For an E-85-type ethanol conversion, there are 3 things you must address: (1) mixture ratio at all power settings, (2) extra timing advance, on the order of 15 extra crankshaft degrees, and (3) quite a bit of extra intake heat to promote adequate vaporization.
If you plan on burning something more concentrated than E-85, there is a fourth item: (4) you will need a start fuel canister of some kind. Otherwise starting at ambient temperatures below 50 F becomes essentially impossible.
For a stiff blend up to E-35, there is no conversion at all! It is just a drop-in fuel. And ethanol splash-blends with gasoline in the tank, so stratification isn't an issue at all. With today's E-10 for unleaded regular grade, use no more than 1/3 E-85 by volume, no less than 2/3 regular by volume, and your mixture will be no stronger than E-35. If it takes 12 gallons to fill up, no more than 4 of those should be E-85, no less than 8 of those should be unleaded regular.
Full E-85 conversions: Mixture Ratio -- With carburetors, increase main jet and idle jet sizes by roughly about factor 1.2 on diameter. You have to sneak up on the right value, it is different for every car. For fuel injection, replace the fuel rail pressure regulator with an adjustable aftermarket unit and set it about factor 1.5 higher pressure, and do not tell the electronics you did anything at all.
Full E-85 conversions: Extra Timing Advance -- First thing to try is +15 crankshaft degrees from gasoline stock. Ethanol's burn speed is similar to gasoline, but its ignition delay is longer so you must start earlier with your spark. What you want to end up with is peak cylinder pressure 1-2 crankshaft degrees before top dead center, no more, no less. This issue is extremely serious with a modern over-square, short-stroke, high-rev automotive engine. With very old under-square, long-stroke, slow-rev technology, this issue makes almost no difference at all.
Full E-85 conversions: Extra Intake Heat -- it is impractical to modify the intake manifold hot spot in most engines. The only viable alternative is hotter intake air. It needs to be above about 70-80 F minimum, before fuel is introduced. 90 F is even better.
Some folks have contacted me for help do-it-yourself converting cars to E-85 or for using "stiff" gasohol blends in unmodified cars. Here's what I know works ----
My "ethanol VW" was a 1973 beetle with an essentially-stock 1600 cc case, jugs, and crank. It was a dual-port head engine, stock valves, stock rockers, no modifications at all. I had long ago replaced the worn-out combination distributor with an aftermarket Bosch 009 all-mechanical unit. (I had also long ago undone the idiot 15-degrees-retarded timing setting, that the factory used in 1973 to try to meet EPA emission standards. This took me from 5 degrees late (mark on pulley) to 10 degrees before top dead center, static timing.)
After I got the Bosch-009, what worked best for ignition timing was a very simple 30 degrees all-in (about 2500+ rpm) by strobe, on gasoline. I used this setting successfully for decades. I decreased the valve lash setting and oil change intervals to 2000 miles from the factory-recommended 3000 miles, and changed from the recommended lash of 0.006 inches intake and exhaust to 0.006 inch intake and 0.008 inch exhaust. These changes enabled me to avoid valve-burning and excess bearing wear problems in the Texas heat. By switching to aviation-grade oil in the 1980’s, I was able to increase the oil change interval to 4000 miles. After the advent of SF-or-better grade auto oils in the mid 1990’s, I was able to return to using auto oils at the longer 4000 mile interval.
The original carburetor was a Solex 34 PICT-3. I went through a couple of them; they wear around the shaft of the throttle plate, and leak air. It upsets the off-idle transition very badly. I had finally replaced it with an aftermarket Solex 30/31 with the adapter plate for the 34 mm manifold. That worked fine for many years, but finally wore out the same way as the 34 PICT-3. I also had available, but had never used, a not-worn-out Solex 30-PICT-2, off a single port head 1600 cc Bus engine.
I went through several combinations of jet sizes with both the 34 PICT-3 and the 30 PICT-2 carburetors, before I settled on the 30 PICT-2, because it did not leak air around the throttle butterfly shaft. I had to use the adapter for the 30/31 to make it fit, and the accelerator pump cover off the 30/31, to find an accelerator pump link bar that would fit. I never even tried the 30/31 with ethanol, because it has idle circuitry that I never really understood: it uses two idle jets of different sizes.
The Converted 30 PICT-2
On the 30 PICT-2 the stock main jet was a "116.0", which is 1.160 mm dia (.045"). On E-85, I settled on a "137.5" from another aftermarket Solex-Brazil carburetor, which is 1.375 mm dia (.054") for good driveability at speed. The stock idle jet was a g55, which is 0.055 mm dia (.022"). I drilled that out to .762 mm (.030") before I was satisfied with idle settings. If the idle jet is too small, you will suck the idle well dry with too much idle circuit air flow, because the idle screw is open too wide. It's a transient effect, with a time constant somewhere around 15-30 seconds. The stock accelerator discharge nozzle is right at .50 mm dia (.020"). I drilled that out to .712 mm dia (.028") before I was satisfied with the off-idle transition.
Ignition Timing Changes
None of this works at all, if you don't first revise the timing. I found that out the hard way. On gasoline with my aftermarket distributor, timing was +30 BTDC all-in at about 2500+ rpm. I set that with a strobe as the most repeatable way. I had to add 15 degrees to that setting, before it showed the same coast-down vacuum curve on E-85. On E-85, the revised timing spec was thus +45 degrees BTDC, all-in at 2500+ rpm. In other words, you need to add right about 15 crankshaft degrees to whatever timing setting you are using on gasoline. Use the minimum that recreates your old gasoline vacuum coastdown curve.
The Converted 34 PICT-3
You may still have a 34 PICT-3 carburetor. If so, here are the best jet combinations I found, before I gave up on it due to the shaft air leak. Stock main is a "127.5", which is 1.275 mm dia (.050"). On E-85, I used a drilled-out 1.57 mm dia (.062"). Stock idle was a "g55", which is 0.55 mm dia (.022"). On E-85, I drilled that out to .965 mm dia (.038"). Stock accelerator discharge was 0.50 mm dia (.020"). On E-85, I drilled that out to .965 mm dia (.038"). I give both metric and US sizes, because it's a metric car, but all I had to work with was a set of the tiny US-sized bits one uses to clean out oxyacetylene torch tips.
Heated Intake Air
The only other thing I had to do (which you might not if it never gets cold where you are) was to fool the intake air into thinking it was always summer. Any time the outside air temperature was under 70 F, I sucked my intake combustion air from a partial sheet metal glove around the muffler, made from scrap metal roofing trim. Above 70F, ambient air works fine. If you don't do this, both driveability and mileage suffer whenever it is cold. This rig worked all the way down to 15 F for me. I did it with a tee made of scrap plastic bottles on the air cleaner intake. I just plugged-up the cold inlet in cold weather, and let it draw from both inlets in warm weather. My hot source was connected to the side inlet of the tee, which has just a tad more flow resistance. Thus it favored cold air with both inlets open.
How It Should Perform
Have fun running your late-model 1600 VW beetle (or bus) on E-85. If you do it right, you should get about 80% of your former gasoline mileage, not the 70% that the fuel energy per gallon says you should get. Ethanol simply burns more efficiently than gasoline in a piston engine. This partially offsets the lower energy per gallon of the ethanol. Tailpipe soot should gradually disappear. Your spark plugs will start looking pristine-clean, too. So also will the carburetor casting look much cleaner, inside and out. It's really amazing how much cleaner E-85 is than gasoline, in so many different ways.
Minor-to-Moderate Compression Troubles, and How to Cope
If you smell ethanol in your motor oil (it'll smell different, anyway, so I am talking about detecting really serious odor here), your rings are leaking. This will show up as uneven (by around 15 psi) or generally somewhat-low (by about 15 psi) dry compression test readings. If your readings are worse than that, you really need to do the overhaul work first. Add about 10 or 20% Lucas Oil Stabilizer to your crankcase oil, and that modest compression defect will correct itself, and the fuel smell in the oil will go away.
Use the "finger test" to reset your oil change interval, it'll get substantially longer with synthetic in the mix like that (mine pretty much doubled from 4000 to at least 8000 miles, on modern SM-rated oils in an 80-20 blend). The Lucas additive really does a good job arresting cold start wear. Before the advent of SF-grade oils, I could not get even 3000 miles without seriously failing the finger test, so I used aviation-grade oils instead. Nowadays, the SM-grade(same as ILSAC-4, by the way) is way better than the aviation grade oil.
If you don’t understand how to run the “finger test”, you better ask me, or a professional mechanic. It’ll tell you everything a lab test can tell, except for a numerical particle identity and count. But, if you see visible metal wear particles, that’s all you need to know anyway (time to overhaul completely).
My 1973 VW beetle is going back into mothballs. Before I was done with it, I reset the carburetor back to gasoline settings by installing a screw on the enlarged main jet, reset the timing back to gasoline-suitable, and undid the heated intake air. I didn’t change the enlarged idle or accelerator discharge. I just reset the idle speed and mixture screw settings as needed, to make it run just fine on gasoline. Then I ran progressively-stiffer gasohol blends until I saw the late timing problem kick-in about E-45-ish on blend strength. The car ran just fine testing blends all the way to E-57 like that. I got the same story (late timing above about E-45) from fuel mileage figures in my fuel-injected unmodified 1995 F-150, and subjectively from my fuel-injected, unmodified 1998 Nissan Sentra.
I did have to reset the VW carburetor screws a little for the blends above about E-40. The fuel-injected Ford and Nissan needed nothing at all, all the way to E-50-something (because closed-loop injection compensates mixture strength automatically, within system flow rate limits).
The requirement for the extra 15 degrees of timing advance (and presumably the warmed intake air) seems to kick-in like a light switch, right about at E-45-ish. If you don’t make these changes, then above E-45-ish, you run weak, smooth, and fuel-consumptive, with a little less intake vacuum on coastdown. That's symptomatic of late timing.
Blend Limits for Unmodified Engines
Cold-weather start "irritations" limited me to E-30 to E-35 max in the unmodified fuel-injected cars. These take the form of starting but dying quickly. A second start then usually works just fine. The problems kick in about freezing. I have tested down to about 10 F here in Texas. This is not serious, just irritating. I have a very old carbureted VW beetle (1960 model) that is running totally unmodified on E-34, and it seems to be doing OK, too. All my completely-unmodifiable lawn and garden equipment runs just fine on E-34, and has for 4 years now.
E-85 is nominally 85% by volume ethanol and 15% gasoline. Its nominal volume fraction ethanol is thus 0.85. These days, unleaded regular gasoline is nominally 10% by volume ethanol and 90% gasoline. Its nominal volume fraction ethanol is thus 0.10. If you know how many gallons of fuel it takes to fill your tank or fuel can (V), and what blend fraction ethanol you want (R), you can use these figures to compute how many of the fill gallons should be E-85 (X):
X = V / [1 + (0.85 – R) / (R – 0.10)]
Examples: for a desired E-35 blend, R = 0.35. Thus X = V/[1 + .50/.25] = V/3.00. Similarly, for R = 0.30, X = V/3.75. For R = 0.25, X = V/5.00. For R = 0.20, X = V/7.50.
Assuming the tank is burned down pretty low (or the can is nearly empty), the residuals will combine with your fill blend pretty close to the R you selected for the fill. In the case of a vehicle fuel tank, this presumes that you have calibrated your fuel gage for gallons-to-fill versus marks on the gage.
This stuff “splash-blends” right in the tank or fuel can. No mixing is required. Just put in your “X” gallons of E-85, and top-off “to the mark” with gasoline. Total gasoline added should come out very close to “V – X” if you did it right.
Calibrating a Fuel Gage
Keep a mileage log over at least three tankfuls of fuel. Record as a minimum the odometer reading and the gallons-to-fill at each fill-up. Fill the tank to exactly the same mark each time. The average mileage between fill-ups is the difference in odometer readings divided by the gallons-to-fill.
While driving on each tank of fuel, as the gage’s needle reaches each mark on the gage, record that odometer reading. The differences in these recorded readings give you miles-between-marks for that tank of fuel. Dividing those by the average mileage for that tank gives you gallons-between-marks. These you average over the multiple tanks of fuel. Listing the averaged gallons-between-marks in a cumulative fashion gives you gallons-to-fill (V) for each gage mark.
Checking Blend Strengths Experimentally
I do this with a simple added-water phase separation test. This requires lab-grade glassware, those being a graduated cylinder of 100 cc capacity for the fuel sample, and a graduated cylinder of about 30 to 50 cc capacity for the added water. You must “abuse” standard laboratory practice and read these to the nearest quarter-division instead of the standard-practice nearest half-division. If you do it this way, your results will come out pretty close to plus or minus 1 or 2 percentage points on blend strength (plus or minus 1 or 2 E-number points). Smaller sample sizes do not work out accurate enough to be useful. I draw my samples from the Schrader fitting located on the fuel rail in most fuel-injected vehicles.
Draw a fuel sample between 58 and 68 cc in volume into the larger cylinder and measure it precisely (bottom of meniscus, or BOM). Compute 1/3 of this volume for the water, put about that much into the smaller test cylinder, and measure what you have precisely (BOM). Record these numbers. Then add the water to the fuel, which will begin to phase-separate immediately. Let this stand 2-4 minutes until all the air bubbles quit decanting. Then measure the total liquid volume (BOM), and the volume below the interface between the separated layers (there is no meniscus, this is a flat plane).
Now, all the water and the ethanol go to the bottom layer, which may grade from cloudy white below to clear right at the interface. The hydrocarbon will all go to the top layer, which is a clear straw-colored liquid. You cannot use the water-plus-ethanol volume directly, because mixed ethanol and water volumes are not conserved, while mixed ethanol and hydrocarbon volumes are conserved.
Subtract the wet ethanol layer volume from the total separated sample volume to determine the hydrocarbon volume floating on top, and record it. Subtract this hydrocarbon volume from the original fuel sample volume, to determine the wet ethanol volume present in the original fuel sample, and record it. Dividing this wet ethanol volume by the fuel sample volume determines the wet ethanol fraction in the original fuel, which in percentage format is a really good estimate of the blend E-number.
I typically find the E-10 “gasoline” to be really closer to E-8; indeed, the placard on the pump usually says “up to 10% ethanol”, not “exactly 10% ethanol”. E-85 typically tests as E-87, which means there is most likely about 1-2% water in the mix. That’s not surprising, as moisture from the air readily absorbs into the ethanol in the fuel. 1 or 2% water is not a problem.