Understeer gradient of modern passenger cars

[hutch325]

I do a fair amount of vehicle testing on the commercial vehicle side of things, and along with that comes lots of rental cars used for travel to the test facility.

Lately I've noticed that it seems like modern passenger cars don't have much understeer, at least not as much as I would expect. In particular the FWD cars seem more neutral than they used to be.

Some examples: Buick Lucerne and Mercury Milan. They seem pretty neutral in quasi-steady state maneuvers above about 70 mph (think a fairly fast interstate exit ramp) and seem to possibly even have some oversteer in some transient situations (think a steering reversal, sine or step input like a quick lane change).

Granted my evidence is anecdotal because I only drive these cars on public roads so I'm unwilling to do anything too daring.

Thoughts? Am I way off base?

 

[JsTyLz]

It is also pssible that your "feel" is affected by the electronic steering assist. In some cases it can make a motor home feel like a sportier suv.

 

[hutch325]

Oops, I accidentally posted this in the wrong forum. Relocate maybe?

 

[cibachrome]

There are very few evaluators who can discern the actual level of understeer in a road vehicle, much less linear range understeer. What they probably react to is the rate of change of understeer per g. Certainly the steer behavior or the sounds from the tires at or near the limit of control is the usual verbage called 'understeer'. Clearly, most 'car' magazines subscribe to this description and comment more on the linear range and steering gain of the vehicle, not on its system dynamics. In today's world, even the method used (constant speed, constant radius, constant steer) to measure a vehicle's understeer can produce different values.

In fact, most cars made in the past 10 years have 2 to 3 degrees per g of understeer. (N = many hundreds) Sort of bi-modal with sportier types low 2 and others 3-ish. Trucks are more towards 3 and higher. Virtually no vehicles below 1 deg/g (well maybe 1 or 2). Thus, no "neutral" cars (sorry, but with good linearity they feel neutral by the magazine translation. The impressions from "feel" things should not be confused with understeer although they are cross linked. Steering torque or force is a tuned deal, with high efforts from hydraulic or electric steer tending to raise understeer a tad because of compliance reactions.

So, I'd vote no historical changes. Steering gains are higher because of lower gearbox ratios, and efforts are changing because of a market begging for more 'feel' via mechatronics. If anything, better tires would allow lower understeer in vehicles with low payload changes (like 2 seaters) but the gain change with speed begs for speed and g sensitive steering systems to make the driving experience pleasant. Better tires usually means higher rolling resistance so ....

BTW: I changed the understeer of my speedboat in a measureable way without it feeling much different using 3 diffrent propellers. it was 4 deg/g, then 3, then 1. The 1 deg/g prop pushed it into oversteer at the 'limit'. But it didn't feel different just acted different. So it even was confusing to me. In God we trust, all others: bring some data...

Also, the use of terms like "transient understeer" will kill the SAE audience because it knows only steady state analysis algebra. We need a better word definition to decribe the transient character of front and rear sideslip behavior. I know what you are talking about, but my collegues (including some retired and some in their graves) roll over and go deaf when these words appear. Be careful out there...

 

[hutch325]

Thanks, cibachrome. Excellent information.

I agree with "...to all others, bring some data..." I have real problems getting accurate subjective assessments from test drivers (meaning their assessments usually don't agree with the data).

Unfortunately, there's no way I'm going to set up the DAQ equipment on a rental car and sneak it into the test facility just to satisfy my curiosity.

Perhaps you are correct, and a majority of what I percieve really is resulting from lower gearbox ratios.

I take it in your last paragraph, that you don't necessarily disagree with the concept of "transient understeer" but some that are confined to the world of linearized, steady-state models would? I was thinking in terms of a yaw rate command by the driver, calculated based on the vehicle speed, wheelbase, steering characteristics, etc. The driver commands a yaw rate and the vehicle yaw rate lags the commanded value, thus for an instant, understeer is occurring, correct?

 

[cibachrome]

You can measure a rental cars understeer in a K-mart parking lot witha VBox using a fixed steer angle method.

A better (in my opinion) control system sequence for a vehicle is that you command a moment on the vehicle. The moment is reacted by steering and tire forces and tire moments. The vehicle goes to whereever it needs to go to resolve this. You don't see any position based coefficients in the equations do you? So, depending on the speed you are going, the vehicles path can be a turn or a sideslip or both. What you don't like is turning at high speed or sidesliping at low speed. Just the right amount produces the best evaluation and the optimum trajectory. At high speed the lateral acceleration lag is the most prominent and twice as big as the yaw rise time. In the past, vehicle dynamicists used understeer to quicken the yaw velocity and lateral acceleration rise times and ignored the resulting settling time and overshoot increases. Some people praising "neutral" steer cars were stymied by the slow response of first order dynamics and used very high cornering stiffness tires to get a decent rise time at the expense of ride and Expense(!). My John Deere handles really good on low profile, shallow tread depth, wide rim R4 tires just like my boat handles well on a 5 bladed ported stainless steel propeller. But, the cost and replacement cost can kill you financially. In the U.S. the legal system caught on to the low underteer issue via 'forseeable misuse'. So spare tires, overload, replacement tires, worn tires must be factored into the understeer recipe.

As far as "transient understeer" is concerned, though, most of us know what we are talking about as drivers, it just doesn'e play well in the steady state classroom playbook. Theoretically the understeer elements appear in the denominator of the yawrate and sideslip transfer function in only one place. The steady state gain term. However, there are an infinite number of front & rear axle sideslip gain values which can make up this understeer net. This also ignores tire transient response behavior. And, virtually all redneck discussions about tire transient response view it as a first order response. This is just not the case. Makes for good Excel math models, just not good correlation.

"You know you are a redneck if you've ever been been in a discussion about truck tires that lasted more than a hour." Jeff Foxworthy...

 

[hutch325]

VBox and a parking lot. After I posted my response, it clicked with me that I didn't need to measure steering angle if I used the constant steering method. I do have access to a VB20SL that could provide some very interesting data.

And yes, I see your point about the understeer gradient only appearing in the denominator of the yaw rate transfer function in one place. This is wholly consistent with the vehicle dynamics models my company uses. So I suppose a better way to describe "transient understeer" may be "yaw rate lag" or "steering - yaw phase relationship" or simply "transient response characteristic".

 

[GregLocock]

Agree with the 2 - 3 deg/g thing. A heavily laden light truck on the other hand can be down to 1 deg/g.

SS linear range understeer is, according to some internal documents, almost impossible to /evaluate/ directly, yet plays a crucial role in defining the feel of the vehicle. I'd like to see someone entangle exactly what that means.

My rule of thumb is, don't even try and get the rest of the linear range steering right if the u/s gradient is wrong, and be prepared to lose up to 0.5 deg/g during the development of the vehicle. Tires, of course have an enormous effect, we'd often see 0.2 deg/g between two consecutive submissions of the same tire, and more than that between submissions from different companies.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[cibachrome]

Vbox or equivalent process for fixed steer understeer test:
Measure speed and yawrate or speed and lateral acceleration while SLOWLY increasing speed. Be careful about shifting gears so pick them wisely. Do this for several radii. Then from ay compute yawrate. A Vbox actually only measures steady state g so its perfect. Then in Excel or Matlab or whatever, make up a function with lateral acceleration on x axis and curvature (yawrate/speed) on y axis. Compute the slope of these multi-radii functions in both directions. The understeer is the negative slope of this function times the wheelbase in radius units. There is a pi thing in there, 2.

The only question I had at the time was what the wheelbase of a boat is. I came up with my own value, what do you all think?
As far as tire submissions, we measure all the submissions before they are screwed on a car. The fun actually begins when compring our tire measurments with the tire company's own measurements. Even more fun is having a tire company measure another tire company's tire. Sounds tiring doesn't it.

Since the understeer defines the steering gain of the vehicle (g/deg), then the amount of steer, the g's, tierod loads, compliance steer and steering booster assistance that result are all mixed up in the feel of the vehicle. Since a .5 deg/g variance means more to a 1 deg/g understeer car or truck than it does to a 3 deg/g vehicle, The low understeer situation is usually avoided unless you like all your cars of the same model to feel different.

No question about laden trucks often having understeer less than 1 deg/g. In fact they will be oversteer after some moderate g level (maybe .3 g). Don't panic, a vehicle is still stable in fixed control until the amount of oversteer exceeds its Ackerman gradient at the speed you are running. That's also why you really don't want to be in a hurry with a spare tire on the back axle.

 

[hutch325]

Thanks, cibachrome. I've done the fixed steer test many times so I already have the necessary Matlab functions. I just usually use different instruments than the VBox.

So what was your method for coming up with the wheelbase of a boat?

 

[GregLocock]

You can get an effective wheelbase by considering the steer angle to get a certain radius, at very low speed.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[cibachrome]

Didn't do the low speed L/R thing for wheelbase because the effects I'm interested were above 40 knots. I presume I had a variable wheelbase vehicle. Mainly used a second observer to tell me where the hull contact points were.

The only reason I mentioned a VBox is because I consider it an affordable tool for this stuff. A reconditioned VBox-I is in the hobby class of devices that serve this purpose. I often now refer to it as the fish finder. The Pi users are upset by this...

 

[gt6racer2]

On the original question, is it not the case that now we all have ESC ( yaw control ) that the cars can be made to respond more agressively, to give the "low g" sense of sportiness, and then avoid the spinout in transients through the ESC ?

 

[GregLocock]

That's possible, but customers don't like aggressive ESC intervention.

So, using ESC to compensate for unusual handling characteristics is probably not a great move.

Having said that, spinning in transients is a pretty unusual failure mode on dry roads or even wet roads. I don't think I've ever spun a car except on a circuit.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.