E85 on a Subaru WRX and STI: What It Takes to Run It Right

What E85 Actually Is

E85 is a spark-ignition fuel made of fuel ethanol and gasoline. The name means a target of about 85% ethanol and 15% gasoline, but the gasoline fraction is there on purpose. Straight ethanol does not vaporize well when the engine is cold, so the gasoline portion is what makes a cold start possible. Pump E85 is also not a fixed recipe. Refiners lower the ethanol percentage in cold weather, the same way they change gasoline volatility between summer and winter. A tank that says E85 can legally sit in the low 70% ethanol range in winter and closer to 80% or higher in summer.

That matters more on a Subaru than the pump label suggests. A tune written for 85% ethanol does not automatically stay safe when the station is pumping a winter blend, or when you top off with premium because the E85 station was closed. The car cares about the percentage in the tank right now, not the percentage printed on the sign.

E85 vs. Pump Premium at a Glance

Why Turbo Subarus Like It

A turbo engine is limited by air, knock, and heat. E85 helps all three, which is why it became a performance fuel on WRX and STI builds instead of only a fleet fuel for flex-fuel vehicles.

• Octane. Pump E85 lands around 100 octane, and ethanol’s blending octane when it is added to gasoline is higher still, near 118. That is race-fuel behavior at a pump price, and it is the main reason a turbo car can hold boost that 91 or 93 would knock on.
• Charge cooling. Ethanol absorbs more heat as it evaporates than gasoline does. The intake charge in the cylinder is cooler, so the engine can fill more effectively and is less likely to detonate.
• It wants to be rich. Gasoline makes its best torque modestly richer than stoich. Ethanol keeps making torque at much richer mixtures. You burn more fuel per pound of air, and at those mixtures you release more heat energy in the cylinder even though each gallon contains less energy.
• Exhaust temperature. At a proper mixture, owners and tuners commonly see exhaust gas temperature drop on the order of 200°F versus a gasoline tune. Peak cylinder pressure is also lower, while pressure stays useful for longer through the power stroke. That is a kinder pattern for a turbo engine than a gasoline tune that is on the edge of knock.
• Burn speed. At the rich mixtures where ethanol is efficient, it burns faster than gasoline. Timing does not need a huge swing versus a gasoline map, but it is not identical either. Copying a gasoline timing table and only changing the fuel numbers leaves power and safety on the table.

On a naturally aspirated engine with no other changes, a correct E85 tune is worth a small gain, on the order of 5%, because you are releasing more energy per pound of air. On a turbo Subaru the larger gain is the boost the octane lets you keep. One long-term 2002 WRX log, with a 16G turbo and injectors around 810 cc/min, recorded 281 all-wheel horsepower on E85 at 17 psi, against 240 AWHP on 100-octane race gas and 250 AWHP on 104-octane race gas on that same car. That is one setup, not a promise for every WRX. It does show the pattern: once the fuel system can feed the blend, E85 competes with race gas instead of with 91 octane.

What Has to Change First

The stock WRX ECU has enough fuel-trim range to add some ethanol. On an early 2002 WRX with stock injectors, that range ran out around 33–35% E85 by volume. Past that point the trims peg full rich and the ECU sets a “too lean” code, because it can tell the mixture is lean and it has no more authority to add fuel. A splash of premium pulls the blend back under that ceiling and the light can clear. That is a drivability limit, not a performance tune.

A real E85 conversion is shorter than the old alcohol-conversion lists from the 1970s. On a port-injected WRX, long-term testing kept coming back to two required changes:

1. More injector flow, about 30% above the injectors that were already correct on gasoline for that power level.
2. A tune that targets E85 mixtures and timing, instead of asking the stock trims to absorb a full tank of ethanol.

An adjustable fuel-pressure regulator can stretch injector flow a little, and it is a useful tuning tool, but it does not replace injectors that are large enough. Raising pressure to fake injector size also changes the map the ECU expects. Decide the hardware, then tune for that hardware.

A Practical Hardware Ladder

Direct-injected engines (FA20DIT WRX and later direct-injected Subarus) still need a tune and enough pump volume, but the port-injector sizing rule below does not transfer one-for-one. Those cars move a large share of the fuel through a high-pressure pump. Have that system sized by someone who tunes that engine. Do not apply an EJ injector chart to an FA and call it done.

Fuel System: Flow vs. Corrosion

Most of the scary E85 stories come from two different eras mixed together. Carbureted cars from the 1960s and 1970s used hoses, needle valves, and floats that were never meant to see alcohol. When oxygenated fuel arrived, some of those parts leaked, swelled, or clogged filters with years of varnish the ethanol had dissolved. Manufacturers then changed hose, seal, and O-ring compounds. They did not design those parts to be “sort of OK” at 10% ethanol. They changed materials.

Methanol is the fuel that attacks magnesium, zinc, and unprotected aluminum and that raises electrical conductivity enough to matter. Ethanol is much milder. Its dielectric constant sits well below methanol and water. On modern OBD-II Subarus, multi-year E85 use — including a 2002 WRX fuel line cut open around 2007 — did not show the hose degradation people expected from the old methanol stories.

Flow is the real constraint.

• Injectors. Plan on roughly 30% more flow than the gasoline injectors that already matched the car. A stock-injector WRX that is happy on premium will go lean, or peg its trims, on a full tank of E85. The same car in the long-running test moved to STI 550 cc/min injectors for high ethanol content, then to about 810 cc/min once a larger turbo and more boost were in the picture.
• Pump. The stock pump is often the first part that runs out of volume once injectors and boost go up. Ethanol-rated in-tank pumps in the 255 L/h class (a Walbro 255 was the pump in that long-term WRX test) are the usual first step. Bigger turbos need more pump than that. A pump that loses pressure after shutdown is an install problem — dry O-rings, a cocked assembly — not proof that E85 dissolved the hanger.
• Filter. Ethanol is a solvent. The first tanks can loosen old deposits and load the filter. That was a common gasohol-era complaint, and it clears with a new filter. It is not ongoing corrosion. Owners who cut filters open after years on E85 have generally found clean fuel-system parts, not rust.
• Lines, tank, and O-rings. On WRX-era cars they have not been the failure point. You do not get to skip a leak check after you open the tank. Replace any seal you damage, use the seals supplied with the pump kit, and do not assume a universal grease is ethanol-safe.
• Oxygen sensor. The factory sensor only cares whether the engine is at stoich in closed loop. It does not need to be replaced because the fuel is ethanol. A wideband used for tuning is a different tool, and it does need the right scale. That is covered below.

Injector Planning Shortcut

The E85 FAQ uses a planning estimate for a four-cylinder that breathes like a WRX:

Turbo airflow at the boost you actually intend to run (CFM) × 1.38 ≈ static flow per injector, in cc/min, on E85.

Use it to see whether a set of injectors is in the right neighborhood before you buy them. It assumes a WRX-like four-cylinder, gasoline-style port injection, and a target mixture — not a direct-injected FA, and not a substitute for dead-time data and a dyno. If the number lands on top of an injector’s maximum, you do not have headroom for a winter blend that is richer in ethanol than you expected, or for a hot day when the pump falls off.

Tuning Targets and AFR

A wideband that only speaks gasoline will lie to you on E85 in a very specific way. It still reports a stoichiometric burn as 14.7:1, because that is the gasoline scale baked into the display. E85’s stoichiometric mixture is about 9.8:1. If you chase 11.5:1 or 12.5:1 on that gasoline scale while the tank is E85, you are not where you think you are.

Switch the meter to lambda. Lambda 1.00 is stoich on any blend. If the meter cannot show lambda, divide the gasoline-scale reading by about 1.47 to 1.50 to estimate the true AFR of straight E85. One dyno session in the FAQ showed a gasoline-scale reading of 11.6:1 on a full E85 tune. Divided by 1.5, that is about 7.7:1, which sits in the max-power range for E85 rather than at the dangerously lean number 11.6 looks like on a gasoline scale.

Reference Mixtures

Read the rich column as the far end of the power range, not as a cruise target. Both gasoline and ethanol reach their best thermal efficiency nearer 15% rich of stoich. Ethanol keeps gaining torque out toward 30–40% rich of stoich, which is where the 7:1 figure comes from. Closed-loop cruise should still be lambda 1.00. Wide-open throttle is where the extra fuel belongs.

Partial blends scale between the rows. A tank that is half E85 and half gasoline is not the E85 row and not the gasoline row. This is the single most common way a “correct” map goes lean: the tune is right for one percentage, and the tank is another.

Timing

At light and moderate load, minimum timing for best torque is nearly the same on E85 and gasoline. At high load, in a closed-loop-style test, E85 wanted a few degrees more advance. At a rich wide-open-throttle mixture the burn speeds up, and the engine may want less advance than the gasoline map, because less of the burn happens while the piston is still rising. Those two statements are both in the research, and they apply to different load cells. The practical version is simple: do not add a pile of timing because a forum post said ethanol is slow, and do not leave a gasoline timing map untouched because another post said it is the same. Log knock and torque, and let the load cells tell you.

Ethanol Content Is Not Always 85%

The ASTM-style volatility classes used for E85 set a minimum ethanol content, not a fixed 85%. Suppliers can run richer in ethanol than the minimum when it is cheap to do so. They drop toward the minimum when cold starting demands more gasoline in the blend.

The calendar is regional. A Front Range Colorado pattern cited in the FAQ runs Class 1 from mid-June to mid-September, Class 3 from mid-October through mid-April, and Class 2 in the months between. Florida can sit on Class 1 most of the year. North Dakota can sit on Class 3 most of the year. If your tune assumes 85% every month, winter fuel is a different mixture than the map.

The Water Test, and Why the Math Is Not Linear

Ethanol mixes with water. Gasoline does not. Adding a measured amount of water to a fuel sample pulls the ethanol out of the gasoline and the two liquids separate. That is a real test, and it is also easy to misread, because ethanol and water bond into a volume that is smaller than the two volumes added together.

The calibration point from the fuel handbook is the one to remember: 10 ml of water added to 100 ml of true E10 does not produce a 10 ml separated layer. The water-and-ethanol layer comes out around 17 ml. A “the extra milliliters are the ethanol percentage” reading is wrong in the direction that makes you trust a tune you should not trust.

If the car is on a single E85 map, measure the fuel or run a content sensor. Guessing from the pump sticker is how winter blend and summer blend get the same wide-open-throttle fuel table.

Mileage and Cost per Mile

A gallon of E85 contains less energy than a gallon of gasoline — about 90,500 BTU versus about 125,000 BTU, roughly a quarter less. Miles per gallon fall. They usually do not fall by the full energy gap, because the engine is more efficient on the blend: cooler charge, more expansion from the extra fuel burned per pound of air, and less throttle opening for the same torque.

In the long-running WRX logs, conservative driving on high ethanol returned about 92–94% of that car’s gasoline mileage. Other stretches, with a larger turbo and harder driving, landed nearer 18 mpg on E85 against about 22 mpg on gasoline. Use a planning range of roughly 15–25% fewer miles per gallon until you have your own numbers. Your turbo, tune, and right foot move that range more than the chemistry does.

When E85 Is Actually Cheaper

Cost per mile is price divided by mpg. E85 is cheaper to drive only when the price cut is larger than the mileage cut. The prices below are a worksheet, not a quote.

In that example the fuel is cheaper per mile because the discount (25%) beats the mileage loss (17%). If E85 is only 10% cheaper at your station and the car loses 20% of its mpg, premium wins on cost and E85 is only “worth it” for the octane. Do the division with your pump price and your dash average before you restructure a commute around an E85 station.

There is a second version of the same physics that explains the power. Per pound of air, at a max-power mixture, E85 releases more heat than gasoline — on the order of 25% more in the standard BTU comparison (about 193,000 BTU versus 152,000 BTU for 100 lb of air). Less energy per gallon, more energy per intake stroke, once the injectors can supply the gallon. Both statements are true, and they describe different questions: operating cost, and cylinder pressure.

Cold Starts and Winter Blends

Ethanol does not want to vaporize in a cold intake port. Below about 20°F, the practical fix used on the long-term WRX was to keep the tank nearer 80–90% E85 by adding a couple of gallons of gasoline, rather than insisting on a summer blend in January. The car starts easier, warms faster, and the winter pump fuel is already headed that direction on its own.

Very cold soaked starts can be slow even with that blend. On conventional plugs, that WRX needed a minute or more of idle in single-digit weather before it would pull itself in gear. A hotter spark plug strategy improved that particular car’s cold-start behavior. Plugs are a supporting detail. They do not fix a tune that is still on the gasoline cranking table.

• Expect the station’s winter E85 to be closer to 70% ethanol than to 85%.
• If the tune is a fixed E85 map, recheck ethanol content when the season changes.
• Do not “fix” a cold-start stumble by going leaner. The stumble is vaporization, not a lack of air.
• Give the engine a short warm-up before boost. E85 does not reward immediate wide-open throttle on a stone-cold block.

Flex Fuel vs. One E85 Map

A dedicated E85 tune is the right tool when every tank comes from an E85 pump and you can measure the percentage often enough to know the map still matches. It is the wrong tool when the tank is a moving mixture.

Owners have built workable blend cars by tuning for one known mix, such as 50/50, and staying near that mix. The ECU can trim closed loop. Open loop will not save you. The FAQ’s own caution, from an early flex-style autotune experiment, is that open loop goes lean if fuel pressure or ethanol content moves and the WOT table does not. Turning the pump or the regulator down on a map that was already matched to E85 is a lean-out, not a drivability tweak.

A flex-fuel content sensor measures the percentage in the line and lets the tune interpolate between a gasoline calibration and an E85 calibration. That is the setup that matches how these cars are actually driven: a road trip through a state with one E85 station, a winter blend in October, a splash of premium when the ethanol pump is down. Hardware-only “flex” kits that do not change the tune with the sensor reading are just a gauge. The ECU still has one map.

Oil, Emissions, and Warranty

Oil

Older studies of engines that ran on neat alcohol, with 1940s and 1950s oils and soft valve seats, do not describe a modern Subaru on synthetic oil. Alcohol that reaches the crankcase boils off once the oil is hot. The long-term E85 WRX logs did not turn into bearing or valve-seat failures. The situation that still deserves shorter oil changes is the one you already know: repeated short trips that never get the oil hot, on any fuel. A 180°F thermostat was used on that test car to speed warmup and was left in through summer without a cooling complaint. It is optional, not a required E85 part.

Emissions

Ethanol tends to lower hydrocarbon and carbon-monoxide emissions when the mixture is under control. It does not give you a pass on the evaporative system. Higher vapor pressure is not what the charcoal canister was mapped for, and there is no bolt-on canister kit that makes a converted car a certified flex-fuel vehicle. A lean code, or any other check-engine light, fails an inspection that requires a clean monitor status. The practical inspection trick from the early tests was to lower the ethanol percentage until the trim-related light cleared. Removing catalysts to “help” an E85 car is illegal on a street-driven vehicle and is not part of this setup.

Warranty and what Subaru actually approved

The owner’s manual language that matters is the E10 limit. The car is warrantied to run on gasoline with up to 10% ethanol. Above that, including E85, you are outside the fuel the powertrain warranty assumed. Methanol is called out separately and more strictly, typically a ceiling around 5%, because methanol is the corrosive alcohol. Ethanol’s material record on these cars is far better than methanol’s. That record is owner experience over years, not a Subaru approval.

Storage belongs in the same honesty column. Cars that sat for long stretches on E85 and later started clean had been kept on a battery tender, treated with a stabilizer meant for ethanol, and not left for years on a half-empty tank that could breathe in water. Bulk-drum fuel handling is a different problem — fire code, vapor, and water absorption — and it is not a driveway project this guide is going to spec. Buy E85 at a pump.

Switchover Checklist

Use this order. Skipping to the pump because the station is on the way home is how stock injectors meet a full tank of ethanol.

1. Decide the fuel you will actually run. Dedicated E85, a single known blend, or real flex fuel. The hardware and the map follow that decision.
2. Size injectors for about 30% more flow than the gasoline setup that already fed the car, then add headroom if boost or turbo size is going up. On a direct-injected car, size the high-pressure side with the tuner, not with the port-injector shortcut.
3. Size the in-tank pump for that same airflow. An ethanol-rated 255 L/h pump is a common first upgrade on a port-injected WRX. It is not the last pump a big turbo will need.
4. Replace the fuel filter after the first tanks if pressure starts to fall off. Ethanol will pick up old deposits once.
5. Tune it. Lambda, not a gasoline AFR you remember from a 91-octane map. Cruise at lambda 1.00. Wide-open throttle richer, in the range the E85 table above describes, verified on a wideband that understands the fuel.
6. Confirm ethanol percentage when you change stations and when the season changes. Class 3 winter fuel can be 70% ethanol. A summer map does not know that.
7. In weather below about 20°F, expect to blend back toward 80–90% E85 or to run the station’s winter fuel, and let the car warm up before boost.
8. Log the first heat cycles. Fuel trims, wideband, knock, and pump pressure. A conversion that only “feels faster” has not been checked.

E85 FAQ

Can a stock Subaru run E85?

Not a full tank. A 2002 WRX on stock injectors and a stock ECU tolerated roughly 30–35% E85 before the fuel trims maxed out and it set a lean code. Above that you need larger injectors and a tune. The car is warrantied for E10, not E85.

How much more fuel does E85 need?

About 30% more volume to reach the same relative mixture you had on gasoline. That is an injector and pump problem first, and a miles-per-gallon problem second.

What AFR should I tune for?

Use lambda. Stoich is lambda 1.00, which reads about 9.8:1 true AFR on E85, not 14.7:1. Best-torque wide-open throttle is richer than a gasoline map, out toward 7:1 true AFR (about lambda 0.71) at the rich end. If your wideband only shows gasoline AFR, divide by about 1.5 on straight E85.

Does E85 destroy fuel lines and O-rings?

On OBD-II Subarus, long-term use has not shown that. Those parts were updated decades ago for ethanol-blended gasoline. Methanol is the alcohol that eats soft metals. The parts that actually have to grow are the injectors and, as power rises, the pump.

Will my oxygen sensor work?

The factory sensor will. It is looking for stoich in closed loop, not for a gasoline-specific chemistry. A tuning wideband must be read as lambda or converted, or every number it shows will be on the gasoline scale.

Why did my miles per gallon drop?

Because a gallon of E85 holds roughly a quarter less energy than a gallon of gasoline. Real-world economy often drops less than that, commonly in the 15–25% range, and sometimes closer to 8% in gentle driving. It will not match premium economy.

Is E85 always 85% ethanol?

No. Winter supply can be as low as 70% ethanol. Summer supply is higher. The pump still says E85. A fixed tune does not track that change unless you measure it or run a flex-fuel sensor.

Can I mix E85 and premium?

Only if the tune was built for that mixture, or the ECU is using a content sensor to move between maps. Mixing on a single WOT map is how open loop goes lean. Small blend changes on a near-stock car, done gradually and under the trim limit, are a different and much milder case.

Is E85 bad for the oil?

Modern synthetic oil in an engine that reaches temperature handles it. The historical valve-seat and oil problems were neat-alcohol engines with old oil and soft seats. Short trips that never warm the oil up are the case that still calls for earlier oil changes.

Will it pass emissions?

A correct mixture tends to help hydrocarbon and carbon-monoxide numbers. A check-engine light will fail an inspection that requires monitors to be clear. The evaporative system was not designed for high ethanol blends. Leave the catalysts in place on any car that drives on public roads.