the time, or at least whenever the rears are making forward force. The car will be fairly controllable if the fronts don’t drive at all until the rears reach a programmed percent slip, provided this is not a percent slip that we will reach in normal throttle steering, and provided that drive to the front doesn’t engage too abruptly. An on-demand four-wheel drive system that just has a computer controlled clutch for the drive to the front, and no center diff, can provide this if the programming is suitable. However, this way we are missing out on the gain in cornering power that all-wheel drive can deliver, and we are losing lateral acceleration capability due to the added weight of the hardware to drive the front wheels.
Full-time AWD increases cornering power because it eliminates the need to transmit the power that drives the front wheels to them from the rear wheels via the road surface. Power to drive the front wheels, in a rear-drive car? Yes! Why? Because when the front tires are running at a slip angle, as they must to make lateral force, they absorb considerable power. If there are no shafts driving them, that power must be transmitted to them from the road surface, and it must come from the rear wheels before that. This means that a portion of the friction circle of all four tires is being used to drive the front wheels, leaving less of the tires’ force capability for lateral acceleration. This is why cars with all-wheel drive will often achieve skidpad performance comparable to rear-drive cars weighing two or three hundred pounds less, with comparable tires.
At the same time, we probably would rather not have the front tires driving on decel, or at least not strongly. If we drive the front with just a viscous coupling, either the front wheels drive on decel or they drag a bit when cornering when they’re tracking on a larger radius than the rears, depending on tire sizes and gear ratios.
Probably the best approach is to use an epicyclic center diff that sends a fixed percentage of transmission output torque to the front and rear at all times, with the percentage of power to the rear at least 15 percentage points greater than the static rear weight percentage. There can also be some viscous limited-slip effect added to this. By juggling tire sizes and front and rear gear ratios, we can make the viscous influence gently drive the front wheels and correspondingly retard the rears when coasting if we wish, and we can control this effect. A pure viscous limited-slip at the front is probably desirable as well. At the rear, all the diff choices that work with rear drive are viable options.
To avoid needing to have the inside front wheel very light in cornering, tire sizes need to be matched to front/rear weight distribution.
Even when all of this is fully optimized, there will still be considerable cost, weight, and packaging penalties, compared to driving only the rear wheels. The car needs to be very powerful, and intended for a very upscale market, for the benefits to outweigh the penalties.