Gasoline Filtration... and Vent Filtration Too?

rev 10-14-2012, rev. 7-30-2024

My last boat had terrible problems with phase separation. Every time I went out I had to remember to drain water. After years of frustration over hard starting I figured it out and installed a Raycor filter and oil/water separator. I became diligent about draining the water before every trip, and the problems became few.  And then e10 appeared in the market, and problems returned, but different problems. Instead of a few milliliters now and then, the gas would stay dry, until i would suddenly separate en mas. When e10 absorbs more than 0.5% water, the ethanol and  water become insoluble and settle to the bottom. Additionally, in the fall when the temperature drops the ethanol/water fraction becomes less soluble; as much as 1 gallon would drop out. The source of the water was simple enough; the boat had terrible fill and vent locations and small amounts of rain water were entering the tank. Humid air certainly contributed, as it was, for the most part, a summer and fall problem.  On two occasions I motored out to the inlet, only to have the engine die between the jetties.  Over a gallon of water had separated each time. Fortunately, we're good at quick anchoring, and we learned to motor back to the dock using fuel bailed from the top of the tank into a small tank (Gatorade bottle) mounted (duct tape) to the transom.


When I got the PDQ, I assumed I was in for the same joy, so I installed an overkill filtration system. The original fuel filter was a tiny in-line fritted element in a seized aluminum casing (junk--had one on my last boat and it was junk too) located in an awkward place near the fuel tank. The engines themselves have little lawn mower filters located under the cowling (keep some extras--they change in moments and are cheap from Advance Auto parts or the hardware store). The engines were stubborn and I thought this might be a part of the trouble, so I removed the fritted element and threw it away and mounted a pair of smaller Raycor separating filters. e10 was new then, I hadn't yet done my studies for Practical Sailor, and I didn't understand the nuances. It just seemed that some boats were troubled and some weren't. Fortunately, the PDQ 32 has much better vent (out the bottom of the bridge deck) and fill (tight o-ring and out of run-off course) locations and I've never had so much as one drop of water find its way to the separator these past few years.  The Raycor is overkill.  I also know that if I should get a batch of wet fuel or leave the filler cap loose, I can deal with that with the equipment I have installed. I like that.

But I did drain water just after installing the system. The boat had seen little used for a few years and these filters did remove a few ounces of water those first few days. I guess we stirred things up, bashing up the Bay in gale conditions. Unlike diesel, subsequent e10 fill-ups most likely absorbed what was left.  Although this initial dose of water almost certainly contributed to the carb problems I had the first 6 months, the water and all memories of it have passed.

I also replaced the West Marine plug-style fuel cut-off valves that were not very "marine." Don't buy these; I've had 3 out of 3 seize within a few years in gasoline service; I suspect they do better with diesel. Common Teflon seat ball valves are a much better choice. There should also be a cut-off mounted on the tank.

The filter cartridges rust quickly from salt exposure, but they last a few years, at which time they should be changed. Note that the clear bowls can ONLY be used on with gasoline on outboard applications and above decks; these are in bulkheaded outboard wells, which amounts to the same thing. The air vent hose I added is not needed or used. 

 Why the posts at this time?  As you can see, I did not simply mount the filter to the bulkhead with through bolts; I mounted the filter to an aluminum base plate, which was then attached by screws to the bulkhead. Through bolts would have been inaccessible on the port side due to cabinetry, and I wanted to be able to unbolt the filter assembly myself, without major surgery. The 5/16-inch bolts that secure the filter assembly are actually carriage bolts that are captive to the base plate, allowing me to unbolt the filter assembly very easily without rear access.  These filter cartridges can get quite tight after a few years, so I like to have a strong mount and the possibility of removing the whole business for leverage and abuse on the dock.

The port side has done fine, the screws remaining solid, but the starboard side was knocked loose a year ago by by an extremely vigorous motor kick-up, pulling some screws loose and twisting the base plate. It looked like hell, and someday it might fall off. Yesterday I replaced all 8 screws on both mounting plates with #10 through bolts and fender washers; now she's permanent. I'm hoping to keep this boat for a very long time, so no half measures, not any more.

While I'm on my rant, I'll mention that the 3.5 hp Mercury on my tender had no fuel filter at all, which was a mistake. There is a course strainer inside the tank--hard to reach and not very effective. A bout 18 months ago I added a small lawn mower element under the cowling. This seems to be a common cruiser addition. If it clogs, it's a 5-minute fix. Between this addition and consistently closing the built-in tank after use (water infiltration), reliability has become quite good.
 
 Would I advise others to invest in such overkill? It is nice knowing you'll never have water or filter problems that can't be solved on the water. Very nice.

_______________

A few notes on phase separation of e10 (gasoline with 10% ethanol). I can't give away everything--you're supposed to subscribe to Practical Sailor--but I can share a few observations you could easily find on the web.

• Phase separation by absorption of water from the air is a myth. It's impossible. Assuming the e10 was dry when received, by the time the gasoline absorbs enough water to become unstable, enough of the light fraction gasoline has evaporated--perhaps 20%--that it won't run properly anyway. At that point water drop-out is noticed, but it isn't true phase separation; so much of the ethanol is gone that the water simply goes to the bottom for lack of a co-solvent. There is no ethanol to keep it in solution. This effect can be seen in small ( typically 0.75 to 1.5 quarts) outboard tanks and is frequently seen in carburetors. In the former case, the engine won't run for beans; new gas is needed and keeping the tank full and the vent closed is the solution. In the later case, the engine is normally fine after a few pulls; there will be only a few drops of water. Running it dry is one solution, sailing often is the other. I'm not guessing. I've spent the past few years doing many experiments, trying to induce separation. Unless there is minor leakage or gross evaporation, it doesn't happen. Mercury Marine has posted the same findings; I find this presentation agrees with everything I've seen.

• The smallest vent or filler leaks do cause separation. Only about 0.5% water is need to make the fuel unstable. Of course, some of this can be provided by absorption from the air, so humidity is still a measurable factor.
• Keep your tank FULL.  A full tank doesn't breath. No breathing = no fuel evaporation, less oxidation, and less moisture absorption. Also less fuel train corrosion and less gum formation.
• Additives don't prevent phase separation. Nothing but snake oil, those that make this claim, full write-up in Practical Sailor. While there are a few that can suck-up water, they do so by adding large amounts of either soap or strong solvents (MTBE, celusolve) to the gasoline. Very bad for a 2-stroke and not for regular use in any engine. Are some of the additives useful in terms of reducing oxidation and corrosion? Probably. Some have the correct ingredients for that, and I will best testing that soon....

I'm starting some more tests on e10. These will run through the summer and be reported in Practical Sailor.

• Corrosion vs. additives. Copper, brass, cast aluminum, steel, cast iron, and solder.
• Corrosion vs. water content (both separated and absorbed).
• Gasoline tank desiccant vent filters. H2OUT is marketing to the marine industry.
• Gasoline tank carbon vent filters. New boats with installed tanks will be phasing-in carbon canisters, similar to those in cars, between July 2011 and July 2013. But would a retrofit actually benefit an older boat? If you would like drier fuel and to save a few gallons from evaporation each year, the answer is probably yes.

I suspect we are going to find that some of the additives do help with corrosion, and that drier fuel is always better. For the desiccant and carbon filters, Shoal Survivor will be our loyal test bed.

_____________

So what do the new vent filters look like?

After surviving a year as a holding tank filter (Practical Sailor tests), I reloaded with activated carbon and this housing will be serving for at least a few more years.  It is mounted to a FRP mounting plate, which will screw and epoxy to a cored bulkhead. It will be placed in the vent, between the Raycor Lifeguard air/fuel separator and the through-hull. This is a modified Vetus No-Smell housing. Holding 500 ml of adsorbent, it is the correct EPA-rated size for a 27 gallon tank,  is fuel rated, easy to switch adsorbents, and I had it.

Note 2024: I tested carbon first, because that is what the EPA favors, but I soon switched to silica gel, after confirming that carbon is all wrong for vent filters on boats:

• Carbon saturates with dew in a marine environment. The EPA tested using the automotive protocol which is lower humidity. Also carbon is what the EPA and vendors know from cars. But it's wrong for boats.
  
• Wet carbon cannot absorb efficiently because gasoline vapors don't like going though water. Obviously.
  
• Carbon is not as efficient in handling water as silica gel. In fact, silica gel absorbs gasoline vapors better than carbon in this application because it does not saturate with water. I did a series of gasoline vapor control tests to confirm.
  

The ethanol in e10 regenerates the silica with every cycle, pushing the water out. The outer portion of the bed guards against moisture and the inner portion control evaporation. An equilibrium is set up that lasts for several years.

Installed. The vent leads up from the fuel tank, through the Raycor Life Guard, through the Vetus No-Smell (with mixed bed carbon/Silica gel fill) and down to the existing through-hull. I had to manufacture a new perforated metal retaining plate for the Vetus unit; this will allow me to test a variety of adsorbents.

A factory installation, in a 2012 pontoon boat.  Instead of mounting the canister in a high loop, they protect it with a filling cut-off valve and sensor in the fill line (so that auto-stop filling pumps will work), an anti-surge valve on the tank, and a p-trap on the through-hull. These canisters are intended to last the life of the boat. All Attwood equipment.


Like my holding tank system, my fuel system--all new gasoline fuel systems--may seem a bit over engineered. I don't think so, not if maintenance becomes less, fuel related problems go away, starting is easier, and a materials amount of fuel is conserved (reduced evaporative losses) to pay for the project. Green, cheap, and lazy, over the longer term. That's just smart.

Try this link, if you would like a better summary of both the new rules and the safety rational behind gasoline fuels system requirements.

Extra bonus. The filter will retain several gallons of gas each year, paying for the project over time. And the volatiles, that help with starting, are retained.

---

Note: After 3-4 years you will want to either regenerate the gel or replace it. It is actually quite easy to regenerate by first letting it completely saturate with water by leaving it in open air (it will change color). This pushes the hydrocarbons out. Then cook it on an outdoor stove until the color changes back. Let it cool and refill.

 Changed color after three years

 

Cooking on the grill.

 

 

Almost done.

 

 

Recharged.