Gas valve kicks off while filling tank

sab said:

Not a leak in the EVAP system, but a clog. The way the system works is that, during filling, the liquid volume in the tank increases, so the vapor volume has to decrease. It can either decrease from an increase in pressure in the tank or from venting of vapors from the tank. The system doesn't allow much pressure increase (only about .02 psi).

What's supposed to happen is that, as the tank is filling, the pressure rises to about 0.02 psi, and the Natural Vacuum Leak Detection (NVLD) pump diaphragm opens up, allowing the tank to vent from the charcoal canister through the vent filter to the atmosphere. So, if there is a clog anywhere from the tank vent to the filter, the system won't vent properly, the tank pressure rises, and that shuts off the gas pump.

Here's what I'd do to diagnose the situation. First, disconnect the vapor line between the tank and the NVLD pump:
View attachment 702958

Now try to fill the tank. If it fills without the pump shutting off, then you know your issue is with the EVAP assembly. If it's still shuts off, then your problem is with the filler neck or tank.

Assuming your problem is with the EVAP assembly, the next thing to do is to remove that assembly so that you can test it. There are six components that could be clogged:
View attachment 702962
1. Charcoal canister
2. Vent filter
3. NVLD pump
4. Hose between tank connection and charcoal canister
5. Hose between NVLD pump and vent filter
6. Hose between charcoal canister and NVLD pump

Here's how the vapors leaving the tank during filling travel through the system:

Tank —> Hose #4 —> Charcoal Canister —> Hose #6 —> NVLD Pump —> Hose #5 —> Vent Filter —> Atmosphere

You need to make sure that air flows through each of these components freely. With the NVLD, the only way you can test the flow is in one direction due to the diaphragm valve. You have to push air in where Hose #6 enters and see if air leaves where Hose #5 exits the NVLD pump. You can do it the other way (Hose #5 to Hose #6) because the diaphragm valve won't open.

If any of those six components won't flow air, that's what's causing your fill pro

Click to expand...

sab said:

I appreciate your expertise. I will definitely investigate to find the problem. thank you very much. tom

Click to expand...

sab said:

Alrighty, fellers. Not completely understanding how the EVAP system works on my 05 LJ has always nagged at me, so I've spent the morning learnin' myself before attempting to give Tom some diagnostic advice. What follows is how I think it all works (based on reading the FSM and looking at my LJ), but before I start, if you've read my previous post, re-read that first because I made a major blunder when I wrote that. I was misremembering that the roll over valve was the check valve. Of course, it's not. The check valve is at the bottom of the filler tube and the rollover valve is next to it. Sorry for the confusion. I didn't mean to spread that contagion. I'd rather spread the cure - knowledge!

Ok, so here we go. Let's start with an annotated version of the tank picture that I posted above:
View attachment 703482

And here's a description of each hose and some of the components in the tank, and how it all works:

• Green Hose - The purpose of the green hose is to collect the fuel vapors that are stirred up while filling the tank. If you look at the filler neck, that green hose connects to a rubber hose (it's actually shown above - sitting on the tank between the green and blue hoses on the right side). The other end of that rubber hose connects to the small steel vent tube on the filler neck assembly. That vent tube terminates at the filler neck, just below the gate for the gas pump nozzle. When you fill up, the nozzle just fits in the filler hole, which mostly seals the system from the atmosphere. The vapors that are stirred up in the tank rise up the filler neck, but they are blocked at the nozzle gate and most of them enter that green hose (the rest sneak out between the nozzle and the ID of the hole in the filler neck for the nozzle.) The other end of the green hose terminates at the flow control valve (the red dot).
• Flow Control Valve - The purpose of the flow control valve (the red dot above) is very simple. It controls vapor flow from the tank to the charcoal canister. It has a float in it, and when the tank is full, it does not allow the vapors to pass into the red hose and on to the charcoal canister. This is done so that pressure in the tank starts to rise once the tank is full (and the float shuts the valve off), which helps shut the gas pump off. The flow control valve also controls tank venting when not refueling through the blue hose, which comes from the roll over valve (the blue dot).
• Roll Over Valve - The purpose of the roll over valve (the blue dot above) is two-fold. Primarily, it is how the tank vents when not refueling. It allows vapors to pass from the tank to the flow control valve through the blue hose and on to the charcoal canister through the red hose. The secondary purpose is how the valve got its name. It prevents fuel from spilling out of the tank's vent system. When the tank hits about a 45° lateral angle, the valve closes to prevent liquid fuel from getting into the vapor recovery system. I don't know how the valve actually works.
• Blue Hose - The purpose of the blue hose is to allow fuel vapors to exit the tank through the roll over valve, blue hose, the flow control valve (red dot), and the red hose into the vapor recovery system. Once at the vapor recovery system at the other end of the red hose, it passes through the charcoal canister, NVLD pump (through a pressure-controlled diaphragm valve), and vent filter out to the atmosphere.
• Red Hose - The red hose serves two purposes. Firstly, it allows the tank to vent into the vapor recovery system, both during refueling (through the green hose), and at all other times (through the blue hose). Secondly, it allows the vapors held in the charcoal canister to be burned by the engine through the yellow hose. This is done only when the engine is running in closed loop mode, and the purge solenoid (in front of the brake master cylinder in the engine compartment) controls when that happens.
• Yellow Hose - The purpose of the yellow hose is to allow the vapors held in the charcoal canister to be burned by the engine instead of entering the atmosphere. When the PCM detects that it's in closed loop mode, it will rapidly open and close the purge solenoid, varying the pulse width (the amount of time it's open each cycle) to control how much extra fuel is added based on running parameters. The oxygen sensors then help the PCM adjust the injector pulse widths accordingly (pulling fuel out to make up for the added fuel from the purge solenoid), to keep the air/fuel mixture correct.
• Check Valve - The check valve is on the end of the filler neck, inside the tank (not shown above, but you can see it in my previous post). It's purpose is to only allow the filler tube to be open to the atmosphere while fuel is flowing into the tank during refueling. It's also the cause of fuel splashing at the end of refueling (when the tank is full) and short-filling (pump shutting off early and making refueling very frustrating). This valve is normally closed by a spring with a very light spring rate. The fuel splashing problem is caused by it sticking open (or partially open), and the short-filling problem is caused by it sticking closed (or mostly closed). The usual reason for it sticking (both open and closed) is swelling of the piston from long exposure to gasoline. For those not familiar with the fix for that, please see the Wrangler TJ Gas Overflow Fix thread.

Simple system, huh? It's actually an interesting bit of engineering! I had bits and pieces of this in my noggin' before today, but now I've been able to put it all together and make sense of it. I hope. If anyone knows this system inside and out, please point out where I've gone wrong.

Now, onto Tom's current query:


As I said earlier, I doubt that a hose clog would cause this. The reason for my doubt is that I can't think of what would clog one of those hoses. Generally, fuel doesn't contain debris in large enough sizes or quantity to clog a hose. I'm not saying it's not possible; rather, that I wouldn't start there with a diagnosis. As I said in my last post, I also don't recommend using high-pressure air to "test" the hoses or unclog them. They're not made for high pressures.

Unfortunately, the best way to perform further diagnosis is going to require dropping the tank. Here's my suggestion for diagnostic procedures, in the order that I'd do them:

1. Re-check the check valve on the end of the fuel filler neck. It's the most likely cause of a short-filling problem because it's the first major obstacle for fuel entering the tank. Before pulling it off the tank, check to see what position it's in by carefully feeling for the orange tip on the piston that protrudes from the bottom of the housing:
View attachment 703530

You shouldn't be able to push it in, and you should be able to pinch the end and pull it down (open). If it moves closed when you push it, it was sticking open. If you can't pull it open, it's stuck. If you can pull it open, but it takes some force, it's sticky. If all is well (it's closed, and you can easily open it), it's likely not the problem, and you don't need to remove it and perform additional sanding.

2. If it's not the rollover valve, then it's the vent system. If the tank can't vent, then the pressure in the tank will increase while refueling, and once the pressure in the tank is high enough to overcome the inertial force of the fuel falling on the check valve from the gas pump nozzle, the check valve will close, and the gas pump will shut off prematurely. Unfortunately, testing the venting system is a little tricky because it's not easy to remove the plastic hoses from the barbs. I can think of two ways to do it without having to remove the hoses: 1) smoke test, or 2) very low pressure test.

Smoke Test
First, make sure your smoke tester is acceptable to use on a fuel system. If it has a system that could ignite the fuel vapors, well, boom! If it's safe to use it on the fuel system, you'll have to do two separate tests with the tank less than full (so that the float in the flow control isn't closing the valve). Both tests require watching for smoke out of the red hose/yellow hose/both above because that's where the tank vents. Smoke coming out of those hoses is a positive test (your tank is venting properly).

For the first test, plug the filler neck and connect to smoke machine output to the green hose. This will test if the flow control valve is allowing vapors from the filler neck to exit. If this test doesn't result in smoke coming out from the red/yellow hoses, then either the green or red hose is clogged, or the flow control valve is stuck closed. To check for clogs in the hoses isn't easy because they are installed using heat, and uninstalling them is tricky. Having never had to do that, I don't have any advice. However, the best way to check for clogs is to remove them and blow through them. Unfortunately, the flow control valve is integral to the tank, so if the flow control valve is stuck closed, you unfortunately need a new tank.

For the second test, plug the green hose and connect the smoke machine output to the filler neck. This will test if the flow control valve is allowing vapors from the tank to exit. If this test doesn't result in smoke coming out from the red/yellow hoses, then either the blue or red hose is clogged, or the flow control valve is stuck closed. Again, I have no advice on how to remove the hoses to check for clogs, and if the flow control valve is stuck closed, you need a new tank.

Very Low Pressure Test
This test is exactly like the smoke test, except that your observation is auditory, not visual. You need to listen for air exiting the red or yellow hoses. The amount of pressure to apply is about 1 psi. You'll need a regulator that allows you to get pressure that low. Follow the two smoke test instructions above, but substitute the smoke machine output for the 1 psi pressure source, and substitute looking for smoke for listening for air escaping from the red/yellow hoses. Everything else about these tests is the same.

Finally, if you are unable/unwilling to do either the smoke or very low pressure tests, I would suggest first checking your check valve as noted above, and if it's working fine, then look for a used tank to replace yours. To my knowledge, no one makes tanks for these anymore, other than Genright ($$$$).

Good Luck!

Click to expand...

sab said:

I'm happy to help because it expands my knowledge of my LJ. Most of my vehicles are "long in the tooth," and I've have EVAP system problems with all of them. I've found EVAP system problems to be very challenging because there's not a lot of detailed information out there about how they work. The FSM for the TJ is actually better than any other I've seen in explaining it, but it reads like a riddle to me sometimes. Now armed with this new knowledge, I'll be ready when my LJ's EVAP system starts failing...


No, vapor lock is a phenomenon that happens in the pressurized liquid fuel lines where excessive heat causes the fuel to change phase into gas. When that happens, it is now compressible, and if that gaseous "bubble" gets to a high point in the fuel line (where it's working against gravity to push uphill), and the pressure in the line isn't enough to compress it and push it over that high point, the fuel pump doesn't have enough push to keep fuel moving.

In your case, we're already dealing with gases, not liquids, so there's no liquid to become vapor locked.


If that is the case, pay attention to changes in behavior with the system. If the culprit is a sticky check valve, the severity of the short-filling problem could change, and you could see splash-back on filling. The reason for that is that the check valve may occasionally move position, and depending on the position it sticks in, the behavior changes. If you do see change in behavior, that would be a strong indicator that the check valve is the culprit.


Yes, please come back and report because it adds to the content here that's helpful for all TJ owners!

Good Luck!

Click to expand...

freedom_in_4low said:

you didn't by chance do the GM fuel hose fix at some point before you did the correct check valve mod, did you?

The reason I ask - in 2021 I did the GM hose fix to correct splashback. It improved, but didn't completely go away.

In 2024, I did the check valve mod. Per advice I found here, I gave the check valve flapper in the GM hose a good whack to break it out, dumped the piece in the floor, and returned the now empty hose back into service. Weeks later I got a CEL for an evap leak, which came and went (usually showing up in winter and going away in summer) and also very occasionally still got weird results at the filling station like early shutoff and splashback.

Finally, by this summer the CEL and evap leak code had become consistent even in warm weather so I decided to do something about it. A smoke test revealed a pinhole leak in the hose at the former pivot point of the check valve flapper, and the filler hose was mildly kinked.

View attachment 708995
(my cheap amazon smoke tester machine has a specific low pressure setting for evap systems but I'm embarassed to admit I didn't even consider the possibility that it could ignite fuel vapors, so either it's good or I got lucky).

I'm not sure I fully understand the mechanism enough to positively attribute the kinked hose with the odd behavior when filling up, but I'm happy to report that the leak code is gone and fillups have been uneventful since replacing the hose with a proper TJ filler hose.

Click to expand...