Vapor lock challenge

After having the engine stall from this issue, the engine would start fine after waiting about 20 to 30 minutes for the fuel to cool down. Until the fuel cooled, the engine wouldn't restart at all.
Sounds like injector heatsoak (do a google on that, happens a lot with aftermarket ECUs).
Injectors have an impendance rating. The ECU calculates opening times based on a given injector impendance.
When injectors get hot (like on a refill fuel stop) the impendance changes and as a result the opening duration shortens (way less fuel) and as the ECU doens't know that, it doesn't react on this temporary change still thinking it sprays the same amount of fuel as injector impendance is a given number and so on the car won't fire up as it actualy doesn't get enough fuel.

In OEM situations the pressure regulator raises fuel pressure on a hot start to overcome this (less opening of the injector + more fuel pressure creates the amount fuel needed).
With aftermarket and standalone ECU you don't have this feature. Holley Sniper is known for this issue aswel as others.

I overcame this by fitting a return line on the other end of the fuelrail so fuel gets circulated on a heatsoak start and the injectors gets cooled by the fuel circulating through the system It also pushes the vapourlock out of the system..
I also raised fuel pressure from 43 psi to 58 psi and recalculated flowrate of my injectors.
My Bosch 044 pump now flushes the system way faster and my heatsoak hot starts are vanished.
 
Sounds like injector heatsoak (do a google on that, happens a lot with aftermarket ECUs).
Injectors have an impendance rating. The ECU calculates opening times based on a given injector impendance.
When injectors get hot (like on a refill fuel stop) the impendance changes and as a result the opening duration shortens (way less fuel) and as the ECU doens't know that, it doesn't react on this temporary change still thinking it sprays the same amount of fuel as injector impendance is a given number and so on the car won't fire up as it actualy doesn't get enough fuel.

In OEM situations the pressure regulator raises fuel pressure on a hot start to overcome this (less opening of the injector + more fuel pressure creates the amount fuel needed).
With aftermarket and standalone ECU you don't have this feature. Holley Sniper is known for this issue aswel as others.

I overcame this by fitting a return line on the other end of the fuelrail so fuel gets circulated on a heatsoak start and the injectors gets cooled by the fuel circulating through the system It also pushes the vapourlock out of the system..
I also raised fuel pressure from 43 psi to 58 psi and recalculated flowrate of my injectors.
My Bosch 044 pump now flushes the system way faster and my heatsoak hot starts are vanished.

I suspect fuel boiling at the fuel pump impeller more than injector heat soak at this point. Maybe I'm wrong but relocating the fuel pump into the tank will tell me. This car is using a GM ECU (early 2000 Camero). I don't know if it is measuring injector impedance or not but I'm hoping the vapor lock issue goes away with in tank pump change. We'll see. This is one of my planned winter projects so it will likely be next summer before I have an answer.
 
Gas Tank Fabrication

The GTO has been undergoing the homebuilt car sorting out process and has just over 2,000 miles on it now. One of the issues it has had since I moved from a location at 800 feet elevation to a location at 5,100 feet has been a “vapor lock condition” that can occur when ambient temperatures are in the mid 90s F and when climbing a steep grade. Given the EFI, it’s not the traditional vapor lock, it happens when the fuel boils and turns to vapor just prior to the fuel pump thus resulting in no fuel pressure and a stalled engine. After a 30 minute or so break, the fuel cools enough that liquid fuel gets to pump, fuel pump starts working again, and the engine restarts. No real harm, just a grumpy, disgruntled and now overheated driver.

The required fix is to move the fuel pump into the gas tank similar to an OEM factory EFI configuration. Having the fuel pump draw fuel directly from the bottom of the tank virtually eliminates the pre-conditions for this type vapor lock.

Back when I first started building the GTO, I had aspirations of using it as a “track car” and built it with a foam filled bladder type fuel cell using an fuel pump external to the cell. After a 20 year plus build, I decided that it was more prudent to just use the GTO as a street car. My first thought when the vapor lock started happening was to just relocate the fuel pump into the existing fuel cell even though it had gone “out of date” many years ago. Subsequently, I was told by several people that doing this is flirting with more issues as the open celled foam inside tends to break down and start clogging up the fuel system with foam debris.

So I thought, what the heck, I built the car from scratch, why not build a new fuel tank from scratch. The fuel tank is rectangular in shape (34” wide, 17 ½” deep, and 9 ¾” tall) so that should be easy enough to build from aluminum sheet. I’ve had a sheet of .090 aluminum gathering dust for a few years so I already have the raw material on hand.

P1170606.JPG


P1170605.JPG


To minimize the welds, I decided to make the bottom and sides from one piece and the top from another. That actually sounds easier than it is unless you have access to the proper folding machines. I found a neighbor with a finger break I thought was larger enough but it wasn’t. Yes, it could bend the .090 sheet but only 2 of the 4 sides could be bent to 90 degrees. The other 2 sides could only be bent 45 degrees as the already bent side hit on the break top bar.

The proper tool for folding a rectangular box this deep is a press brake and I should have sought out one of those. Fortunately, I was able to use the flanging dies on my TM Tech power hammer and some muscle to move the 45 degree folds over to 90 degrees in the .090 thick sheet. The corner seams on the resulting piece were close but had too much gap for good welds. I broke out my TM Tech flow forming tool (i.e. rivet gun with hard plastic hammer head), clamped heavy metal bars inside the box, and “moved” the corner on the short folded sides about 1/16” inward so vertical seams were now tight together.

My plan is to reuse the fill plate and fuel collection box from the fuel cell. The fill plate already has the 2 ½” fitting for tank fill, -8 outlet, -8 return, -8 vent with rollover protection, and fuel level sensor mounted on it. I just need to drill and mount an electrical wire bulkhead fitting into it for fuel pump power.

The plastic fuel collection box has one way valves on 3 sides to let fuel in but not back out. The EFI return line will be plumbed so it drains directly back into the collection box. I’m using a compact Walbro fuel pump which can be mounted with a simple 90 degree bracket inside the collection box. The pump is positioned with a fuel sock directly on the box bottom. This way fuel is gravity fed to the pump impeller until the tank is almost completely empty which should prevent the vapor lock issue.

P1170607.JPG


Onto tank baffling and this is where I could use some input. In my mind, the baffling has two purposes: 1) minimize fuel slosh during acceleration and cornering, and 2) reinforce the fuel tank from the inside out against deformation or bursting. This tank mounts to the chassis with 3 under hung 1 ¼” straps and is free standing in that there are no external bulkheads on the sides or bottom. I believe the .090 thick aluminum is plenty strong for this tank but want to reinforce it well for extra strength/protection. 20 plus gallons of fuel when tossed about with fast acceleration, cornering, and braking can pack a wallop.

My first thought for the baffling was to use 3 across the short direction and 1 across the long direction egg crate style.

P1170600.JPG


P1170601.JPG


The collection box would be held in position between 2 of the baffles with openings in the baffles for 2 of the one way valves. Thus the collection box is fed from 3 different sections of the tank. For assembly and servicing, an access way is needed to the collection box so pump, hoses and wiring can be accessed. For this, the baffle closest to fill plate needs to have a portion that can be easily removed and reinstalled through the fill plate opening.

My first thought was to weld all the baffles to the tank sides and bottom directly. Upon thinking about this some more, I think it might be wise to weld some small angle brackets to the tank sides and bottom then attach the baffles to the brackets with either machine screws or pop rivets. The reason being that there isn’t much room for a TIG torch, filler rod and my big hands inside once the baffles are in place. Also, sightlines to the weld zones will be very restricted and I hate trying to weld without direct line of sight. Anyone have experience with something like this, please speak up with advice.

The other part about the baffles where I could use some input is design. In other words, hole quantity, size, and positioning. I want fuel to readily flow from chamber to chamber during tank filling but resist movement between chambers except for when flowing to the collection box as the tank empties. I’m thinking a couple of larger sized holes up high in each baffle for purposes of tank filling and fuel slosh shouldn’t be a factor when the tank is greater than ¾ full. Then a few smaller holes per baffle, maybe ½” in size, down low so fuel can move via gravity from one chamber to another as tank empties but not a lot of movement due to fuel slosh.

Any input on baffle design would be greatly appreciated. Please include the “why” along with the “how” so we all can learn something here.

In addition, if there’s anything I’ve described as part of this fuel tank project that doesn’t sound right or safe, please don’t hold back. This isn’t my first fuel tank project but I certainly don’t claim much expertise on the subject either.
 
On my Cobra, I made trap doors like you see in an oil pan.

In the center section, I had a 1 gallon collector tank fed with a venturi pump powered by the return fuel.

The sender unit is on one side, and with 1/4 tank, I can swerve left and right a couple of times, and make the gauge read empty for 10-20 second intil the fuel leaks leaks back to evenly distributed.

I've ran out of gas twice, but NEVER had a pump starved stall or noticeable hesitation.
 

Attachments

  • sm1201.jpg
    sm1201.jpg
    95.9 KB · Views: 35
  • sm1196.jpg
    sm1196.jpg
    173.5 KB · Views: 37
On my Cobra, I made trap doors like you see in an oil pan.

In the center section, I had a 1 gallon collector tank fed with a venturi pump powered by the return fuel.

The sender unit is on one side, and with 1/4 tank, I can swerve left and right a couple of times, and make the gauge read empty for 10-20 second intil the fuel leaks leaks back to evenly distributed.

I've ran out of gas twice, but NEVER had a pump starved stall or noticeable hesitation.

Now that's a very sophisticated gas tank!! I've had fuel starvation issues upon hard acceleration in my Cobra when the fuel level gets low. That's what I'm hoping to prevent with the GTO.

Trap doors in the baffles sound like some fabrication work but should be simple enough to be maintenance free. Once the tank top is welded in place, there won't be access to other than the baffle next to the fill plate. So I need to make sure all the other baffles are bullet proof.

What size trap doors did you use? I mocked up a 3" by 3" in cardboard just to get a sense of the fabrication work. I'm guessing smaller, maybe 1" by 1" would probably work fine.
 
Going from memory and ball park scale size, I think they were about 1.5" tall and 3" wide. They're far from a perfect seal, but tweeked to lay flat with a little pressure.

Actually, cutting a window into the baffle and adding the door wasn't a big deal. Do it all on the bench before you weld the baffle in and it's easy.

It's a blessing and a curse. There is no warning. The internal swirl pot scavenges the last gallon from the tank and the trap doors keep fuel in the center when there are twisties, and with that first hesitation, your dead dry and stranded. You MUST watch the gauge, and the last 1/8 goes the quickest.

I couldn't find the original full res pics. 1196 shows a 3/8" return line that's pinched down to a 3/32" jet and it blows into a 1" tube to act as a venturi pump. I noticed a similar feature in a car I had that I had to replace the fuel pump and integrated it. Pic 1201 you can see the box then tube in the right side of the swirl pot. I had to add the hood on the top to roll the returning fuel plus what the venturi pump added back into the pot. I ran it in the swimming pool with a junk pump and it was a surprisingly huge flow of water being sucked up by the venturi pump. I would guess 10X what the pump was moving.
 
Gas Tank Fabrication (Part 2)

After thinking about the feedback I’d received from multiple sources, it lead me to these choices: extra bracing for the side corner weld seams, addition of a sump in the tank bottom, inclusion of a venturi jet pump powered by the EFI return fuel, and using baffles attached to tank sides via welded on brackets.

P1170625.JPG


The tank side corners were first welded on the outside. I then did a welding fusion pass on the inside to ensure complete weld penetration through the sheet. These gussets were made from the same .090 sheet as the tank and welded into place. I don’t know that they were actually needed but better safe than sorry.

P1170620.JPG


A blister was metal shaped into the tank bottom forming a 3/8” deep sump. I also curved the tank bottom to the center and sloped so fuel will gravity feed to this sump.

P1170628.JPG


The sump is located under one of the baffles so it can serve as a mount point for the pump. In testing out this venturi pump, it draws about a quart of liquid up for every gallon pumped.

P1170627.JPG


Here’s the tank with all the baffles installed and fuel collection box in place. Each baffle has two 1 ¼” holes near the top so fuel can easily pass through during tank filling. There are no other holes in the baffles except those going into the collection box. There are 3 one-way valves in the collection box, each going into a separate baffled tank section. There is a small gap under the baffle to let fuel pass by as the tank is drained. The gaps are fairly small but widest in the center where the tank bottom is sloped. The baffle next to the fill plate is two pieced so the fuel pump and collection box can be serviced through the fill plate opening.

I did some testing using about 3 gallons of water to see how the baffles would work. My first observation is that the one-way valves in the collection box are very good at what they do. After a couple of swishes of water around the tank, the water level in the collection box was two inches higher than the rest of the tank. I felt the baffles were letting too much liquid under them to fast so I bent the lower edges downward to lessen the gap. I then hooked up an old electric fuel pump to test the venturi pump. The tank emptied out quickly with all the water going to the sump via gravity.

Ok, I feel good about the tank internals from the testing I did. Now it’s on to fitting up the fill plate to the tank top and welding it on.
 

Ian Anderson

Lifetime Supporter
Will the baffles need a hole at top edge to allow air to exit when filling or the level will be limited to the top of the transfer hole?

Easier to drill them now

Ian
 
Gas Tank Fabrication (Part 3)

The top panel is now welded in place with big fat weld beads on all four sides. My lesson learned from prior gas tank fabrication experience is not to skimp on the filler rod. You’re more likely to get pin holes in the TIG weld if you don’t have a nice “fat” puddle adjoining both sides of the weld joint so I used 3/32” filler rod with frequent dabs into the puddle.

P1170629.JPG


Now on to the Quality Assurance part of the tank fabrication with leak testing. I cobbled together some PVC fittings to adapt a low pressure gauge to the 2 ½” filler hose. The closest PVC fitting I could find had a 2 ¼” OD so I wrapped 1/8” thickness of duct tape around it which mostly worked.

You can see from the soapy water test where the leak issues occurred. The fuel level sending unit held air about as good as a cheese cloth balloon. I now know the source of the “gas smell” when the prior fuel cell was full. I tried thicker nylon washers on the hold down screws and even added nylon washers and nuts on the backside of fill plate but the sending unit continued to leak. A second hose clamp on the fill hose helped slow the leak there but I couldn’t stop it completely.

I now have a new fuel level sending unit on the way and built a block off plate to continue the leak test. The leak test consisted of using a bicycle pump to bring the tank up to 2 psi of pressure. Soapy water was sprayed over the tank seams and fill plate to check for leaks. After 24 hours, the gauge read 1 ¾ psi and the only soap bubbles were on the duct tape around the fill hose adapter. I concluded from this test that none of the weld beads were leaking and the tank is now good to go.

P1170631.JPG


Given the exhaust pipes that run just under the fuel tank, I decided to apply some insulation. I used the DEI self adhesive tunnel/floor insulation that has an aluminum skin on the outside. The insulation edges were sealed with aluminum duct tape to keep water and oil contaminates from getting into the insulation.

Once the replacement fuel level sending unit arrives, final tank assembly can be completed. It’s then onto installing the tank, filling it up and conducting some field testing for the fuel pump, etc. The real test for the vapor locking issue will be this coming summer when hot temperatures arrive but even if the fuel pump “noise” is reduced then that will be a nice improvement.
 

Howard Jones

Supporter
Great stuff for sealing those fill large plates and fuel sender. It can be disassembled but you will need new gaskets and it takes some time to scrape off all the surfaces. It will clean up with Acetone.

Don't overdo it. A little thin film goes a long way. I had to find something to seal my big fill plates and fuel sender.

Read the TDS tech doc. It specifically says it will work with Gasoline.

 
Last edited:

Neil

Supporter
Great stuff for sealing those fill large plates and fuel sender. It can be disassembled but you will need new gaskets and it takes some time to scrape off all the surfaces. It will clean up with Acetone.

Don't overdo it. A little thin film goes a long way. I had to find something to seal my big fill plates and fuel sender.

Read the TDS tech doc. It specifically says it will work with Gasoline.

I've used Hylomar with good results.
 
Back
Top