Ideas for solving fuel smell

[mikee30]

Hi all,
I am chasing down a fuel smell that is only noticeable when I park indoors or sometimes when I walk by the car. There is no smell in the cabin and I don’t smell anything when I am driving.
I removed the trunk carpet and fuel sending unit cover…no pooling of gas and no smell is noticeable in the area.
In the engine bay, there is no fuel pooling anywhere near the fuel rail or injectors.
I would like to go through the most common sources of fuel smell with the hope of finding the source. I have a spare evap canister although no smell near the driver door / fender area. Some fuel related items have been replaced but didn't help…such as the fuel filter/check valve/fpr/high pressure fuel hose near fuel filter.
I was told the previous owner changed the injectors at one point.
Looking for ideas on where to start
Thanks!

[cda951]

Hi Mike,

Your best bet for finding a fuel smell is to use a smoke machine. The various fuel vapor hoses harden and crack, and this can be hard to see. I remember a thread here or on another forum where a group of DIYer enthusiasts local to each other pooled money to buy a smoke machine to be shared amongst the group as needed. This is a great idea to amortize costs, as such a thing is quite useful for tracking down vacuum/boost leaks and fuel vapor leaks.

[dr bob]

A quick gargle search finds a good feature on the system in a Clarks Garage posting. Follow this link https://www.clarks-garage.com/shop-manual/ts-03.htm and you'll very quickly know a lot more about 944 vapor recovery systems than I do.

I'll share some somewhat generic learnings to the mix --

-- Your suggestion that the PO swapped injectors is not a bad place to start. Most troubleshooting is triggered by something that changed. This is a known and easily related change, so the effort might start at the injectors, and specifically at the fuel connections and o-rings.
---- Start with the basic fuel system leak-down test. Add a fuel pressure test gauge to the end cap of the fuel rail, and use a switched jumper in the fuel pump relay socket to pressurize the system and verify the system holds pressure correctly.
------ My handy-dandy fuel pressure test package now includes a spare rail cap, drilled and tapped for 1/4-20 to match a GM test port thread. Then a bargain-cheap Harbor Freight fuel pressure test kit that includes the GM port adapter hose. Instead of the ball, I put a 1/4" brass compression fitting sleeve inside that cap, so fuel from the rail can make it to the gauge, yet the sealing on the cap is maintained. You can probably free-rent a fuel pressure test kit from a better local parts store, and it's likely to have the GM port adapter.

After that, you might be searching to seepage around injector o-rings at the intake. My nose has been somewhat neutered from some early-employment episodes with perfume manufacturing, so I depend more on local 'touchy-feely sample-then-sniff' methods for isolating leaks. Table your first thoughts from that description, then consider: A tab of paper towel held against a suspected fuel leak might actually get wet from the fuel. Or it might absorb enough vapor for my nose to detect up close. Wear Nitrile Gloves for this; You really don't want to have fuel absorbed into skin. Sniff your test sample swabs progressively closer to your nose, for the same reason-- Both nasal and optical body parts are very HC absorbent and close enough to risk brain damage. I'm stumbling evidence of that risk, so trust me on this.

If the system fails the pressure and leak-down tests, you get to find out why and where the fuel is leaking. FPR's and dampers have internal diaphragms to get tired, and leak fuel through vacuum ports. As an easy example. Other connections, including any that were opened for the injector swap, are of course suspect. Also filter and pump connections that are opened as part of normal maintenance. I add a film of PTFE pipe thread sealant/lubricant to the ball-and-seat connections as well as to the threads where the nuts engage. Saves on leaks and fitting damage from misalignment and over-torqueing the pieces trying to stop a leak. Can't say enough for lubricating the injector o-rings top and bottom where they fit the rail and the manifold. Saves scuffing and o-ring distortion during assembly.

If the system passes the pressure and leak-down rate tests, move on to the vapor cannister and the related plumbing pieces. The vapor system is functionally extremely simple, with a hose functionally from tank vent to engine airbox, also connected to a port on the vapor cannister if the front wheelhouse. There are some interesting pieces in between, naturally.
-- At the tank filler, there's a sort of liquid trap intended to keep filler spillage or overfill liquid out of the vapor system. Design varies some between early and late cars but for our purposes the fumes diagnosis is the same -- sniff test supplemented by visual inspection for leaks.
-- There are hose connections from that section, connecting to steel lines under the car, forward to the engine bay right side. From there it's a hose to a T, one leg to the cannister, and the other to a vacuum-operated valve connected to the intake. I don't know if that T fitting is plastic in the 944; it's a plastic Y in my 928, and is a popular failure point that generates the fuel smell symptom you describe.
-- The system really depends on having the hydrocarbon vapors from the cannister get pulled into the warmed running engine. If that purge doesn't happen, the activated carbon/charcoal in the cannister will be overwhelmed by the HC load, and you'll start to get fumes from the cannister vent to atmosphere.
-- The vapor system is intended to work under slight pressure and vacuum, as the fuel temperature changes with engine off or cold. Fuel gets bigger or smaller with those changes, and the pieces need to contain the vapors while the engine is off regardless of those changes. Chasing leaks in the recovery system is easy if you have access to the smoke machine mentioned by @cda951 above.

I've had some luck using a very small amount of air pressure to test for and sometimes find leaks in vapor recovery systems. In every case though, the leaks I eventually found were large and obvious to the eye and nose. The pressure needed is fractions of a PSI, like in the 6-11 inches-of-water-column range similar to what you might see in a domestic natural gas system. With that in mind, you can use a hardware-store gas pressure gauge to do a little leak-down testing. Use a rubber tire valve to pressurize with air with the gauge connected (hint- fits inside the 15mm vapor hose at the cannister...). The vapor hose in the system is unreinforced/unlined, so it's generally OK to pinch to isolate sections for testing.

Commercial 'smoke machines' use a heated element to oxidize common mineral or 'baby' oil for smoke, and a small fan or blower to push that smoke into and through the system under test. Were I thinking emergency DIY smoke machine, the electric ignitor from a home gas furnace and a surplus blow-drier for the fan/blower might step up for the duty. Do all this outside so the smoke and possible fire don't risk the car or your home. Or worse-- Do Not return your wife's favorite blow-drier after using it to blow smoke. No amount of further 'splainin' (read: smoke-blowing) will save you.

[barnwerks]

You have not shared the model year but if an early car (up to 85), the steel fuel tank is a common source of fuel or fuel vapor leaking from cracks in the top skin of the tank.

[mikee30]

Thanks all for the replies and help - the car is an 86 951.
On a different note, while I am chasing this down, I noticed almost all my fuel lines are original.
Which would be the recommended lines to replace to avoid fires in the future? I see the two that go over the engine as the most obvious and the smaller hose near the dampener....anything else? Is Lindsay Racing and the Porsche dealer the only options for those hoses?

[dr bob]

<<...>>

On a different note, while I am chasing this down, I noticed almost all my fuel lines are original.
Which would be the recommended lines to replace to avoid fires in the future?

<...>

My personal recommendation is to replace ALL of them. At this point the rubber bits in your car are 40 years old. There have been way too many otherwise nice cars lost to car-b-que events from old hoses and owner-induced leaks and damage. Fuel hose technology has advanced quite a bit in the time since your originals were the best available.

On my 928 anyway, I was able to source premium modern hose and have the original lines rebuilt using the original end fittings. I ended up buying my own crimping tool for the AC hoses in my 928, and used the same tool for fuel, power steering, and the automatic transmission cooler lines. On the 928, there's a critical cross-connect hose between the cylinder banks at the rear where the original bend radius is too small for common hose without risking collapsing the plastic hose core; Probably not an issue in the one-bank-of-cylinders 944, but a consideration wherever there are tight-radius bends in hoses in your car.

The vapor recovery hoses are generally slip-and-clamp connections, and the hose is pretty generic fabric-wrapped stuff used on many other cars from the corporate family. I use a cutter intended for PVC sprinkler pipe to get clean square cuts, as they look a bit prettier than the shredded hose ends I get with scissors/shears or diagonal-cutter pliers.

-----

I'd kind of forgotten how fun and exciting it was replacing/rebuilding all the fuel system hoses. I may need to offer a local clinic on that sometime soon, so other central Oregon locals can at least see what's involved and plan/schedule their own work if they aren't a clinic car.

I often see cars with some but not all of the hoses replaced or rebuilt; while some may be be a bit of a challenge accessing, particularly those at the fuel tank itself, It's really cheap insurance for our generally irreplaceable toys. PCA HDR members, watch for a spring clinic invite. It will force me to spring-clean the workbay if nothing else.

[Tom]

<<...>>

On a different note, while I am chasing this down, I noticed almost all my fuel lines are original.
Which would be the recommended lines to replace to avoid fires in the future?

<...>

My personal recommendation is to replace ALL of them. At this point the rubber bits in your car are 40 years old. There have been way too many otherwise nice cars lost to car-b-que events from old hoses and owner-induced leaks and damage. Fuel hose technology has advanced quite a bit in the time since your originals were the best available.

On my 928 anyway, I was able to source premium modern hose and have the original lines rebuilt using the original end fittings. I ended up buying my own crimping tool for the AC hoses in my 928, and used the same tool for fuel, power steering, and the automatic transmission cooler lines. On the 928, there's a critical cross-connect hose between the cylinder banks at the rear where the original bend radius is too small for common hose without risking collapsing the plastic hose core; Probably not an issue in the one-bank-of-cylinders 944, but a consideration wherever there are tight-radius bends in hoses in your car.

The vapor recovery hoses are generally slip-and-clamp connections, and the hose is pretty generic fabric-wrapped stuff used on many other cars from the corporate family. I use a cutter intended for PVC sprinkler pipe to get clean square cuts, as they look a bit prettier than the shredded hose ends I get with scissors/shears or diagonal-cutter pliers.

-----

I'd kind of forgotten how fun and exciting it was replacing/rebuilding all the fuel system hoses. I may need to offer a local clinic on that sometime soon, so other central Oregon locals can at least see what's involved and plan/schedule their own work if they aren't a clinic car.

I often see cars with some but not all of the hoses replaced or rebuilt; while some may be be a bit of a challenge accessing, particularly those at the fuel tank itself, It's really cheap insurance for our generally irreplaceable toys. PCA HDR members, watch for a spring clinic invite. It will force me to spring-clean the workbay if nothing else.

Have to agree with all of this -- any gas leak is a fire risk. That said, if you need to prioritize, I'd start with the lines under the hood going to the fuel pressure regulator and damper, and the jumper from the rail to the damper. Rennbay makes a nice kit (and factory parts are still available too). Those lines sit above the header, so have been exposed to more heat than most and are the ones you most often hear about failing. And, of course, when they leak, they drip gasoline directly onto the hot headers with predictable results!

[cda951]

Have to agree with all of this -- any gas leak is a fire risk. That said, if you need to prioritize, I'd start with the lines under the hood going to the fuel pressure regulator and damper, and the jumper from the rail to the damper. Rennbay makes a nice kit (and factory parts are still available too). Those lines sit above the header, so have been exposed to more heat than most and are the ones you most often hear about failing. And, of course, when they leak, they drip gasoline directly onto the hot headers with predictable results!

I'm not a fan of the aftermarket stainless braided fuel hose kits for the 944 fuel rail feed and return hoses----there is no way to inspect the rubber inner portions of said hoses.

What I have done on my own car and for dozens of customer 944s is (carefully) cut the crimped ferrules of the factory hoses at the portion where they exit the inner fender, and use high-quality fuel injection hose and stainless fuel injection clamps, along with either re-using the factory heat sleeves or aftermarket versions. This way the hose sections can be easily inspected and replaced as needed (every 5-6 years or so). Then pay the money for the Genuine Porsche short jumper hose between the fuel rail and damper, and rest assured that you have done what you can to prevent engine bay and forest fires.

[Tom]

Have to agree with all of this -- any gas leak is a fire risk. That said, if you need to prioritize, I'd start with the lines under the hood going to the fuel pressure regulator and damper, and the jumper from the rail to the damper. Rennbay makes a nice kit (and factory parts are still available too). Those lines sit above the header, so have been exposed to more heat than most and are the ones you most often hear about failing. And, of course, when they leak, they drip gasoline directly onto the hot headers with predictable results!

I'm not a fan of the aftermarket stainless braided fuel hose kits for the 944 fuel rail feed and return hoses----there is no way to inspect the rubber inner portions of said hoses.

What I have done on my own car and for dozens of customer 944s is (carefully) cut the crimped ferrules of the factory hoses at the portion where they exit the inner fender, and use high-quality fuel injection hose and stainless fuel injection clamps, along with either re-using the factory heat sleeves or aftermarket versions. This way the hose sections can be easily inspected and replaced as needed (every 5-6 years or so). Then pay the money for the Genuine Porsche short jumper hose between the fuel rail and damper, and rest assured that you have done what you can to prevent engine bay and forest fires.

For what it's worth, the Rennbay one's aren't covered. They are the simplest of the aftermarket patch kits, with a custom coupler to make it way less bulky. I have one Rennbay line and one new factory hose in my car. The factory ones come with a section of hard line under the car and the rubber already crimped on. I would have done both with factory replacements but, despite what PET says, one of my hard lines goes all the way back to the pump area with no coupler in the middle, meaning I'd need to drop the transaxle and/or suspension to replace it. The factory parts comes with a shiny heatshield sheath (unlike the original versions with the thick rubber casing), so I guess they suffer from the same problem as braided lines unfortunately. My old kit was indeed covered in a nylon braiding, and sure enough it sprung a leak despite looking like new. On the up side, I found the leak without any special tools.

[dr bob]

Consider that modern-spec injection hoses are really plastic tube with various reinforcing and protection layers added. I'll suggest that wanting to see the condition of the outer rubber-ish layers is interesting, but unless there's fuel flowing out through cracks it's not a reliable inspection strategy.

The plastic tube fits over specific shape 'barbs' on the tube fittings that slide inside. The clamping force applied by a crimped-sleeve is pretty much right at the end, just enough to hold the end small. Adding regular clamps right at the end is OK, but counterintuitively you really really don't want to clamp around or even close to the fitting barbs like you might with a rubber-lined hose. The plastic liner in the modern FI hose will fracture over those barbs. You might not notice it right away, because the supporting layers will contain and eventually redirect the leaks. Of course, when we see such a leak near a connection the first thing we try is tightening the clamp that caused the leak.

The plastic liner has limited tolerance for sharp bends, as the plastic tube liner will kink inside the supporting outer layers. You might not see it, but again the outer layers can contain and redirect the direct flows from the eventual fractures that happen at the kink. They might look like end leaks, and tightening clamps there might slow or temporarily stop the apparent leaks.

These plastic lined hoses have absolutely zero tolerance for 'hose pinchers' that get used to isolate sections for testing. A pinched hose that way is a ticking fire-bomb.

This is really my long-winded recommendation that we pay close attention to the installation requirements for the modern hose. My local hydraulics shop refused to even crimp the end sleeves on a fuel hose for me. Use your best sense when you replace any of these hoses. SAE publishes some pretty detailed guidance on the various hose and fitting options, along with use and installation warnings.