Well, this was kind of frustrating. I was out running errands today with the AC blowing ice cubes out the vents (not really, but the cold air was very refreshing) when I get a strong odor of gas that is really bad. I was almost home so I rolled down the windows and switched fuel tanks and the smell went away. Switching fuel tanks is my go-to solution when I have an issue and this time it worked and let me get home.

When I get home I shut it off and start looking under the hood. Everything looks ok so I open the trunk and see puddles of fuel. I turned the ignition on and see fuel spraying out of the electrical connector on my 6 port Pollak valve that is used to direct excess fuel to the return tank. Not good.

I pulled the valve out of the trunk and I am amazed at how much fuel is in this thing. It keeps leaking out; the housing was obviously full of fuel.



I want to see what's wrong so I drill out the 4 rivets and immediately see the problem. The washer that you see on the left side is out of position by a large amount which let fuel bypass the seal, fill up the valve housing, then spray through the electrical connector. I am not sure what caused this to fail - the valve is rated for 65 psi and I am below that. It could have been an electrical failure but it worked perfectly for over 1,000 miles.


I am not sure what my play here is. I can pick up a new valve and cross my fingers it doesn't fail again. I think I am going to explore my options and see what else is out there. I'm pretty disappointed because I had the same setup on my XJ6 and never had any issues with the valve.

 

Yikes, that a little scary but glad you got it found and sorted fairly quickly

 

Could it be a manufacturing defect and that washer was never installed in the correct position to begin with?

 

Don,
My opinion - if it was a manufacturing defect it would have failed quickly. This worked for over 2 months and 1,000+ miles. I suppose the seal could have gradually worked itself out of position over time.

Thomas

 

Are there any witness marks on the plastic body showing the washer was pushed out of its original position?

Maybe the valve body is sealed and it took a while for the gasoline to break through the seal?

 

Thorsen, as PM'd.

 

Now that I have had time to think things over, I have come up with the following potential solutions with the pros and cons:

#1 - Replace the failed Pollak valve with the exact same thing
+ Fast
+ Easy
- I have lost faith and will never trust it

#2 - Replace the failed Pollak valve with electrically activated solenoid valves
I am considering valves like this
+ Viton seals are compatible with gasoline
+ Will provide positive sealing
- Failure modes are unknown but possibly similar to #1
- Not rated for continuous duty but they say don't use it for more than 8 hours. On long trips I switch tanks every hour so this is less of a concern - but still a con

#3 - Abandon the twin fuel tank setup and switch to a single fuel cell in the 20 gallon range
I am considering a fuel cell or https://www.summitracing.com/parts/sum-295220-s+ Great simplifies the fuel system and eliminates the need for valves
+ New tank removes the worry about the condition of the original fuel tanks, specifically the liner falling apart
+ The one from Summit Racing uses an in-tank pump which will be quieter
- The Evil Energy tank will need an external pump (louder than in-tank pump)
- Having to open trunk to fill fuel tank (could possibly be offset by using a filler neck connected to driver's side fill port)

#4 - Greg's solution in post 726
Using two high pressure fuel pumps and electrically controlled valves
+ Simplifies fuel system by eliminating the need for the low pressure fuel pumps and the swirl pot
- Still relies on some sort of electrical valve to switch between the pumps as well as directing the excess fuel to the correct tank

I am curious to hear the group's thoughts and opinions, as well as solutions I might have missed or pros/cons I have not considered. I'm still pretty salty about yesterday and the fact I will be missing tomorrow's Coffee and Classics where British Classics are the feature, but I will share that I am leaning towards solution 3.

 

Thorsen
Many thanks for posting the very clearly explained options. I too will be interested to hear the collective thoughts. My own are, so far anyway, to take it in stages, rather than try to fix it in one hit.

Now the idea as in my PDF has only just occurred to me, and I apologise to Thomas for not having thought of it when we were discussing the system two months ago. Given the tank shapes I reckon a swirl pot is not needed as long as the tanks gravity feed to the pumps. The tanks pretty well are swirl pot-shaped.
So I suggest hooking up one tank to a pump under the car and testing the idea, easy to do and what's to lose? If it works OK and no acceleration starvation etc etc, the battle is won - providing a decent valve can be sourced. In the event of hard acceleration slight starvation for a second or two, there is a dead simple remedy: the hose from the tank to the pump needs to be quite a large diameter, even if it needs adaptors each end, say 5/8 or 3/4. The hose needs to be gasolene resistant, obviously, but as thin-walled as can be found. Then when the engine FPR calls for more fuel, the big hose can act like a mini-swirl pot as it can flatten slightly to provide the extra fuel to the pump for a second or two while gravity catches up.
I have actually very successfully used this idea on a Range Rover I converted from carbs to F.I. It works and is reliable.

Valves:
I followed the link to the solenoid valves, and the "do not energise for more than 8 hours warning is pretty explicit". I think any system that relies on a solenoid being continuously 'hot' is a breakdown guaranted to happen. If there were electrically activated valves where the motor did the work of a manually turned valve, that would be far better. My quick search revealed that there is such a thing as a 12v motorised ball valve which would perhaps do the job more safely, if one can be found for gasolene?

So let's say the test does not work, or that reliable valves are not to be found, then I vote for the new tank idea.

 

I do not want to commit heresy, but I found this thread:
https://forums.jag-lovers.com/t/alte...r-valve/367081

 

Greg,
I did find these valves which have viton seals (safe for gas) and are rated at 100% duty cycle.
I also found some of the 3-port valves but I just realized - with all the motorized ball valves the engine can't be running to switch tanks. The motorized ball valves take a few seconds to full open/close and I don't think I want to dead-head the fuel pump against an opening valve. It would also cause the engine to run lean for a few seconds which isn't great.

Thomas

 

I don't know much about fuel systems, but how about modifying the tanks so that you have :
-one "main" tank with a LP pump towards the engine and the return system
-one "auxiliary" tank with a manually activated pump that transfers fuel to the main tank

This would eliminate the need for a switchover valve. You could go further down the rabbit hole and add a "lock" to prevent running the aux -> main pump if the main tank is full. Or automate the fuel transfer entirely.

I'm always a bit sad when I see a fuel cell in a non-racing car, as you lose that "wow your mods are so sorted this car could have left the factory like this" factor, which your car still has!

 

So ball valves no good, pity. The spec sheet on your link says in tiny print "*These are not intended for use in medical life support, combustion, aviation, aerospace, automotive or similar applications". Maybe this does not actually matter, but I wonder why it says that.
https://www.electricsolenoidvalves.c...Spec_Sheet.pdf

 

I had missed the fine print - thank you for pointing that out!

 

Since the car is non-operational for a bit, it's time to start addressing some things.

First up - since I had to pull everything out of the trunk to air it out from the gasoline bath it took I decided this was as good a time as any to address the surface rust on the trunk floor.



After wire-brushing, vacuuming, and cleaning it I put down a layer of rust converter. I don't understand the chemistry but this changes the iron oxide (rust) to black oxide (not rust) or something like that. I used it 33 years ago on my MGB and it's held fast so I trust it.


While I was waiting for the rust converter to dry, I moved on to the part that I was planning on getting to tomorrow after Coffee& Classics. New window channels and NDV (No Draught Ventilator) seals. First up was removing the door panel.


I had to switch from 3/16" felt to 1/8" felt, otherwise the window would not go up and down.
I'm not sure why they advertise this as grey felt - it's not even close to grey.

I had planned on using more modern window channel with flocked rubber, but they did not fit.


I laid a piece of tape across the top of the door to mark the hole locations for the beaded trim that goes between the door and the outer weatherstrip.


Here is a picture showing the beaded trim and the new outer fuzzy weatherstrip.


The beaded weatherstrip continues to the front of the door where it turns 90 degrees and is between the top of the door and the NDV aka vent windows. Here it is being glued in place.
I didn't get a picture but while I had the door apart I pulled the glass and put a light tint on it. It's 70% tint (ie it allows 70% of light to pass) which isn't dark at all, but it's ceramic and has UV blockers to keep the sun from baking you on long drives.


After giving the rust converter 4 hours to dry, I top-coated it with black paint.
After the paint dries I need to pull the roll of heavy Hardura off the shelf and make a new trunk liner.