Drop unsprung mass, gain serious MPG

[goingplacesanddoingstuff]

If you're using the onboard computer, the tire size has made the mpg read outs inaccurate. Even calculation by hand would be difficult to equalize to the previous sized tires.

But if the tires are larger diameter, then that would serve to make the MPGs appear lower than they actually are.

 

[Hohn]

Hard to say because you went with a significantly narrower tire which has less friction. Why did you not choose say a 225/50R16?

The IS250 is a sport sedan. 205/60 16's are underwhelming

Underwhelming according to whom?

I would have gone with 185/70/16 if such a tire existed. But it doesn't that I could find, 205 is about as narrow as you can go in 16" wheels and have some options for profile and brand and overall design.

The narrower tire was a major point of the upgrade because it's hard to find narrower tires in larger wheel sizes. They should indeed have less rolling resistance.


What they will *not* have is less friction in terms of traction, at least not to the degree people mistakenly seem to believe. The formula for static friction has no term for surface area. It's just the coefficient of friction and the amount of force acting on it. In other words, larger contact patches do not translate to more traction automatically. Rather, the increased traction largely results from the softer compounds of "performance tires." So then you go to an all season-compound on a wide tire on a large wheel and you get 1) more weight 2) stiff sidewalls that are instable over rough sections 3) worse performance in terms of fluid dynamics-- passing through air, water, and snow.

Moreover, for a given amount of contact patch surface area, a longer narrower patch is preferable to a shorter wider one for my purposes. Short and wide is preferable on track days with high corner loads on a dry traction surface. For civilian car usage on public roads in varying weather conditions, tall and narrow is a better choice. As it is for straightline drag racing and it is for many offroad scenarios as well.

 

[kschachn]

Exactly. Which is why my initial observations took time to trust. Only after it was consistent across fuel stations, driving routes, and cold/hot engine type conditions did I come to the conclusion that the phenomenon was real.

But then again that is my point. It is never consistent.

 

[Hohn]

You can pay more money to add lightness in wheels, using the same diameter where available, such as Forged wheels ($$$) or Flow-formed wheels ($$).

And/or you can get LRR tires and also pick up some mpg as well.

Certainly. Boy did I really want to get some nice Enkei RPFs that are 15# each.

But alas, this is a near-beater commuter with 176k miles on it. I can't justify wheels that are $200+ each when the car itself is worth maybe $3000.

But a basic black steel wheel that's only $90 each? Well, it turns out that the fuel saving alone will pay for these wheels around month 22 in terms of ROI. That felt like somewhat low risk for cost recovery.

 

[goingplacesanddoingstuff]

Underwhelming according to whom?

I would have gone with 185/70/16 if such a tire existed. But it doesn't that I could find, 205 is about as narrow as you can go in 16" wheels and have some options for profile and brand and overall design.

The narrower tire was a major point of the upgrade because it's hard to find narrower tires in larger wheel sizes. They should indeed have less rolling resistance.


What they will *not* have is less friction in terms of traction, at least not to the degree people mistakenly seem to believe. The formula for static friction has no term for surface area. It's just the coefficient of friction and the amount of force acting on it. In other words, larger contact patches do not translate to more traction automatically. Rather, the increased traction largely results from the softer compounds of "performance tires." So then you go to an all season-compound on a wide tire on a large wheel and you get 1) more weight 2) stiff sidewalls that are instable over rough sections 3) worse performance in terms of fluid dynamics-- passing through air, water, and snow.

Moreover, for a given amount of contact patch surface area, a longer narrower patch is preferable to a shorter wider one for my purposes. Short and wide is preferable on track days with high corner loads on a dry traction surface. For civilian car usage on public roads in varying weather conditions, tall and narrow is a better choice. As it is for straightline drag racing and it is for many offroad scenarios as well.

This. There are complex interplays between traction, width, and pressure per square inch of contact patch. Wider doesn’t always provide more traction and better performance.

 

[Hohn]

But then again that is my point. It is never consistent.

Well, it is and it isn't. It's consistent in the way that matters to me, which is trending and averages-- regression to the mean is a thing, and with enough fuel burned from enough different stations and enough different uses, the variability converges on something like a true average.

SO when you see a change immediately, the sample size is much too small for any convergence to occur and you can dismiss the results to the point-to-point variability which is real and which you correctly mention will occur.

It's only after the datapoints converge on the new nominal that you can start to trust that your new nominal is truly different than the previous one. Think in terms of T-testing/Z-testing and sample size. At some point, the sample is large enough that the error terms shrink enough that you can justify concluding the new nominal is different than the prior.

 

[kschachn]

Sure. Care to share the details of that statistical analysis with us?

 

[goingplacesanddoingstuff]

Well, it is and it isn't. It's consistent in the way that matters to me, which is trending and averages-- regression to the mean is a thing, and with enough fuel burned from enough different stations and enough different uses, the variability converges on something like a true average.

SO when you see a change immediately, the sample size is much too small for any convergence to occur and you can dismiss the results to the point-to-point variability which is real and which you correctly mention will occur.

It's only after the datapoints converge on the new nominal that you can start to trust that your new nominal is truly different than the previous one. Think in terms of T-testing/Z-testing and sample size. At some point, the sample is large enough that the error terms shrink enough that you can justify concluding the new nominal is different than the prior.

Exactly. If you were to graph the results over time, after sufficient sampling size is achieved, you’d see a distinct shift in the median value before and after the tire/wheel change.

Reducing width and cutting significant weight from a vehicle’s tires can absolutely increase MPGs, all other factors being equal.

In your case, other factors could include the rolling resistance of the new tires and the inflation pressure.

 

[kschachn]

Exactly. If you were to graph the results over time, after sufficient sampling size is achieved, you’d see a distinct shift in the median value before and after the tire/wheel change.

Reducing width and cutting significant weight from a vehicle’s tires can absolutely increase MPGs, all other factors being equal.

In your case, other factors could include the rolling resistance of the new tires and the inflation pressure.

Graphing (observing) something is the easy part. Attributing what portion of that observation is due to which variables is the hard part. Here the OP is attributing all of the observation to one variable which isn't warranted when there are many variables. Even you note this.

 

[Hohn]

In your case, other factors could include the rolling resistance of the new tires and the inflation pressure.

Certainly. The new tire pressure is surprisingly close to the old because the contact patch ends up being nearly identical in size at the same pressure. (the narrower width is offset almost 1:1 by a longer patch).

There's probably less rolling resistance too, but I'd have never expected that a softer-riding tire with more sidewall would have LESS rolling resistance.

Which leads me to conclude that a major element of the rolling resistance is in the tread area-- not just the sidewall.

 

[Hohn]

Graphing (observing) something is the easy part. Attributing what portion of that observation is due to which variables is the hard part. Here the OP is attributing all of the observation to one variable which isn't warranted when there are many variables. Even you note this.

While this is technically true, when all the other variables are constant and have regressed to their respective means, they are essentially controlled. In a regression model, you'd throw those out of your model because the coefficients would be too small.

It's easy to say "muh other variables" when you can't point to WHICH of the other variables. Neither can I, because literally nothing but the wheels and tires have changed. I buy fuel at the same 2-3 places year round, drive in the same weather where I live, and generally have a pretty good feel for how this car behaves after 9 years of ownership and commuter use.

But I'm sure it's possible a total stranger with no experience with this car in this area with my duty cycle using my fuel sources with my driving style has secretly discovered the mysterious X factor variable that explains the observed MPG increase and discredits the attribution of the gains to the only thing that has changed on the car in 3 years.

But I wouldn't take that bet.

 

[edyvw]

I'm not sure I'd want to be contentious over something as subjective as ride quality. I'll be my own counsel as to what constitutes hard riding and you can be yours.

These are not run flat tires, which past experience soured me on.

True, it is subjective.
Why not 205/55R16 which are replacement for 225/45 R17? Plus, much better availability.

 

[IndyIan]

While this is technically true, when all the other variables are constant and have regressed to their respective means, they are essentially controlled. In a regression model, you'd throw those out of your model because the coefficients would be too small.

It's easy to say "muh other variables" when you can't point to WHICH of the other variables. Neither can I, because literally nothing but the wheels and tires have changed. I buy fuel at the same 2-3 places year round, drive in the same weather where I live, and generally have a pretty good feel for how this car behaves after 9 years of ownership and commuter use.

But I'm sure it's possible a total stranger with no experience with this car in this area with my duty cycle using my fuel sources with my driving style has secretly discovered the mysterious X factor variable that explains the observed MPG increase and discredits the attribution of the gains to the only thing that has changed on the car in 3 years.

But I wouldn't take that bet.

Small driving habit changes can make a pretty large difference in mpg and if it stays up for several tanks, you've either found efficiency from the tire swap, or now consistently drive more efficient. Also you've changed several factors at once, so its hard to pin down what actually is causing the increase in mpg. But in the end, the main thing is that your mpg has gone up, and your replacement tires are probably cheaper too.

 

[goingplacesanddoingstuff]

Certainly. The new tire pressure is surprisingly close to the old because the contact patch ends up being nearly identical in size at the same pressure. (the narrower width is offset almost 1:1 by a longer patch).

There's probably less rolling resistance too, but I'd have never expected that a softer-riding tire with more sidewall would have LESS rolling resistance.

Which leads me to conclude that a major element of the rolling resistance is in the tread area-- not just the sidewall.

The meaningful variables that changed are thus:

• Wheel/tire weight
• Tire rolling resistance
• Tire width

Reducing any of these variables will serve to increase MPG. It’s hard to determine which has the largest effect in your case, but I believe you when you say there has been a significant improvement.

 

[edyvw]

I don't bother hand calculation with this car anymore because the onboard computer was consistent. It was consistently wrong of course, but the offset was repeatably ~-1.2mpg (i.e. it was a bit optimistic vs true) over 3 years and I stopped bothering with hand tracking (well, using an app, but whatever).

Also, hand calculation has variables that people often fail to account for-- like the degree to which the tank is filled. Unless you fill up at the same pump at the same station facing the same direction at the same time of day every time-- your hand calculations are also prone to error because of the invalid assumption that the definition of "full" is always the same for each tank. It isn't. Even if you top off and "fill" your tank, the amount of air in retained in the tank (as it is even when you top off to spilling out the neck) will vary.

Outside of a dyno and lab conditions, the best any of us can do is come up with a useful approximation of true MPG. And the observed gains I'm seeing are well outside the measurement noise of my somewhat crude measurement system. Indeed, it's exact *same* measurement system and it's indicating gains despite the minor size difference which would cause error in the opposite direction (tend to show MPG loss).

8lbs on such vehicle is BIG thing. There is no doubt major contribution to higher mpg is less unsprung weight. Toyota and Honda for example always went smaller rotors on family SUV’s, minivans to bump mpg.
On Atlas for example, I switch from 255/50R20 to 235/65R18 snow tires which hardly have better rolling resistance. But gain in is 2-3mpg’s.

 

[meep]

It takes a lot of guts to experiment with tire and wheel sizes - they aren’t cheap. Good on you for doing the math and taking a shot at it. I’ve gone counter-culture myself on a couple of vehicles with smaller wheel / taller sidewall and less commonly, reduction in width. I’ve also done the opposite with bigger wheels/tires and 4x4 lifts.

in vehicle lifts, I’ve always lost 2-3 mpg with the lift and tire combo. My lifts were between 1.5 and 2”, never anything more than barely noticeable, and I kept my tires inside the wheel well. The lift hurts aero and the tire does the rest. Rough guesswork was 40% aero and 60% rolling losses on the two vehicles I did the 2” increase on, as I did each step at different times.

on my Lexus GS, I gained about 1 solid mpg simply reducing the wheel from an 18 to a 17 while maintaining the same diameter, width and offset.

a 205 width on yours, also, I find perfectly acceptable, and could very well be a more appropriate width for the size and weight of your car. 225 looks sweet but IMO even the slightly larger GS doesn’t actually need all that.

I don’t think all of your mpg increase is from weight. There could be a revs-per-mile component, less heat lost in the suspension from vertical motion due to more sidewall action absorbing irregularities, as well as reduced air drag (both from vehicle motion and simply spinning in the wheel well stirring the air).

Not to mention, that’s got to be a significant improvement in NVH due to unspeung weight. either way, enjoy the increase!

 

[GRUMPYDUDE2]

Light and narrow will make a diff..

 

[edyvw]

TIRES, WHEELS AND BRAKES ARE SPRUNG WEIGHT, NOT UNSPRUNG

Really?

Unsprung vs Sprung Weight

Lightweight? Forged? Flow Formed? You need new wheels but you're confused as to what the hype is surrounding wheel weights and construction right? We take a deep nerdy dive on why lightweight wheels matter on your UTV or ATV. Valor Offroad engineers lightweight UTV Wheels that can improve your...

www.valoroffroad.com

 

[GRUMPYDUDE2]

Really?

Unsprung vs Sprung Weight

Lightweight? Forged? Flow Formed? You need new wheels but you're confused as to what the hype is surrounding wheel weights and construction right? We take a deep nerdy dive on why lightweight wheels matter on your UTV or ATV. Valor Offroad engineers lightweight UTV Wheels that can improve your...

www.valoroffroad.com

I am wrong.sorry about that,

 

[Hohn]

True, it is subjective.
Why not 205/55R16 which are replacement for 225/45 R17? Plus, much better availability.

I wanted more sidewall. I've had pinch flats on this car due to potholes (again, commuter duty) and going down to 16" as well as going slightly taller on sidewall are both directionally helpful with that.


This size was the lightest overall setup I could get in what I felt was the tallest sidewall I wanted to risk. 55 profile was never under consideration, but 65 was.