The peachtrees are killing my gas mileage!

October 9, 2007 10:30 AM

How can I maximize fuel economy in a car with an automatic transmission?

I know the general tricks about reducing weight, keeping tires inflated, being light on the gas and brakes, etc. (as discussed previously), so this question is targeted specifically at how the automatic transmission could factor into gas mileage. This is my first car with an automatic transmission (having driven manuals my whole life until now) and I am curious about what transmission setting I should use to maximize gas mileage. In case it matters, I have an 04 Honda Civic (not hybrid).

Having moved recently to Atlanta, I find myself in a lot of stop-and-go traffic and constantly driving up and down hills. This leads to a lot of up- and down-shifting when I leave the car in "D". It occured to me that perhaps I would be a bit better off by leaving the car in "D3" (one notch below "D") so it wouldn't shift so much. Is it better for gas mileage to keep the RPMs in the low-to-mid-range all the time rather than to be pivoting between low and high-ish RPMs? Is gas consumption directly proportional to engine RPMs?

posted by wondercow to Travel & Transportation (23 answers total)

Overdrive only saves gas when you're not in stop and go traffic. My mechanic told me to use D3 for traffic, and D for the highway.
posted by solipsophistocracy at 10:41 AM on October 9, 2007

The whole point of an automatic transmission is to keep engine RPMs in that optimal "sweet spot" for best efficiency based on load. Manually shifting it isn't necessary.

About the best thing you can do is to simply perform routine maintenance...oil, fluids, filters, etc. Keep your tires inflated properly (according to your owner's manual).
posted by Thorzdad at 10:53 AM on October 9, 2007

Is gas consumption directly proportional to engine RPMs?

I don't believe so. The amount of gas used more closely related to the amount of power you're using. Whether you're climbing the hill at 3000RPMs in 3rd gear (with less throttle applied) or 2300RPMs in 4th gear (with more throttle), you're probably using a similar amount of fuel. (Of course, you want to keep the engine in its efficient RPM range, which for a Civic is probably fairly broad.)

Beyond the standard fuel-saving tips, I don't think there's anything special you can do related to the automatic transmission.
posted by knave at 10:56 AM on October 9, 2007

I've heard that the ScanGauge II is a good way to improve your fuel efficiency. It connects to the engine computer and shows you the MPG in realtime, so you have direct feedback on how your driving style impacts it. OTOH, it's $160.
posted by smackfu at 11:20 AM on October 9, 2007

knave, I have to respectfully disagree. I've spent numerous hours tuning racecars, which involves establishing the big RPM vs. load matrices that a car uses to determine how much fuel to inject into the engine for optimal [power/efficiency/torque/life].

Without writing a treatise on engine tuning (I have neither the time nor the inclination today, but there's info all over the interweb), let me just say that the only way to keep fuel consumption down is to lower the revs. An engine is most efficient in terms of power-per-unit-fuel-consumed at peak torque (probably about 3000 RPMs in this case), but that doesn't matter a whit when you're talking about efficiency in terms of distance-travelled-per-unit-fuel-consumed. If you drive a manual transmission, shift as low as you can, even if it means going wide-open-throttle. In an auto, put it in D and baby the throttle. All the time.

It's all you can do.
posted by TheNewWazoo at 11:26 AM on October 9, 2007

Automatic transmissions in modern cars are controlled by computer to shift to deliver the best of all worlds. Tromp on the gas and it downshifts. Go light on the pedal and you'll get the best possible economy under the circumstances.

Unless your owner's manual specifies otherwise, leave it in D and let the car do the work.

I'm curious - what kind of mileage have you been getting with your Civic?
posted by SteveInMaine at 11:28 AM on October 9, 2007

Agreeing that you should leave it in D for all normal driving. Your owner's manual problably says to use D3 in certain situations only, such as towing or climbing steep grades or driving through heavy snow. You might find in those situations that the transmission is switching back and forth a lot between gears, and downshifting to D3 or 2 would eliminate that. Especially in snow that would give you better control, though you won't have that problem in Atlanta.
posted by beagle at 11:54 AM on October 9, 2007

@TheNeWazoo: An engine is most efficient in terms of power-per-unit-fuel-consumed at peak torque (probably about 3000 RPMs in this case), but that doesn't matter a whit when you're talking about efficiency in terms of distance-travelled-per-unit-fuel-consumed.

Your first statement is correct, but clearly more efficient power output means more efficient travel. All other things being equal, a higher power output means a higher speed, so more distance travelled per unit time. Therefore, maximum efficiency (in distance traveled) occurs at the same time as maximum power efficiency.

Secondly, shifting as low as you can in a manual transmission is silly if you have to go wide open on the throttle to achieve that. The reason is the same as above: if you go below the maximum torque level, efficiency drops off.
posted by ssg at 11:58 AM on October 9, 2007

Response by poster: Thanks all for your responses! Unfortunately, it sounds like I will just have to leave it in D and do my best not to get too antsy at stoplights.

SteveInMaine - I bought the Civic (used) this summer in Montana and was consistently getting 46-48mpg (way higher than the EPA rating!), even when a substantial part of the tank was not highway miles. Now that I am in Atlanta, my mileage has plummeted to around 24mpg (worse than the EPA rating) and I suppose I will have to chalk the difference up to the amount of time idling in traffic. Makes me wish I had hunted for a hybrid...
posted by wondercow at 11:59 AM on October 9, 2007

TheNewWazoo...really? So if I'm driving a 5-speed and crusing along at 70 in 5th, and approach a steep hill, it's better to "floor it" to maintain speed (bogging the engine since i'm down around 2k rpm) then to downshift to 4th, and easily clime the hill with 1/2 throttle, although at a higher rpms?
posted by lohmannn at 12:01 PM on October 9, 2007

Secondly, shifting as low as you can in a manual transmission is silly if you have to go wide open on the throttle to achieve that. The reason is the same as above: if you go below the maximum torque level, efficiency drops off.

Not to mention that going wide-open-throttle invokes a whole different set of engine-control parameters, like injecting extra fuel to keep the combustion chamber temperatures down.

I'm not a race car engine tuner, but I am versed in the application of common sense.
posted by knave at 12:02 PM on October 9, 2007

Wow, wondercow, that's a horrible drop. Maybe coincidental with the move, you got some kind of tuning problem that needs to be looked at? Consider checking our your air filter, fuel filter, spark plugs, wires, distributor cap & rotor (all cheap parts to replace). Also, you can use lighter weight motor oil in Atlanta than you would in Montana.
posted by knave at 12:07 PM on October 9, 2007

Also, you can use lighter weight motor oil in Atlanta than you would in Montana.

Sorry, ignore that, it's backwards.
posted by knave at 12:13 PM on October 9, 2007

ssg: Don't forget gearing.

lohmann: Generally, yes.

knave: Your "common sense" is wrong. There are no "whole different set of engine-control parameters". I know, I've set those parameters up.

Listen, here's how it works. There's a base matrix that controls how much fuel to inject based largely (and at times, solely) upon RPM and something called "load". Load can be either engine vacuum (mass-air flow calculation), air flow through a sensor (air-flow volume or mass metering), or throttle position alone (alpha-n, if you're keeping track). Excepting edge cases like cruising at constant speed (which incorporates a feedback from the O2 sensor, but at NO OTHER TIME except idle) and when knock is detected (which shouldn't happen if you've got good gas and a good fuel matrix), the engine measures RPM, measures how much air is being consumed, and looks up how much fuel to inject. That's it. That's the black magic that makes engines go. I should note that there's generally a fairly linear relationship between air flow through the engine and fuel injected, excepting turbocharged cars, but that's not what we're talking about here.

Now, if you're tooling along at 3000 RPM, your engine is going to inject a certain amount of fuel. The amount of fuel injected doesn't vary very much with load. That is, you can't greatly increase the amount of airflow by stomping on the throttle (again, provided that RPMs don't change). But if you double the RPMs, you'll almost certainly double the fuel injected. So, if you are at 2k RPM and injecting 1.5 ms worth of fuel under cruise, you won't be able to get above, say, 2 ms by stomping the throttle. If you double the RPMs but keep the same accelerative rate, you'll inject twice the fuel by virtue of frequency alone.

That was scattered, and maybe I'll come back later and write something a bit more lucid, but trust me; The math is all there. The best way to save gas is to keep the revs down.
posted by TheNewWazoo at 12:58 PM on October 9, 2007 [1 favorite]

The transmission can't anticipate loads--it only reacts to load, so I do most of the downshifting myself, then put it in the higher gear and let the transmission upshift when it's ready. Control is not just in the shifting, but in the throttle as well. Anticipating hills with a little more fuel and a downshift keeps the revs in the sweet spot and prevents that last minute lugging and burst of fuel that is so wasteful and hard on the engine.
posted by weapons-grade pandemonium at 2:57 PM on October 9, 2007

@TheNewWazoo: Don't forget gearing.

Gearing has nothing at all to do with it. For example, say you need 10HP to continue driving at constant speed. You can get those 10HP at 1500rpm, 2800rpm or 4000rpm, by injecting the necessary amount of fuel (assuming your engine can handle it). However, the engine is most efficient at 2800rpm (assuming that is max torque), so it makes most sense to run it at that speed.

I've driven around in my car with a laptop attached to watch the fuel consumption change with different driving. I do have a turbocharged car, but even at low RPMs (when the turbo is supplying very little boost), wide open throttle will cause the fuel consumption to jump by a lot (even a factor of 10!) I'm pretty sure the turbo isn't suddenly supplying 10 times as much air.
posted by ssg at 3:46 PM on October 9, 2007

My husband read an article on just this subject in some British car magazine about two years ago, and the main thrust was to keep the RPM below 2 at all times, 2.5 on the highway. In our old Volvo this puts us at around 69 MPH.

Other tips included turning the car off at long stoplights, and using neutral when going downhill and at short stops. I wonder whether that's bad for the transmission, though.
posted by tempest in a teapot at 4:10 PM on October 9, 2007

TheNewWazoo is correct. The maximum efficiency for an internal combustion engine is when the throttle is wide open and the RPMs such to provide the maximum torque. But cars are seldom operated at this optimum value so there are trade offs.

The engine is basically an air pump -- it sucks in air and pushes it out. The throttle is the means of controlling the speed of the engine. It does this primarily by restricting the flow of air into the engine. The greater the restriction, the more the engine has to work to suck air. Whenever the throttle is less than fully open, energy is wasted drawing in air against the throttle obstruction. This is the same whether going uphill, downhill or on the level. You only notice it more when going downhill because the engine is putting out less power to overcome the drag.

So you waste more energy when cruising in a lower gear with a partially closed throttle than in a higher gear and the throttle wide open. So yes, when going up a small hill, you are better off pressing harder on the throttle than downshifting, but only up until the point that the engine starts to lug. Only then should you downshift.

This is one of the reasons over-sized engines are so much less efficient in a car. They spend most of their time cruising with an almost closed throttle, struggling mightily to suck air through a soda straw.

People who drive gas endurance contests know that you shift through the gears at as low an RPM as possible and keep it in the highest gear possible at all times -- even if this means 35 mph in fifth gear with the pedal floored.

As far as the OP's question, just leave the transmission in drive. The transmission controller will try to maintain the highest gear.
posted by JackFlash at 4:26 PM on October 9, 2007

ssg: You're confusing brake-specific fuel consumption (BSFC) with miles per gallon - it's okay, we all do it. :)

Here's the thing - "peak efficiency" only matters when you're talking about WOT. That is, peak volumetric efficiency (VE) can only happen when the throttle is wide open. When you're dealing with partial-throttle situations, that all goes right out the window, for the reasons JackFlash outlined and a few others.

So for a given desired power output, considering that BSFC is inversely proportional to rev under part-throttle, you can see that keeping the revs as low as possible is the best strategy. There is, of course, a tradeoff with intake charge velocity and fuel atomization which results in more fuel needed for proper combustion, but that's talking about 1000 RPM full-throttle operation - outside the realm of normal driver technique.

Also, I excluded turbocharged cars from this discussion for a very good reason... because when you're seeing boost, even low boost, at low revs in a forced-induction situation, the AFR has to be REALLY low, or else you build up heat in the cylinder. This is because your combustion events take forever relatively speaking, and there's more time for the cool intake charge to soak up heat from the valves, and the intercooler is less efficient, and on and on. So you get the ECU dumping fuel into the cylinder to cool things off much more than a normally aspirated car ever would. I love you and all, but your observations just don't apply.
posted by TheNewWazoo at 4:51 PM on October 9, 2007

fascinating. sorry for the semi-derail OP. thanks tnw and jackflash, i feel like i learned more about engine performance in the last ten minutes than the past 2 years of hanging out with OMG BOOST kids around college.

i just bought a 2002 trans am, and now i'm beginning to understand why cags helps with gas mileage so much.
posted by lohmannn at 6:53 PM on October 9, 2007

Lohmannn, as you know, the Corvettes and Trans Ams after 1992 have a device called the CAGS (computer aided gear selector) installed on manual transmissions. It is intended as an emissions control device to improve fuel economy. What the device does is lock out second and third gears forcing you to shift directly from first to fourth gear under low acceleration conditions. It does this when the throttle is opened less than 21%, which would be the case if you are accelerating normally since the engine power is much more than normally needed. Since the engine is so powerful it can easily go from 1st to 4th without lugging, even at 15 mph, which causes you to open up the throttle and reduce the restriction. This is just one more example of improving fuel economy by keeping the gear high and the RPMs low in order to open up the throttle.
posted by JackFlash at 9:21 AM on October 10, 2007

Other tips included turning the car off at long stoplights, and using neutral when going downhill and at short stops. I wonder whether that's bad for the transmission, though.
posted by tempest in a teapot at 7:10 PM on October 9 [+] [!]

coasting in neutral is illegal in many states, including georgia.

i've heard switching to neutral at long stoplights will reduce transmission wear.

also from hearsay, air conditioner usage does play a large part in gas consumption-- in stop-and-go traffic, roll down the window. on the highway, use the ac. rolling down the window on the highway creates a lot of drag and kills any savings from not running the ac.
posted by Ziggy Zaga at 1:56 PM on October 10, 2007

I missed what tempest had said. In addition to being illegal, idling during deceleration actually uses more gas than engine braking, during which time the computer actually stops injecting fuel altogether. This is done for emissions reasons, and is usually transparent to the driver.
posted by TheNewWazoo at 2:12 PM on October 10, 2007