The Science of …Fuel Intake Problems

Drew Donnelli asks: Can you offer any insight into the problems JGR seems to have with fuel supply?

Thanks for the question, Drew. As usual, with the help of a couple incredibly patient friends, I can provide a little information on the possible causes of the fuel intake problems that probably lost Denny Hamlin the Bristol race. Here’s a flowchart for the path the fuel takes from the fuel cell to the engine.

fuel flow

There has to be a constant supply of fuel for the engine to run properly. The sign of a fuel pickup problem is when a car that had been running just fine all of a sudden–often on a restart–sputters like it’s out of gas, but then recovers and continues.

Let’s start where the fuel starts, which is in the fuel cell. The picture below is from the ATL catalog. ATL is one company that supplies NASCAR-legal fuel cells and they have a very helpful catalog.

ATL Fuel Cell

The fuel cell is a rectangular box filled with foam to prevent the fuel from sloshing around too much. There are one or two fuel pickups inside the box that are sometimes called “duck feet” because they are shaped like, well, duck feet. The pickups have doors at their entires so that fuel is trapped when it enters the fuel pickup.

The duck feet are located on the right-hand sides of the fuel cells on oval tracks. A car turns because the tires exert a force that prevents the car from going straight. That force (the centripetal force) points toward the center of the turn. The problem is that, because you are in the car, you feel like you’re being pushed towards the outside of the turn. Some people refer to this as a centrifugal force, but there is no force pushing you outward–it just seems that way. (I belabored this point in the book, so I won’t go into it in detail here. Suffice it to say that unexpected things happen in non-inertial reference frames.) The same phenomenon happens to the fuel: It tries to go straight while the car is turning, so the net effect is that the fuel is pushed toward the right-hand side of the fuel cell, hopefully, right into the pickups. The doors prevent the fuel from coming out again on the straightaways.

This strategy sometimes backfires on the more steeply banked tracks. Liquids always seek their own level. Gravity acts on them and doesn’t care what they are contained in. I’ve shown a drawing below of what the fuel would look like if the car were just sitting on a banked track. Amusingly enough, this is actually a picture I was asked to draw during my doctoral comps. The request was to show what happens to water in a glass that is tilted, but the picture is the same regardless of whether it is a glass of water or a fuel cell.

When there is a lot of fuel in the tank, this isn’t so much of a problem, as the top picture shows; however, when the fuel level gets low (as in the bottom picture), even the turning action of the car may not get enough fuel into the fuel pickup, causing the engine to sputter. This would be even more of a problem on a restart because the force with which the gas enters the pickup changes with the square of the car’s speed. Under caution, the cars are going more slowly, while the force of gravity is the same regardless of the car’s speed. It becomes a tug of war and if gravity wins, the driver whose car is experiencing this battle often loses.

A second problem is when there is something in the fuel line that isn’t liquid and/or isn’t fuel. Fuel vapor, air bubbles or water (remember second Atlanta last year?) all cause problems if they get in the fuel line. One of the more common culprits is vapor lock, which occurs when liquid fuel becomes vaporized before reaching the cylinder. Under caution, fuel is in the fuel lines for a longer time because it is not being used as quickly. The inlet lines that bring fuel to the engine are close enough to the exhaust that they get hot. If the fuel is heading to the engine quickly, there isn’t enough time for the heat to affect it significantly; however, some of the hydrocarbons in the fuel vaporize at relatively low temperatures. As one of my engine experts put it, “Vapor is difficult to pump!” (I need to check–I think the unleaded fuel, which has higher concentrations of toluene, is more likely to vaporize at lower temperatures than the old, leaded fuel did.)

Fuel vapor makes it difficult for the fuel pump to properly regulate fuel pressure, which again leads to sputtering on the restart. The problem is compounded when air gets into the fuel lines because, in addition to messing up the pressure regulation, combustion requires gasoline as well as air (in a pretty picky ratio).

UPDATE: In response to Lou’s comment (and I had to put this in the blog because the comments section wouldn’t let me add the link!):

I should have mentioned that the fuel pumps on NASCAR cars are mechanically driven for safety reasons. You don’t want the fuel pump on a crashed car to continue to pump fuel. The Waterman fuel pump, is a cable-driven pump that is mounted on the fuel cell (as opposed to on the engine). This type of fuel pump is run on the Chevy R07 engines, but I understand that the 2008 Toyota engines (which I assume Hamlin would have been running) also use this type of pump. I’m also told that Waterman originally was the only one offering such a pump, but that other manufacturers (like CV products) have started providing this type of fuel pump. One argument in favor of mounting the fuel pump on the fuel cell is that it lessens the chances of vapor lock; however, if you have any type of pressure regulating device near the engine that is expecting liquid and gets vapor, you’re still likely to have issues with fuel supply. Another positive of this type of fuel pump is that it should be more efficient in getting fuel out of the fuel cell, although I don’t have any testimonials as to whether that is actually the case or not. The web link to Waterman racing Components above has very useful pictures and information about the fuel pump.

The big question, of course, is why this would affect some teams and not others. How much fuel is in the fuel tank, the type of fuel pump (some are mounted back by the tank and others are mounted by the engine), and even the way components are mounted relative to the effective net force (i.e. the vector sum of centripetal force and gravity) can impact whether or not the driver suffers from a fuel intake problem on the restarts.

UPDATE: Jayski has a link with information from JGR through ESPN that mentions some of the same issues. One might ask why JGR seems more susceptible to the problem. It may be that we only take notice when it happens to a car that is in the front. Other teams may be experiencing similar issues, but we just haven’t noticed. Sorry to hear that this is making Denny age prematurely!

6 thoughts on “The Science of …Fuel Intake Problems”

  1. The Chevy R07 engine utilizes a cable driven mechanical fuel pump that has to push fuel to the motor. These cable drive fuel pumps are notorious for not wanting to prime and restart which aggravates any fuel interruption condition. Just about all the Cup teams now use that type of fuel pump (Waterman??)

  2. Of corse JGR is not going to announce the amount of fuel that was left in the tank. My guess is that the G/W/C finish is as much a problem as the steep banking, there was little fuel left.
    Using your drawing, imagine if the 2nd drawing had just 5% of the gas you showed, closer to what happened at Bristol.
    I also have to wonder about the effect of slow re-start–the state of the fuel at the green was probably like your drawing (only much lower fuel level). When you get into the first corner, fuel runs to lower right corner of tank. The real solution has to be to put one of the duck feet in the lower left of the tank? And always figure Cup will have GWC overtime, put more fuel in the tank!
    Thanks for publishing some real information about NASCAR, keep up the good work!

  3. The information above is some of the most accurate I have seen to date on what could have happened at Bristol.
    There are a number of factors that affect to fuel delivery at low levels, including the type of “collector” in the corner of the tank, the type of body on the fuel filter, the pump itself, the line diameter, heat, humidity, and the best factor yet, the venue. Also a factor in the design of a fuel cell collector itself is the driver style. It was not too long ago that a certain two car team had different collector styles to complement the two different driver’s styles. This team won three championships in the current decade, and know how to solve these little issues.

  4. Hi LD71:
    I consulted with Dr. Andy Randolph, my engine guru and he confirmed that it is legal to have more than one pickup in the car. The problem, Andy says, is that pickups are always picking up. So around the corners, when the fuel is sloshing to the right (in the reference frame of the fuel tank), a left-mounted fuel pickup would be sending air to the engine.
    Teams aren’t allowed to have anything electronic in the fuel cell, so you couldn’t get around this by having, say, a valve and a sensor that would tell you if you have liquid or vapor in the pickup. Perhaps there is a clever way of accomplishing the same thing mechanically…

  5. diandra,
    one of the issues we contend with when using two pickups is that one is exposed the pump will typically choose the path of least resistance and move air rather than fuel. You can duplicate this experiment today at lunch, just use two straws in your soda. Try to separate them and pull one from under the surface out into the air and apply vacuum…. See how much liquid refreshment you get?
    That’s the major reason I discourage folks from running two pickups, one to each rear corner.

  6. Gary: Thanks for your comments. Your website ( has some great information about fuel cells and safety. Your experiment about using two soda straws is a great way to illustrate why multiple pickups can actually make a situation worse rather than better. Thanks again for the comments.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: