Gas Mileage Tips
The internet has much information on getting optimal fuel mileage. What follows are the tips I feel are relevant to circa 1988-2000 Hondas. Sometimes it costs a lot to remedy low fuel mileage. Sometimes it costs nothing but a little labor. At other times, conditions are such that the fuel mileage will go down, and the owner/operator can do nothing about it. For example, in colder temperatures or at highway speeds, fuel mileage will go down compared to warmer temperatures or speeds around 50 mph. Also, starting sometime in the Autumn, some parts of the United States switch to a lower heating value of gasoline. This may reduce mileage but only around 3%. Things you can do that cost nothing (or almost so):
  • Check the tire pressure every two weeks and particularly during season changes. Pressures higher than those specified by the manufacturer may result in better mileage. My 91 Civic tires are supposed to be maintained at 26 psi, but since about 2005, I have maintained them higher: 32 psi for summer and 35 psi for winter. This yields better fuel efficiency at the expense of a more comfortable ride, which I do not really notice anyway. It is a Civic, after all. Two caveats: (1) Traction, especially in slick conditions such as snow, ice, and rain will be worse with higher tire pressures; and (2) I am not sure whether higher pressures are meaningfully detrimental to tire wear.
  • Check the PCV valve. Wrap the hose connecting to it with cloth, and squeeze there with pliers to cut off flow. Listen for a click, which should happen within 30 seconds. No click, then spend the bucks and replace the valve. Click, and the valve may or may not be operating optimally. Remove the valve, and spray carburetor/PCV valve cleaner into it, or try PB Blaster or WD-40. Re-install. The engine control system relies on a particular metered flow through this valve for different operating conditions. If something's clogging it or the spring is old, then it will not work correctly, and the engine computer may be fooled into admitting incorrect fuel amounts. In 2003 I inspected, for the first time, the PCV valve on my 91 Civic. At 138k miles and never having been touched before, it was chock full of waxy buildup. I replaced the PCV valve, and the mileage went up significantly. Now I remove and clean the valve at every oil change. I will probably replace it every five years or 60k miles.
  • Check the level of the coolant in the reservoir, per the owner's manual. Consider doing an air purge of the cooling system. Certain engine control components are cooled by this coolant. Wrong levels mean improper cooling, which means sub-optimal operation. If you recently bought a used Honda, strongly consider a complete coolant drain, flush, and refill, using only OEM coolant or Havoline Orange Dexcool.
  • Check the ignition timing. This assumes you have a timing light. A used one will cost around $15. You want the center red mark on the pulley or flywheel lining up as exactly as possible with the fixed pointer. I experimented with this a bit in December 2005. I am convinced that being off more than two degrees from the center red mark will have a noticeable effect on fuel mileage. I believe the three red marks cover four timing degrees on my 1991 Civic. 

Things you can do that cost something, but not too much, and are probably worth the money for a Honda you bought used and do not know well:

  • Replace fuel filter and air filter.
  • Replace distributor cap, rotor, ignition wires, and plugs with OEM ones. I would not bother with platinum plugs unless the owner's manual says to. OEM wires and plugs, for one, will also help preserve the life of the ignition coil.
  • Wait until the gas tank is near empty and you are a week or so away from an oil change. At the gas station, add a bottle of the fuel injector cleaner "Chevron Techron," and fill the tank. (Note: I called the 800 number on the bottle's back and asked if I could add the Chevron Techron after filling the tank. The representative said this would be fine. I think the directions on the bottle are written so as to maximize the probability that the dilution ratio is optimal. I understand Chevron Techron is one of two fuel additives that the magazine "Consumer Reports" says actually can make a difference. I tried STP fuel system cleaner in late 2005, and it seemed to foul my PCV valve very quickly. I think this is more evidence for why one should add these fuel system cleaners about a week before an oil change.
  • Ensure the correct viscosity oil is used in your engine for your season and climate.
  • Make sure the thermostat is OEM and, say, less than 12 years and 150k miles old.
  • Consider a new oxygen sensor. A bad one will not necessarily set off a "Check Engine Light" code. I replaced my 1991 Civic's pre-emptively at about 160k miles, 13 years, based on reading at internet sites. The oxygen sensor sends a signal to the car's computer, setting the amount of fuel injected into the cylinders. Only buy an OEM-make sensor. This would be either Denso or NTK. Amazon has the best deals on Densor and NTK oxygen sensors. I have used Amazon for oxygen sensors three times and have never been disappointed. Look up the correct model number of oxygen sensor for your Honda at or the NTK site.
  • Consider a new distributor housing. The housing holds three sensors: The CYP, CKP, and TDC sensors, inputting signals to the engine computer. My sense is the sensors do degrade over time. My Honda's mileage shot up after it received a new, OEM distributor housing, and I think it may be because the old sensors had wore down.
  • Under investigation, November, 2008: Consider a new TW sensor (a.k.a. ECT sensor). This senses engine coolant temperature and sends a signal to the engine's computer. The signal controls the "basic discharge duration" of the injectors. I have a report that changing it out, even with a junkyard one, may make a big difference on older cars. Testing it first can also throw some light on its operation. See testing TW sensor, circa 93 Civics. This sensor is located on the engine block under the dizzy housing. It is not the only sensor there, so make sure you have the right one.
  • Check that the fuel pressure meets the specs given in the shop manual. A failed fuel pressure regulator can cause excessive pressure, causing the car to run rich. A non-stock pump may also cause the fuel pressure to run high. In 2009 Harbor Freight was selling a perfectly good fuel pressure gage for about $10. It screws into the service bolt fitting on the fuel filter. The shop manual directions for checking fuel pressure are easy to implement.

To purchase OEM parts at competitive prices, see the online parts companies listed at Maintenance & Repair Resources.

Many of the items listed above are routine tune-up items. These must be replaced every few years anyway for optimal engine performance. The issue becomes: Does one spend the money on parts now and have better fuel mileage, et cetera right away? Or does one spend the money later, suffering through worse engine performance and likely greater wear and tear on other more expensive engine parts?

For a discussion of optimal driving speed, see . Note that the typical passenger car drives most fuel efficiently when cruising between about 40 and 55 miles per hour.

Personal Fuel Mileage Report

I did not track mileage until early 2003. From about 2003 to 2007, my 91 Civic achieved 40+ miles per gallon. This was in states that do not use "winter fuel" (a.k.a. an ethanol blend) and driving a combination of highway and suburban. This mileage is consistent with others' reports for this relatively tiny, 1.5 liter engine Honda. My Civic's mileage dipped to maybe around 38 mpg, worst case, when it was really cold.

In late 2007 I began driving in a state where the gasoline has a minimum of 10% ethanol. Also I started doing more city driving. My Civic was averaging about 35 mpg. Internet reports say the ethanol gas should cause a mileage decline of not more than around 3%, so I figured the increase in city driving was the problem. Then in August 2008, my Civic failed (by 2.4%) its NOx emissions test. The most likely causes of the poor emissions, and what I did in response to them, are as follows.

  1. The EGR system. Response: None. My Civic does not come with an EGR system.
  2. Carbon deposits in cylinders, especially for older, high mileage cars. Response: Italian tuneup. Ran at high speed on long, steep hill for about an hour. Revved some to red line. Anecdote to counter this theory: On Sep. 20, 2008 the Car Talk guys said this advice applied to carbureted engines only. Fuel injected engines do not experience any serious carbon buildup on pistons or cylinders.
  3. Incorrect timing. Response: None. I check the timing often. It is fine.
  4. Speculation by me: The distributor housing was filthy and so the sensors were not working optimally, messing up the timing, fuel mileage and emissions. I wiped out the housing and blew compressed air through it to get rid of dust. I base this theory on getting a new distributor housing in 2003. I also put in a new PCV valve both in 2003 and in mid-2008. The fuel mileage shot up in 2003 and subsequent years, but then went down again for a year.
  5. Bad catalytic converter (though not excusing what caused it to go bad.) Response: I replaced the 17-year-old cat with a Magnaflow direct fit converter that I bought online from "Direct fit" here means it has OEM flange connections and so bolts right on;  no welding. It cost about $94. It is not "high flow." It fits fine. I bought the inlet and outlet gaskets on the net, too. I did find the inlet gasket would not fit and ended up re-using the old gasket. I bought new nuts and bolts from a local True Value store, too, since the old ones would not work with the aftermarket cat.
  6. Car not warmed up enough at time of test. Response: I thought this might have been the problem until I had the second emissions test done. At the first emissions test, the testers took a long time getting the numbers. The testers seemed studious about it. With the benefit of 20/20 hindsight, now I think they were trying to get my cat converter etc. hot enough so my car would pass, and they almost were successful. For the re-test a few weeks later, I drove in summer heat for an hour prior to the test. The test guys took like 30 seconds to get the now fantastically low emissions numbers. It was way faster getting in and out of the test facility.

Interestingly, following all these actions, my 91 Civic's mileage is back up to 40+ mpg. The emissions troubleshooting above also seems to have corrected the mileage decline. claims my 1991 Civic should get 27/32 mpg (city/highway, air conditioning assumed, 2008 method). Discrepancies with the government site are usual. Mileage depends on driving habits, the condition of the car and climate. These will vary a good deal from one person to another.