User Rank
Faulty Fuel Pump or Faulty Driver?
Critic   9/21/2012 6:31:15 PM
Automotive fuel systems are generally not very tolerant of large volumes (more than a few ounces) of water in the fuel tank.  This is true regardless of the make, model, or year.  In some geographical areas, it is advisable to periodically dump some "dry gas" additive in the tank to dissolve at least some of the moisture that may have accumulated in the tank.

Most production cars and light trucks are not very tolerant of being submerged in water, for many reasons, not all of which I will mention here.  It is generally best to not drive into water deeper than the height of the wheel center.  For example, if the wheel center is 15 inches from the ground, then don't drive into water deeper than 15 inches.  One bad thing that can happen in deeper water with a rear-wheel (or 4-wheel) drive vehicle is that the axle vent(s) can be submerged, flooding the axle housing(s) with water.  When this happens, the axle(s) must be drained and re-filled with gear oil.  Another bad thing that can happen is submersion of the exhaust pipe, potentially stalling the engine.  Submersion of the gas tank is a really DUMB thing to do- you would be just asking for a flooded gas tank.

When a car or truck gas tank gets flooded, the engine will stall as soon as water gets into the injectors (similar story for carbureted vehicles, but I won't go into that now).  When this happens, it is important to ASAP drain the tank and lines, replace the fuel filter(s), and flush the system with fresh gas.  It is a good idea to again replace the filter(s) after the flush.

If you don't drain the system right away and get it flushed out ("I'll do it when I have time, in a month or so"), there will be lots of corrosion damage to the fuel pump, fuel gauge sender, and possibly the tank and injectors as well.

Several years ago, I bought a used Ford F-150 truck in non-running condition.  The truck had dual fuel tanks, with separate fuel pumps.  Both tanks had considerable rust in them, and neither fuel pump or fuel gauge sender was in working order.  The front (side) tank had a slit/hole in the filler hose which had allowed water spray, when the truck was driven on wet pavement, from the left front tire to enter the tank.  This water, left in the tank for a long period of time, apparently destroyed the pump, sender, and tank.  The rear tank had much less rust in it, but the fuel pump and sender had both died.  It cost about $800 to buy new replacements for all of the failed fuel system components.

I don't believe any of the failures were caused by bad design or bad manufacturing.  Most of the problems were caused by poor maintenance:  extreme water accumulation in the tanks was allowed.

User Rank
Re: MPI HP Fuel Pums in a Can
betwys1   9/7/2012 4:18:19 PM
Pay attention Corey872!   :-)    my initial post mentioned that my fuel tank swallowed half a gallon of water - so you don't need to speculate on this point - it's a fact!   But Corey brings up TWO interesting features of fuel handlers; let me discuss both of them.

He mentions the hydrophobic qualities of some fine cell filters. There is a notable example of this quality still known to old-time pilots (like me!).      If you are forced to use questionable gasoline in an aero reciprocating engine, you really REALLY don't want the engine to swallow water in the fuel - it can stop an engine dead! The "Classical" method of filtering such fuel was to line the bowl of the tundish or funnel with a square of chamois leather - it passes gasoline, traps particles, and stops flow DEAD when it is touched by water.      So if a pump/can design admits fuel though a small orifice, the hidden filter media had better not get water-logged if it is hydrophobic.    (I was not sure whether this can filter was blocking with sediment of was waterlogging - but the mechanic fished it out for replacement.)

But did you notice Corey's second point?   In a returned flow system, you pump out x ounces of fuel, and a fair fraction of that fuel is returned by the fuel pressure regulator, still at reasonable pressure. (The vehicle in question regulates at 40 psi over manifold pressure so the return varies from 25 psi or so, to 38 psi. )      What an interesting insight to notice that this  flow is able pull more fuel into the can than a small head difference could provide! If its done right, the proprtion returned could activate a jet pump input more than the engine needs, so the can COULD fill up!

Oh yes, those S korean engineers might have found this was  a helpful way to avoid routing the inlet filter sock outside the can.   But it looks like they were not thinking about water contamination.       Fuel with up to 10% alcohol has this advantage - it is hydrophilic, so in moderate quantities the fuel can mop up water and pass it to the engine without problems. Those bottles of fuel dring fdluid usually have a good proportion of ethanol for just this reason.   Most  light aircraft feature tank sampler valves, which can let out fuel from the lowest point in a tank - this is easy to observe for water (at the bottom of a sample cup) because of the differewnt refractive index.  

Cars don't feature such valcves, but they MAY use a plug screw in the bottom of the tank - and its there for a purpose!


Thanks for these valuable points...

Brian Whatcott

User Rank
Re: ???
tekochip   9/7/2012 11:27:36 AM
That's a pretty well detailed failure scenario and sounds quite plausible.

User Rank
corey872   9/7/2012 10:22:15 AM
It's interesting you mention all this jet ski / water activity, then the truck dies, then suddenly it's the fault of the fuel pump?  Sketchy at best.

I have seen fuel systems such as you describe.  This 'plastic can and return orifice' do have functions.  The return fuel orifice is generally arranged as a venturi so when high pressure fuel returning from the engine is injected into the plastic can, it also vacuums up fuel from the bottom of the tank.  This venturi/vacuum action allows you to use most all fuel from the tank and as you say, always keeps the fuel pump covered in fuel.

I suspect your issues began by burying the truck in water.  It's likely you got some water in the gas tank as well.  A warm tank of fuel plunging into relatively cold lake water has a way of creating a vacuum and pulling water into the tank.

In running the truck a few miles, the 'plastic can' eventually fills with water because the fuel pick-up sock is a small enough mesh size and made from hydrophobic material, it allows fuel to pass, but not water.  So the can fills with water and chokes off the fuel to the engine.  Engine stops, fuel stops and the water slowly drains out of the can and spreads back across the tank bottom because it is heavier than the fuel.  With the sock exposed to fuel again, you can drive a few miles before the can fills with water, chokes off the fuel and the cycle repeats.

In rebuilding the pump, you probably cleaned out the old fuel / water and everything is back to normal.   The pump obviously worked before the jet ski adventure, and there are likely thousands of copies of this truck out there working fine with the exact same design, so it's a bit hard to fault that IMHO.


User Rank
Fuel Pumps for MP fuel injection
betwys1   9/6/2012 1:50:24 PM
Ah Naperlou: you get me nostalgic for that Old MGA with the clicking SU solenoid pump - that sooner or later would burn its contacts.       I went on to tour Europe in an XKE ('E type')  - but I am glad that pumps went to impeller types - even though they then buried them in the fuel tank - and on more miserable designs necessitated dropping the tank to replace the pump. There's a job for ya!      Those impeller pumps usually take a pump relay, and its THAT that now burns its contact!

Brian Whatcott

User Rank
Re: Do not Understand Title
betwys1   9/6/2012 12:20:20 PM
Hi Toolmaker - sorry the story was less than transparent.    I bought a salvage 2002 Kia Sportage: this was the first design series, and experienced a poor service record. The maker evidently learned quite a bit from this design, because the second series (2003 on )was  considerably better.      Not only was the first series  not that good overall, Kia had already changed fuel pump designs at least once on the way to the 2002 model  fuel system.

How did  I knew the earlier fuel pump layout was an open layout ?    Facing a $300 hit for a new replacement, I sourced several used pumps, of which one was skeleton style (year 2000), and one was canned       Both shared the very NICE design feature that the pump could be accessed by removing a cover on the rear passenger floor.

Anyway: if you put a pump in a can with a little hole in the foot of the can, then even though you return excess fuel to the can, the fuel level in the can will be below the fuel level outside the can.      The designer wanted (as I think you felt too) there to be a reservoir of fuel at startup, even in a low tank: indeed there was a little flap on the can orifice to let fuel in, but to stop it leaking out again.    The orifice had its own little filter arrangement - hard to see - that was the problem.    Without drawings, you will be struggling with this, I know...

Oh, the other question: how did water get in?  there was a vent valve which allowed air into the tank on controller command - this vent pipe was not high enough to breath air when I was swimming the truck!   :-)



Brian Whatcott

User Rank
Do not Understand Title
Tool_maker   8/24/2012 9:20:15 AM
I guess this thing has just gone completely over my head. Obviously the original design worked or you could not have gotten to the launch ramp. Perhaps your alteration has improved the performance, but it does not sound as though everyone who owns one of these vehicles has to make this alteration for the vehicle to run. Unless of course there are dealers across the nation with inoperable vehicles due to faulty fuel pumps.

I would be more concerned how I ended up with gallons of water in the fuel tank. Also curious as to the brand of this vehicle.

User Rank
more to it
DougA   8/23/2012 1:37:10 PM
I'm not sure you can blame the fuel pump design for this; every design has its tradeoffs.  Don't know your design, but the fuel level in the can is probably above the tank level when your fuel is low.  The pump suction fills the can with whatever fuel it can suck off the bottom of the tank.  It then holds it there so if your tank is low you don't run out of gas parking or driving on an incline.  Putting the sock inside the can protects it from damage from assembly and service handling and whatever debris you manage to get into your tank, and filters any debris that might be coming back in your return fuel lines say from a failing regulator seal or fuel system service.  That $600 repair could easily triple if the inectors get plugged. 

Did the techs figure out how lake water got in the tank in the first place?  Your tank should be sealed, otherwise your check engine light should be coming on and it'll smell like gas on hot days.  There has to be a vent somewhere with a check valve that allows air into the tank when the gas cools down and the vapor pressure drops (like if you dunk it in the lake or turn it off).  My guess is that vent was submerged, so you might want to add a hose to bring the opening above your waterline.    Not designing the car for boat launching might be where the real monkeys are.


User Rank
difficulty in repair
naperlou   8/22/2012 11:10:06 AM
One thing this story highlights is the difficulty of finding a problem with a vehicle, much less fixing it.  You often hear about repair shops doing an expensive repair when a less expensive fix exists.  That is, assuming you know what the less expensive fix is.  It is often easier to just replace a whole unit then to spend time to figure out what the problem is in detail. 

I had an older car that had a dodgy fuel pump.  It was an old English sportscar.  The pump was right behind the seats (it was a two seat convertible).  All I had to do was whack it with a hammer and it would restart.  Of course, I should have just replaced it, but that would have been much too easy.  If you were moving the procedure was to depress the clutch and then reach back and whack the pump.  The engine would pick up and you would re-engage the clutch and go on your way.  If you had a passenger who had never experienced this before, it was quite a trip.  Frankly, as I write this I am laughing at the silliness of it.

Later I had a European car with a fuel pump assembly much like the one you describe.  The car had 110K miles on it and had never had a major problem.  Then, one night, a relay to the pump went.  I had to have the car towed, but the repair was not very expensive.  My wife decided that we needed to replace the car based on that and the mileage.  I think that was probably an over reaction.

Partner Zone
Latest Analysis
General Motors is putting an off-road twist on hydrogen fuel cell technology with an imposing new pickup demonstrator called the Chevrolet Colorado ZH2.
Fine powder printing of industry-standard metal and ceramic powders with a grain size of less than 10 microns is now available from industrial 3D printer maker ExOne for its Innovent printer.
At ARM TechCon 2016, CEO Simon Segars will discuss how he sees billions of devices scaling to trillions as IoT applications proliferate. We know it’s happening. How do we prepare?
The term “autopilot” is now at the heart of a growing debate between Tesla Motors Inc. and Germany’s Federal Motor Transport Authority.
A make-your-own Star Wars Sith Lightsaber hilt is heftier and better-looking than most others out there, according to its maker, Sean Charlesworth. You can 3D print it from free source files, and there's even a hardware kit available -- not free -- so you can build one just in time for Halloween.
Quick Poll
The Continuing Education Center offers engineers an entirely new way to get the education they need to formulate next-generation solutions.
Oct 10 - 14, Embedded System Design Techniques™: Getting Started Developing Professional Embedded Software
SEMESTERS: 1  |  2  |  3  |  4  |  5  |  6 |  7 | 8 | 9 | 10

Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.
Next Course September 27-29:
Sponsored by 3M
Learn More   |   Login   |   Archived Classes
Twitter Feed
Design News Twitter Feed
Like Us on Facebook

Technology Marketplace

Copyright © 2016 UBM Canon, A UBM company, All rights reserved. Privacy Policy | Terms of Service