2008 CanadianDriver 50-litre Challenge

[chrischasescars]

It would have been easier to start from the "low fuel" light. These are very accurate on new cars.  Consider the "low fuel" light as empty.  Put in X number of liters.  If it's less than 50, when the light comes on again, add the leftover and drive till light comes on again.  More accurate than the way you guys did it...

I see what you're getting at. I think the problem with doing that way would be having a way to accurately dispense the right amount of fuel en route. What would be great, in all seriousness, would be to have a small tanker convoying with us, which we could use to top the cars off, as you suggested. I think this would be a little impractical, if only for the fact that it would likely be expensive. We need a fuel sponsor for next year, maybe. 

Paul's explanation does admit the technique we used isn't scientific, but we were consistent from car to car in driving off the estimated amount of fuel to get each car down to 50 litres.

The way we'd like to be able to do this test is to run each car dry, fill them all with 50 litres from the same pump and then go again. Problem is that manufacturers usually supply cars with a full tank; give that we got most of them just a few days before the test, it would have been tough for the four of us who live in Ottawa (where we held this event) to drive the eight or nine cars with larger-than-50-litre tanks dry so that we could refill with the required 50 litres.

[chrischasescars]

9th generation Corollas were known to hold close to 58 liters.

What? My dad has a 9th gen 'Rolla, and he thought for the longest time the car only had a 45 litre tank, cos he could never get more than 42-43 litres into the thing, even when he ran the car till the gas gauge was pegged on "E."

On my '05, when the low fuel light lit, I'd put only 40 liters.  I once drove 104 km on the highway with the light on and put 47 liters in.  An act of faith I would'nt have done had I not known beforehand that my car could swallow more than its stated 50 liters.

So to conclude, this test is more of a range test where the cars with the bigger tanks stand more chances of winnning although I fully admit the Corolla and Civic to be the more economical of the group.

Can you show me where you found out that 9th-gen Rollas can take 58 litres of gas? I have a hard time believing that.

[rrocket]

I see what you're getting at. I think the problem with doing that way would be having a way to accurately dispense the right amount of fuel en route.

Ummm...don't the fuel pumps have readouts where you live?   

And barring that, I don't think carrying a gas can with the "X" number of remaining liters to add would be unreasonable, since for most cars it would be under 10...

[chrischasescars]

I see what you're getting at. I think the problem with doing that way would be having a way to accurately dispense the right amount of fuel en route.

Ummm...don't the fuel pumps have readouts where you live?   

And barring that, I don't think carrying a gas can with the "X" number of remaining liters to add would be unreasonable, since for most cars it would be under 10...

Ummm...do gas stations magically appear where ever you want them to where you live?    How the hell do we ensure we just conveniently pull up next to a gas pump right when the low fuel light comes on?

?
?
?
?

[rrocket]

Ummm...do gas stations magically appear where ever you want them to where you live?    How the hell do we ensure we just conveniently pull up next to a gas pump right when the low fuel light comes on?

Certainly wouldn't have been any more vague than the way you guys burned off fuel.

Like I said...using a gas can with 10 liters in the trunk wouldn't have been terribly difficult....And this way, you could have also used cars with fewer than 50 liter tanks to show us the best, most frugal car each company offered...

[vdk]

Okay really... I say fill up the tank... and drive... calculate the fuel economy from there... X liters burned in Y kms... XL/100km... tada!
It's alot easier.

[rrocket]

Yea, exactly....but you have to know how many liters are being burned...which is why I say start from the "low fuel" light..

[Hyksos]

How we made our calculations

Five of the cars in our group have 50-litre capacity fuel tanks. Although some of the cars have 51, 53 or 55-litre tanks (one had 59), we only used 50-litres for our "Challenge." The range of these vehicles (Volkswagen City Golf, Mitsubishi Lancer, for example) would certainly be extended by their larger fuel tanks, had we started filled them to capacity.

Our technique to equalize the tanks was to fill all the tanks, then "drive off" the required number of litres based on the combined city/highway rating for litres/100 km as supplied by the manufacturer to Transport Canada. For instance, the Dodge Caliber has a 51-litre tank; the combined L/100 km rating is 8.24 which is equivalent to 12.1 kilometers. Consequently we "drove off" 12.1 km before starting the event in the Caliber. We used the same method for each of the vehicles with larger-capacity fuel tanks.

Note: We understand that our method is not scientifically defensible. In order to cover all possible factors that impact an event like this would be beyond our resources. In fact, Transport Canada uses a simulation to generate its ratings precisely because of the numerous factors that can affect results. We feel that we have used the same methodology for each of the vehicles, driven them on the same route on the same day, and offer our results with some certainty that they are representative, and accurate to a plus or minus two-percent margin of error.

.... I must say that I'm shocked .... reading your article really let me think that you guys really only had 50L per car ...

That is about the worst way to make this test accurate. I even wonder why you even considered using GPS to make the test more accurate when you totally destroyed any accuracy using such terrible technique to know that you should have "about" 50L gas in the car ....

Tank capacity and real capacity is totally different, all the plumbing can increase a tank capacity considerably, so you might think that this car only have 50L tank but you can actually fill 54L into it when it's 100% dry. Or it can have "air bubble" on top of the gas tank so you can't fill it full without insisting, continue filling even when the gas pump stop until all air bubble are out.

The way we'd like to be able to do this test is to run each car dry, fill them all with 50 litres from the same pump and then go again. Problem is that manufacturers usually supply cars with a full tank; give that we got most of them just a few days before the test, it would have been tough for the four of us who live in Ottawa (where we held this event) to drive the eight or nine cars with larger-than-50-litre tanks dry so that we could refill with the required 50 litres.

Then how can you know the car came with a full tank of regular gas?? If manufacturers fill them with supreme the car can do more mileage.

Next time you should simply use a manual gas pump (15$ value) to empty all the car then fill them with exactly 50L.

[wing]

I think we should just fill them with 1L of gas and see how far they go, much more accurate since it's only 1L and would take a lot less time!

[initial_D]

Interesting. Sideshow is better the main attraction.

[UmroAyyar]

I think we should just fill them with 1L of gas and see how far they go, much more accurate since it's only 1L and would take a lot less time!

Now, you are thinking like Jeremy Clarkson. To make a point about fuel economy, they did exactly that with supercars in Top Gear. I like your thinking. 

[JBrill]

How we made our calculations

Five of the cars in our group have 50-litre capacity fuel tanks. Although some of the cars have 51, 53 or 55-litre tanks (one had 59), we only used 50-litres for our "Challenge." The range of these vehicles (Volkswagen City Golf, Mitsubishi Lancer, for example) would certainly be extended by their larger fuel tanks, had we started filled them to capacity.

Our technique to equalize the tanks was to fill all the tanks, then "drive off" the required number of litres based on the combined city/highway rating for litres/100 km as supplied by the manufacturer to Transport Canada. For instance, the Dodge Caliber has a 51-litre tank; the combined L/100 km rating is 8.24 which is equivalent to 12.1 kilometers. Consequently we "drove off" 12.1 km before starting the event in the Caliber. We used the same method for each of the vehicles with larger-capacity fuel tanks.

Note: We understand that our method is not scientifically defensible. In order to cover all possible factors that impact an event like this would be beyond our resources. In fact, Transport Canada uses a simulation to generate its ratings precisely because of the numerous factors that can affect results. We feel that we have used the same methodology for each of the vehicles, driven them on the same route on the same day, and offer our results with some certainty that they are representative, and accurate to a plus or minus two-percent margin of error.

.... I must say that I'm shocked .... reading your article really let me think that you guys really only had 50L per car ...

That is about the worst way to make this test accurate. I even wonder why you even considered using GPS to make the test more accurate when you totally destroyed any accuracy using such terrible technique to know that you should have "about" 50L gas in the car ....

Tank capacity and real capacity is totally different, all the plumbing can increase a tank capacity considerably, so you might think that this car only have 50L tank but you can actually fill 54L into it when it's 100% dry. Or it can have "air bubble" on top of the gas tank so you can't fill it full without insisting, continue filling even when the gas pump stop until all air bubble are out.

The way we'd like to be able to do this test is to run each car dry, fill them all with 50 litres from the same pump and then go again. Problem is that manufacturers usually supply cars with a full tank; give that we got most of them just a few days before the test, it would have been tough for the four of us who live in Ottawa (where we held this event) to drive the eight or nine cars with larger-than-50-litre tanks dry so that we could refill with the required 50 litres.

Then how can you know the car came with a full tank of regular gas?? If manufacturers fill them with supreme the car can do more mileage.

Next time you should simply use a manual gas pump (15$ value) to empty all the car then fill them with exactly 50L.

I agree.  Though I don't think the results would have been significantly different in vehicle ranking, distances were probably within a +/- 25km range of error based on amount of fuel in the tank.  However, as you have pointed out, it would have been easy enough to siphon the tanks dry, run them around the parking lot until they stalled, and then filled them with 50l.

The bigger concern is individual driving techniques.  As an example, my wife seems to be incapable of coaxing more than 500km out of a full tank on the minivan.  Yet, when I drove it for several weeks, I had no problem getting 600-700km on a full tank doing similar city driving (even though I'm a much heavier user of AC  ).  We simply have different driving styles.  I'm 100% sure I could probably have made any of the vehicles achieve 100-200km better results.  I know there was probably a pep talk to all drivers before starting, but there is no way driver behaviour for acceleration, slowing to a stop, etc... can be taken out of the equation unless it's the same drive who tested all the cars.

In the end, the test did show that on average, the Toyota and Honda surpass the competition in this area, it's a nice publicity stunt, but should not be taken as scientific fact.  Buy the vehicle that meets your needs.

JB

[mrthompson]

Could you weigh each vehicle to determine how much fuel was consumed (knowing the density of gasoline)?  Perhaps +/- 1 lb. is adequately precise?

[airbalancer]

but they change drives every 100km

[mrthompson]

Weigh each driver and allow for that in the calculations, before and after their time in the car.

[johngenx]

Why don't they just invent an atomic measuring device that can break it down to the subatomic particle level so we can more precisely measure how much gas is being used?

As each driver's weight changes during the course of the day, they will be forced to either consume liquids to compensate for weight loss, or have appendanges amputated to compensate for weight gain.

The fuel will come from one source and be carried in a tanker along the route.  To ensure no tampering takes place, highly trained special forces soldiers will be hired to observe the tanker, the cars, and the drivers.

The driving style of each driver will be monitored by computer for throttle position, braking usage and steering angle deflection.  As each car moves over a determined section of highway, this data will be compared, and drivers deviating by more than 2% of the standard determined by the average of the group, will be "removed."  (by the Special Forces team...)

All cars will have their OE tires replaced with identical models.

To ensure no "ringers" from manufacturers, all cars will be car-jacked from little old ladies fitting an identical demographic to ensure consistency.

...And most of you will still rag on about the results...

[initial_D]

^^ That works for me.

I think Wing shoulda just write a simple simulation program that can run on a Blackberry. 

[airbalancer]

You could install a GE TransPort PT878 flowmeter http://www.gesensing.com/products/pt878.htm?bc=bc_ps+bc_flow_pana_port  on each car to measure the gas flow
Everyone carries one in their truck, right

[libraman]

So the testers assumed that the measurements of the tank size was accurate and that the burn off of fuel was accurate.  

With respect, the easiest way to proceed  would be to drive a known distance, (500km?) and then refill the tank. Use the same gas for all cars. Refill at the same station. This would reduce the variables. Then do the math.

The test does not bother me. What bothers me are the misleading statements/conclusions in other articles which are based on the faulty test. See for example:
http://www.canadiandriver.com/ctc/blog/sedans/2009-toyota-corolla.htm  and
http://www.canadiandriver.com/50litre/tech-opinion.htm

If you use the data from the Consumerreports tests for example you can say that the Corrolla is more efficient but you can't say that the Civic is significantly more efficient than say, a Mazda 3.

[chrischasescars]

Why don't they just invent an atomic measuring device that can break it down to the subatomic particle level so we can more precisely measure how much gas is being used?

As each driver's weight changes during the course of the day, they will be forced to either consume liquids to compensate for weight loss, or have appendanges amputated to compensate for weight gain.

The fuel will come from one source and be carried in a tanker along the route.  To ensure no tampering takes place, highly trained special forces soldiers will be hired to observe the tanker, the cars, and the drivers.

The driving style of each driver will be monitored by computer for throttle position, braking usage and steering angle deflection.  As each car moves over a determined section of highway, this data will be compared, and drivers deviating by more than 2% of the standard determined by the average of the group, will be "removed."  (by the Special Forces team...)

All cars will have their OE tires replaced with identical models.

To ensure no "ringers" from manufacturers, all cars will be car-jacked from little old ladies fitting an identical demographic to ensure consistency.

...And most of you will still rag on about the results...

 

Thank you, John. Even if we had done this test according to r-r-r-r-r-rocket and librarian-man's suggestions, someone out there would question the results.