The unreported downside to higher gas prices

[msmercenary]

Each time the price at the pump goes up a nickel, I know about it. It's always in the news around here because Western WA and OR have some of the most expensive gasoline in the U.S (as I write this, premium is in the $3.55/gallon range at the Chevron station near work). But there's a hidden downside to high prices that never hits the media

Comments

[chalain]

Without trying to say that this isn't a hassle, I just want to say that this is very cool. I mean, from a decidedly nonsubjective viewpoint.

Think about it as a computational problem. The mechanics of the pump can only track and resolve at a certain rate. We tend to think of this as a closed transaction without a time component: we bought X gallons for Y dollars. But look at it from the viewpoint of dispensing gasoline fairly: if the pump has a fixed resolution for dispensing fuel, and it meters volume/time rather than price/time, then as the scalar for price increases the more gas the pump gives you for free. I would presume that we're talking teaspoons here, but if this really were the case I could easily see this getting exploited--somebody would post a lifehacker video of them "stuttering" the gas into their tank and getting 20 gallons for like $5. And even if not, then I can still see beancounter accountant whining that they pumps dispense 10.01 gallons for the price of 10.00, which is 0.1% in lost profits. For Exxon-Mobil that's $347M in free gas in 2006. So I can see them throwing $60,000 at a programmer to "fix" it at the expense of consumer time.

By scaling down by gas price, the oil company makes sure that they get paid fairly for their fuel. What I want to know now is if you 'stutter' gasoline, do you still get an accurate reading? Or does it unfairly penalize the consumer? I think about all the times I get 29.87 in fuel so I'll sit there jiggling the handle to try to get an even 30.00. Am I short-changing myself and getting say $0.10 of fuel for $0.13? I bet the pump testing people don't test for that.

[jerith]

My proposed solution: Make the flow rate constant (volume/time) but let the pump dole that out in packets of price. If you push the lever you get a minimum of $0.01 worth of petrol or something. Then again, this is probably too fragile to work in the real world. I'm going to go back to metering bandwidth instead of petrol...

[kazriko]

I'm not sure how consumer gas pumps work, but the metering we do of fuel on the pumps that put gasoline into the tanker trucks that service the gas stations has one big pump with valves. The pump can operate at variable speeds but as long as the rack is on, the pump is keeping the pressure up on the line. You open the valves to the dispensing arm and then meter it after it goes through the valves, and the meter measures the gas flowing to the tank regardless of how fast it is flowing.

I suspect they have something similar, a separate pump that can operate at variable speeds to keep a certain pressure and a separate flow meter that measures the fuel going out to the dispensing nozzle. Why? There's a sticker on the side of all the pumps around here saying that they're accurate at any pressure or speed... If they relied on counting how long the pump ran then multiplying that out, it wouldn't be accurate at any pressure or speed.

I'm also not sure about the instrument, but the common choice would be a fluid turbine meter that pulses for every X volume. If you intermitted it, that partial teaspoon would probably turn that turbine meter wheel a certain percentage of the way around to its next pulse, where the next teaspoon would finish pushing it around to the pulse.