How well does "coasting" work?

November 11, 2005 8:38 AM

Does "coasting" in a standard transmission vehicle really save gas over maintaining your speed?

I don't know how many times I've heard this. One of my friends is low on gas so they say if they "coast" more then it will magically extend their gas millage enough to make it to a gas station.

For those who aren't familiar with the term, coasting is basically getting your vehicle up to speed and then pushing in the clutch effectively going into neutral and letting your cars momentum carry you for awhile. The idea is, since you aren't on the accelerator for that extended little bit of road then you are effectively saving gas.

It seems to be a pretty common belief but it always bothered me for two reasons. 1. it takes more energy to get your vehicle back up to speed then it does to just maintain your speed, 2. your vehicle is still using an amount of gas while in neutral just to keep the engine running.

So what do the experts say? Does coasting really save your gas? How effective is it, really?

Thanks!

posted by nickerbocker to travel & transportation (36 comments total)

Coasting on flat ground (or uphill, heh), no. Coasting down hill, yes.
posted by LordSludge at 8:45 AM on November 11, 2005

What LordSludge said, with the addition that coasting to a stop uses less gas than braking.

Braking turns gas into worn-out brekepads, the less you use your brakes the better your gas milage will be. And no, engine braking isn't a magic loophole, it just turns gas into a worn-out engine instead.
posted by revgeorge at 8:58 AM on November 11, 2005

I'm no expert, though I have toyed around with homebrew fuel injection for a while.

My initial reaction is that fuel consumption is proportional to load on the engine, so if there is any savings it would be minimal.
posted by kableh at 9:02 AM on November 11, 2005

heh. i was listening to a track on the mecha last.fm station and there was someone talking about a certain bearing, which i googled. anyway, fyi, there's this bearing, whose name i cannot remmeber, which whizzes round + round only when the clutch is depressed and the transmission moving. and freewheeling downhill with the clutch in puts wear on this bearing, since presumably it is intended for momentary use.

anyway, the singer in question had problems with this bearing. fwiw. after freewheeling downhill to save money (which is completely valid physics, btw)
posted by andrew cooke at 9:05 AM on November 11, 2005

oh, sorry - only valid for downhill. on the flat, you are right. steady state (maintaining constant speed in this case) is always(?) most efficient, all other things being equal. there must be a proof of that, but i can't see how at the moment.
posted by andrew cooke at 9:08 AM on November 11, 2005

andrew cooke; would that still apply if you release the clutch after moving the gearshift out of gear? The clutch would re-engage whatever piece you're talking about.
posted by odinsdream at 9:09 AM on November 11, 2005

Remember that when reengaging the engine after coasting down hill if your not careful you can put a lot of stress on the engine as well.
posted by alexst at 9:09 AM on November 11, 2005

The real bottom line here is RPMs.. How many RPMs your engine is doing is directly proportional to how much gas you're using - which seems (and is) pretty obvious.

If you can maintain speed without higher RPMs from your engine - you will use less gas. Therefore, the answer to your question is "sometimes, specifically when going downhill".

Now, the only place it gets complicated is when you throw in a transmission - whether it's manual OR automatic.

If you watch your tachometer (which measures RPMs - Revolutions Per Minute for the not so car savvy), you will see that it goes from low to high as you accelerate, and then back to low again as you shift gears.

You want to be at the low RPM "range" of whatever gear you're driving in to get the most out of your gas. It means the engine is doing as little effort as possible to maintain that speed.

Where it gets tricky, and this isn't related to coasting, is "what speed should I go to get the most out of my gas?". The answer to that question is a little more difficult because in 3rd gear at low RPMs, you may be going X miles per hour, and in 4th gear at low (but higher than the low end of 3rd gear) RPMs, you may be going Y miles per hour. You want the lowest RPM/x ratio possible -- if it takes 1500 RPM to maintain 35mph, and only 1600 RPM to maintain 50pmh -- you're getting way more speed (and therefore distance) for using just a bit more gas.
posted by twiggy at 9:13 AM on November 11, 2005

Oh I should also add:

Your engine is still running when you put your car in neutral, and your RPMs do not drop to 0, so consider that, too. If your engine "idles" at 1200rpm, and it only requires you to be at 1300rpm to be in gear and getting a bit more push to your car to maintain speed -- you may not really be saving gas.

This will vary greatly by car.
posted by twiggy at 9:15 AM on November 11, 2005

Ok... coasting downhill and coasting to a stop instead of breaking to a stop seem like reasonable gas saving tricks. The person I'm thinking of actually thinks he saves money by taking his car up to 45mph then letting it coast till it drops to 35mph and then taking it back up to 45mph and so on. That just seems ridiculous to me. Not only is it taking him longer to get to his destination (extending the time his vehicle is on and running) but accelerating the vehicle from 35mph to 45mph would be taking more gas then just maintaining the 45mph (not to mention pissing off anyone who is behind him). He drives a "boat" that gets horrid gas millage anyway.
posted by nickerbocker at 9:16 AM on November 11, 2005

Isn't this called "riding the clutch?" Neutral (or "coasting in neutral") means that the car isn't in gear *and* your foot is off the clutch. I do this alot, but not to save gas.

looks like opinions are mixed here, too.
posted by whatnot at 9:18 AM on November 11, 2005

The bearing andrew cooke is referring to is the clutch throwout bearing. And yes, keeping the clutch depressed will wear it out faster than normal use should. You also shouldn't keep your foot lightly on the clutch pedal at stop lights for this reason.

twiggy, RPM is just one part of the equation. 3000 RPMs on a flat stretch of road and 3000RPMs while going uphill arent using the same amount of energy, right?

You most basic EFI determines how much fuel to inject based on engine RPM and load. Load is typically determined by how much air is flowing into the engine, measureed either directly (Air Mass Meter) or indirectly (Manifold Absolute Pressure). Throttle angle can also be used as an indicator, tho this is a bit less accurate.

I dont know the math behind it all. It's tied to BSFC, linky here.

Nickerbocker, sounds like you're in the right here. Steady state is definitely going to save gas, since we get better mileage on the highway instead of in stop and go traffic.
posted by kableh at 9:20 AM on November 11, 2005

I always thought "Riding the clutch" is where you lightly place your left foot on the clutch so you don't have to keep on moving your foot on/off the pedal.
posted by jmd82 at 9:32 AM on November 11, 2005

"Riding the clutch" means to keep your foot resting on the pedal when the car is in gear. People think "I am barely touching it, how can it make any difference?" It does, by introducing even the most miniscule slip.

Resting your foot on the clutch when it is NOT in gear is at worst putting strain on the throwout bearing although I am personally unconvinced it could ever cause wear that would preceed the clutch plate's demise. I come to that opinion largely unencumbered by facts or research, however, so you're probably well off just not doing it.
posted by phearlez at 9:49 AM on November 11, 2005

here's a sketch of the argument i was trying to come up with it. it's not as general as i thought, but i believe it.

generally, when we think of a process, there's something that dominates costs. for driving a car at a reasonable speed, it's probably air resistance.

anyway, no matter what it is is, this thing typically varies as a high power of some variable in the system. for air resistance, it increases as the square of velocity (iirc). so going twice as fast loses four times as much energy.

now how far you get depends on velocity (speed). or, if you like, velocity raised to the first power (linearly). so going twice as fast quadruples your loss of energy, but only doubles the distance travelled (in a fixed time). so you end up losing out (losses go up by a factor of 4, gains by a factor of 2).

more generally, there's almost always going to be something that gets worse more strongly than the gains (which are typically linear, like distance with speed).

now over a whole journey, you want to minimise the total losses (total fuel used/available). so you want to minimise the integral of the losses, while keeping the required result (distance travelled in this case) constant. so, looking at velocity v, you want to minimise the integral of f(v) wrt time, given that the integral of v wrt time is a constant.

and if f(v) is dominated by higher-than linear terms then i'm pretty sure this has to be a straight line (more exactly, for a polynomial with +ve coefficients). but my maths isn't up to proving it at the moment.

but you can see the basic idea from the drag argument. any excess speed really hurts, because your losses go up more than your gains. so you want to keep the speed as low as possible. that means constant speed, assuming that you have to get there within some time interval (can't take an infinite amount of time).

however, this argument breaks down if the source of losses changes, because then f(v) becomes more complex than a polynomial with positive coeffs. or, indeed, if f(v) does anything other than increase with v. for example, you might imagine that there's a minimum speed below which it's very inefficient for an engine to run. in practice that's probably very low - like 5mph. but if you have almost limitless time then travelling at 1mph is going to lower losses (to air resistance at least). so it's likely that coasting until you stop, then accelerating to 5mph and coasting again could be more efficient than running at a constant 1mph. but, again, this is probably only at speeds lower than people actually do coast, human nature being what it is.

so, in conclusion, when a single source of losses dominates, the argument sketched above strongly suggests that a constant speed is most efficient. but there may be exceptions when the loss function becomes "complex", and one might be that engines run inefficiently at very low speeds.
posted by andrew cooke at 9:53 AM on November 11, 2005

I thought driving in neutral, or coasting, was illegal. I have used it in emergencies, like north of Lake Superior, where gas stations are a brazillion km apart, but in traffic you won't have the same control over your car, and one accident will more than wipe out your puny gas savings. Don't do it.
posted by weapons-grade pandemonium at 9:56 AM on November 11, 2005

BTW, I'm not so sure that engines run inefficiently at low speeds. The efficiency graphs are skewed towards the high end because low speeds are normally those used for acceleration, which does waste gas. If you're driving a constant 20 mph on the level, I would bet you'd get far better mileage than someone covering the same ground at 60 mph.
posted by weapons-grade pandemonium at 10:03 AM on November 11, 2005

nickerbocker, your friend is kidding himself. Coasting downhill uses less gas; coasting (and periodically speeding up again) on the flat uses more.

The controlling parameter for fuel consumption is throttle opening. If you can drive at a given speed in second gear with a smaller throttle opening than in third gear, then second is more thrifty. Your friend is interspersing periods of closed throttle (which is thrifty) with periods of large throttle opening (when he accelerates), which is wasteful. The wasteful in this case overwhelms the thrifty, when compared to steady throttle opening.

I don't think coasting is illegal, at least not everywhere. When you took your foot off the gas in old Saab 2-strokes, they went into freewheel, which is the same as coasting.
posted by Kirth Gerson at 10:12 AM on November 11, 2005

sure, wgp, (assuming you're adressing me), but at very low speeds, there's not going to be a suitable gear to keep the rpm high enough to work properly - whatever the speed is in first gear that's slower than the engine ticking over, for example. i have no idea exactly what it is, hence my guess above of around 1mph.
posted by andrew cooke at 10:13 AM on November 11, 2005

trying again, because what i wrote might seem like greek. here's an example.

drag increases as the square of speed, ok?

now let's decide to, instead of going at constant speed, go for a while 10% faster, and then for the same time 10% slower. so on average our speed stays the same, but we wobble it around a bit.

this is like coasting, but by making the change in jumps of 10% we make the maths easier.

now when we're going 10% faster we're getting 21% more drag (because it's square of speed, and 1.1^2-1 = 0.21, or just trust me). and when we're going 10% slower we're getting 19% less drag (1-0.9^2=0.19).

so overall, there's (slightly) more drag. 21% losses on the high speed bit and 19% gains on the low speed bit (and 21% > 19%). and this difference is going to come out whenever the losses (drag in this case) are speed "squared" or "cubed" or anything that's "stronger" than just proportional to speed itself.
posted by andrew cooke at 10:31 AM on November 11, 2005

Metafilter: I come to that opinion largely unencumbered by facts or research.
posted by OmieWise at 10:59 AM on November 11, 2005

If you push the clutch while going downhill you are actually using more gas then when you just roll downhill in gear without pushing the accelerator. Why? Because if you are rolling in gear and don't accelerate, the motor is kept running by the rolling car. Modern motors shut off the gas intake in this case - you don't use any gas at all. If you push the clutch, some gas is needed to keep the motor running.

(Rolling downhill without pushing the clutch is also recommended for steep declines, btw, because you use the motor as a brake and keep the brake pads from wearing out. Shift to a lower gear to brake harder.)
posted by amf at 11:03 AM on November 11, 2005

Omniwise, that's AskMetafilter: I come to that opinion largely unencumbered by facts or research.
posted by zpousman at 11:48 AM on November 11, 2005

Unless the hill is very steep, taking your foot off the gas will make you slow down, because of compression braking. You'll have to use more gas to get back up to speed. So no, declutching or shifting to neutral does not use more gas than taking your foot off the gas.

Related note: if you have a Subaru (and this may apply to other AWD cars) do not coast; it damages the power-transfer mechanism.
posted by Kirth Gerson at 12:04 PM on November 11, 2005

kableh: RPM is just one part of the equation. 3000 RPMs on a flat stretch of road and 3000RPMs while going uphill arent using the same amount of energy, right?

I must say this counfuses me*. But from the link provided by kableh, what I do find interesting is the idea that different grades/octanes of gasoline will affect power delivered from combustion therefore changing gas mileage. Am I reading this correctly?

*If the same engine is turning the same crankshaft connected to the same wheels by the same gearset -- how can it burn more gas for the same RPM's. Of course it would not be able to climb said hill in same gearset… *head explodes*
posted by Dick Paris at 12:13 PM on November 11, 2005

Re: Not coasting in a Subaru.
This does not make sense to me. It would seem the transmission would be designed to be back driven same as any other car (not saying your wrong just it does not make sense where the damage would occur.)
posted by Dr_Octavius at 12:16 PM on November 11, 2005

But to answer the question, on a flat coasting would not save gas. It has always been my understanding that cars are designed so that the maximum efficiency (in terms of miles per gallon of gas) is in the speed range of 40-50 mph with larger cars on the lower end of this range. This is because low speed efficiencies are canceled out by the peripheral loads on the engine (fans, alternator, pumps.)
posted by Dr_Octavius at 12:27 PM on November 11, 2005

If accelerating and then decelerating repeatedly were fuel-efficient, then EPA mileage figures would be higher in the city than the on the highway instead of the other way around (except for hybrids, of course, which actually turn off the gas engine while stopped and are assisted in accelerating by electric motors, which in turn recharge the batteries by acting as a dynamo during deceleration). It's continually having to accelerate back up to a given speed that uses the most fuel, and as pretty much everyone has said so far, your friend is wrong.
posted by mr_crash_davis at 12:48 PM on November 11, 2005

Doc_Ock, it isn't the Subaru transmission, it's the transfer cases that move power from slipping wheels to wheels with traction. The cases are filled with a special grease that gets more viscous when there's wheel slippage, transferring power to other wheels. (This explanation is probably a gross oversimplification.) Subarus are not supposed to be towed on their wheels, or emission-tested on a dynamometer, for the same reason.
posted by Kirth Gerson at 2:51 PM on November 11, 2005

Years ago, I hitched a ride in an old pickup where the driver claimed if he could shut his engine off for longer than 30 seconds he was saving money. He would coast engine off, for about 5 miles, right into his driveway.
posted by pointilist at 2:52 PM on November 11, 2005

How many RPMs your engine is doing is directly proportional to how much gas you're using - which seems (and is) pretty obvious.

Uh, no, actually. This is the trouble with kids these days; these fancy electronic fuel-injection drive-by-wire gizmos insulate the driver from what he's actually doing when he presses the throttle.

Let's go back to the carburated engine for a moment. From top to bottom:

Air intake and fuel pump -> carburator.
Carburator -> intake manifold.
Intake manifold -> intake valve -> cylinder head.

Basically, the first downstroke of the engine cycle, as the sealed pistons move down in the cylinder, creates a vacuum. At that moment, the intake valve is open, creating a conduit to the intake manifold. The carburator sits atop the intake manifold, mixing air and gasoline and spraying it into the manifold, where the vacuum sucks it down into the cylinder. The carburator is a passive pump - the vacuum created by the cylinder causes it to work. The 'throttle body', which is linked to the 'throttle' pedal, obstructs air from entering the carburator; this prevents the carburator from spraying very much air-gas mix into the intake manifold. When you tromp on the pedal, the plates in the throttle body open up and allow air to flow unobstructed.

Result: air and gas flow freely through the carburator into the cylinders. This causes more gas to be burnt per engine cycle, which causes the engine to produce more power and, given a constant load, to accelerate in RPM.

Also result: you can hook a vacuum gauge up to the intake manifold and measure what's going on in there. High vacuum = high instantaneous fuel economy. Low vacuum = unobstructed throughput of gas/air = poor instantaneous fuel economy. It so happens all the cars I've ever driven have been equipped with a vacuum gauge, and you can see that when you declutch and let off the pedal, the engine enters a high vacuum, high-economy state.

I have been an obsessive-compulsive monitor of gas mileage at points in my life. Some data in around-town driving:

1965 Pontiac GTO, 4-speed standard transmission:

Clutch-in coasting whenever possible; minimal acceleration: 12 mpg.
Driving sanely without paying attention: 8 mpg.
Keeping throttle floored at all times, burning rubber, shredding clutch: 4 mpg. 2 speeding tickets per tank.

1995 BMW M3, 5-speed automatic transmission:
Popping into neutral whenever possible, minimal acceleration, 'economy' mode: 14 mpg.
Driving sanely, not paying attention: 13 mpg.
'Sport' mode, driving like maniac: 12 mpg.

Point being, it doesn't make much difference, and with modern computerized engine management, it makes even less difference.
posted by ikkyu2 at 4:02 PM on November 11, 2005

But it's fun!

I think anybody learning to drive a standard transmission should first have experience riding a bike with gears, so you'd understand what's going on with the engine. Don't you always coast, on the bike, going downhill? Why? 'Cause it uses less energy! More on coasting -- I've heard it's illegal, 'cause you have less control, but c'mon. And,

Isn't this called "riding the clutch?"
This question just came up on on Car Talk; they said those who believe this are idiots. phearlez is correct, and they went into detail, as to why: increased wear -- sometimes, a lot.
posted by Rash at 4:19 PM on November 11, 2005

Sorry but I think that is a ridiculous idea. You will end up using more gas by doing this "coasting" bullshit.

First of all, when you are driving normally down the street without depressing the accelerator pedal, all cars made in the last 10 or 20 years or so go into "fuel cut" mode which means that the fuel injectors do not run AT ALL and you use NO GAS WHATSOEVER. This works because the momentum of the car will continue to turn the engine, so it does not need to use any fuel to do so.

But if instead of just taking your foot off the gas pedal you depress the clutch, you disengage the engine from the drivetrain. Now the engine will come to a halt unless some gas is used to keep it turning, just as it is necessary at idle when you are at a stop. So here you are actually using some small amount of gas where you were using precisely NONE before.

There is also the argument of steady-state. Fuel is used in a larger proportion the harder you accelerate, and you will find that you get the best possible fuel efficiency -- everything else being equal -- if you maintain a constant speed without speeding up or slowing down. By doing this ridiculous thing of putting in the clutch, you just cause yourself to have to constantly slow down and speed up, which will take more gas than just maintaining speed.

Finally there is the throwout bearing argument, which says that if you constantly keep the clutch disengaged (pedal always depressed) you will prematurely wear out the bearing that is used to hold the clutch in that state. And clutches can be expensive.

So for the love of god, if someone tells you to disengage the clutch and coast to save gas, tell them they are full of crap.
posted by Rhomboid at 9:35 PM on November 11, 2005

And I forgot to mention:

Someone will probably try to say "but by disengaging the clutch and coasting, the car will not be slowed down as much due to the drag of the engine, so it will be more efficient." I've got news for you: unless you plan on turning the engine off completely, it will still be running and causing that drag one way or another. You can either "pay" for that drag by using the momentum of the car with the clutch engaged, or by consuming gas and firing the spark plugs with the clutch disengaged. Either way the engine is turning, and the energy to do so has to come from somewhere.
posted by Rhomboid at 9:43 PM on November 11, 2005

I'm surprised no one has mentioned Ecodriving (pdf) yet. A few highlights are: accelerate at half-throttle up to 3000RPM (ie, your engine's maximum torque point); skip gears if you can; at constant speed, use your highest gear; coast to stoplights - and hope they turn green as you're coasting.

The training manual doesn't mention alternating between coasting and running at maximum engine torque as a strategy.
posted by mediaddict at 11:18 PM on November 11, 2005

Some folks here need to study up on compression braking, because they clearly do not understand it.
posted by Kirth Gerson at 4:54 AM on November 12, 2005