Vacuum Leak From Gas Tank?

[Audiophyle]

00 S4 Tip, with a small vac leak.
DTC code specifies a small EVAP leak, which my boost gauge confirms. Idle sits at -11" of vacuum.
I have literally tested EVERY vac line in the car, and there is only one thing I can find causing the leak, the gas tank.

This is a picture of where the charcoal canister sits, and the 3 lines that connect to it. (not my car)


#1 connects to the LDP, it is one of the 2 vac lines the LDP uses to test for leaks. This line is intact, LDP is intact, & line from LDP to vac canister up front is intact.

#2 connects to the gas filler neck, this is where the fuel vapor is introduced to the charcoal canister. With vac pump on this line the max I get is -11" vac, telling me the leak is there somewhere & identical to what the boost gauge reads. I get the same results with & without the gas cap on, so its not the cap. There has to be some kind of check valve or something, IDK. There is no rush of air when removing the cap when refueling either (due to no vac in tank)
I am stumped about this leak for the time being.

#3 connects to the N80 purge valve on top of the engine. Line holds full vac, and purge valve is working fine. No issues here.


My idea is to cap line #1 & #3, eliminating the vac leak & the car should think all is copacetic.
Leave line #2 connected to the canister so fill ups wont be an issue & the canister will still trap fuel vapors as intended.


Feedback? Possible issues? Ideas about the leak?
If anyone has a diagram or breakdown of the fuel tank & its components I'd love to see it.
TIA

[BlkdoutS4]

an evap leak will not cause your vacuum to be at 11mmhg, if that is where your vacuum is at idle then you have a vacuum leak and need to pressure test the car.

have you tried a new gas cap? a bad cap can throw a evap leak code. also check line one in your diagram.. sometime with aftermarket exhaust that line can be too close to the piping and the heat will cause a small pin hole to appear and cause a leak. this happened to me with my SSAC cat back..

[The_Jerbel]

So, I assume you've been doing boost leak testing with soapy water?

[bigern45]

EVAP leak doesnt mean vacuum leak. evap has to do with fuel tank (fuel system)and the ability to hold pressure and vacuum, and to change pressure and vacuum, and that type of vacuum at idle means you have an intake leak or timing belt jumped a tooth, but usually youll get a code for cam sensor position off or something.. so me thinks you have a vacuum leak .. but also youd think you would have a bank 2 or 1 lean code.. or both.. hrm... drankin.. cant think straight..

[Audiophyle]

DTC specifies vac leak in the evap system, boost gauge confirms there is a vac leak somewhere. That is the only code I am getting.
Generally a car will run rich with a vac leak, as the fuel pressure increases as vac decreases (& increases even more under boost). The car adjusts the injector duration to compensate for the higher FP so AFR is still about where it should be.

I have already physically tested & inspected EVERY vac line under the hood, one at a time. This is the only thing it could be. I have not been using boost leak tester + soapy water, I have been using a vac pump & systematically checking each & every possible location for leaks. Everything is intact except what I am finding at the fuel tank. & what I am finding there exactly matches what my boost gauge has been showing me. That cant be just coincidence.

[bigern45]

dude, its a separate system, you have a purge valve that controls pressure/vacuum in the tank. believe me.. and it may be called something different or have another component, but its separate...

[Audiophyle]

Do tell please, because I have 2 vac lines that connect directly to the engine.

[The_Jerbel]

Yes, you do, but that is a pretty small air source to cause the engine to be running 11 in-hg of vacuum. I would do a boost leak test. You may have a torn F-hose or something (I've seen this cause EVAP vac leak DTC's before). How does the car run, other than the vacuum at idle?

[sCeRaXn]

dude, its a separate system, you have a purge valve that controls pressure/vacuum in the tank. believe me.. and it may be called something different or have another component, but its separate...

Its not a seperate system. The n80 uses vacuum from the motor to draw vapors from the tank and reintroduce them into the combustion cycle. If this valve fails then there will either be no vac to the ldp/tank or there will be constant vac...either way it should throw a code for the n80 and a large vac leak(unless its stuck open). Basically everything that runs off of vacuum is related on this car. There is a continuous loop going from the n80 to the charcoal canister, filler neck, LDP, and then back to the vac source on the drivers side of the motor. If any of these arent completely sealed then youll get a code.

Audiophyle-I worked on a buddies car that had a similar problem a while back and it ended up being the small line that makes a 90deg turn out of the LDP and goes to the drivers side of the engine bay. The little connector thats next to the vc got brittle from so many heat cycles and cracked. I replacd it and everything went away. But it sounds like you already know of one leak at the filler neck. Like i said before if there are any leaks in the Evap vac loop then you will get codes. Replace that hose and see what happens. Also, be sure to clear all your codes after doing this. If you have the ability to log you can try blocks 70(n80 pulse rate) and 71(LDP test system reed switch). The n80 should stay around 15-20% when the car is warmed up and LDP reed switched should be closed when there is vac present.

You cant really cap the LDP and n80 lines or else youll get more codes.

[Audiophyle]

^^ Thank you!

I know exactly what line you are talking about & its completely intact. The line from the LDP is a metal core with a plastic cover, it connects to the vac reservoir in the front of the car, and then that crappy plastic line runs from there to the mess of lines just above the FPR.
I could see capping the 2 lines causing an issue with the LDP not having a vac source on the Evap side, maybe if I loop the two lines together? So the N80 line provides vac to the backside of the LDP. Temp fix I know, but I need to pass emissions soon.

With the car parked in my garage there is no gas smell, only when I remove the gas cap, & then its an instant gas smell, telling me there is not a hole in the tank. I found a thread a while back that a guy had an issue with the line from the charcoal canister. Its a shitty plastic line like the spider hose, but right at the gas fill neck there is an aluminum fitting & then the line changes to rubber. The problem he was having was the aluminum fitting failed & created a leak. I couldnt get in there to unhook the line from the filler neck, but I will have to try again this weekend. I need a friend like the guy in the BK commercials, his tiny hands would greatly come in handy. lol

[sCeRaXn]

I dont see why that wouldnt work. Of course these cars are finicky as hell so you never know. After i posted i went and dug up an old AW post that might help out. Theres no telling how long it wouldve taken me to diagnose my buddies car if it wasnt for this post. Its a long read but its worth it. I hope it helps out man.

http://forums.audiworld.com/showthread.php?t=1595413

btw- lol at little hands guy!

[Audiophyle]

Found that thread a while ago, practically have the whole thing memorized already. lol

[YBLEGL]

Does anyone near this guy own a smoke machine, if so lend it out to him, and find the EVAP leak please, if no OP goto a shop and have them test the evap system with a smoke machine and you will know exactly where your problem is. If you want to drive to MA I can lend you one to use and show you how to use it.

[bigern45]

Its not a seperate system. The n80 uses vacuum from the motor to draw vapors from the tank and reintroduce them into the combustion cycle. If this valve fails then there will either be no vac to the ldp/tank or there will be constant vac...either way it should throw a code for the n80 and a large vac leak(unless its stuck open). Basically everything that runs off of vacuum is related on this car. There is a continuous loop going from the n80 to the charcoal canister, filler neck, LDP, and then back to the vac source on the drivers side of the motor. If any of these arent completely sealed then youll get a code.

Audiophyle-I worked on a buddies car that had a similar problem a while back and it ended up being the small line that makes a 90deg turn out of the LDP and goes to the drivers side of the engine bay. The little connector thats next to the vc got brittle from so many heat cycles and cracked. I replacd it and everything went away. But it sounds like you already know of one leak at the filler neck. Like i said before if there are any leaks in the Evap vac loop then you will get codes. Replace that hose and see what happens. Also, be sure to clear all your codes after doing this. If you have the ability to log you can try blocks 70(n80 pulse rate) and 71(LDP test system reed switch). The n80 should stay around 15-20% when the car is warmed up and LDP reed switched should be closed when there is vac present.

You cant really cap the LDP and n80 lines or else youll get more codes.

but my point is everyone and their mom lives with a evap code, but will this cause his drivability issue?? that low vacuum?? that was my point.. everything is connected in some way shape or form. so it is the n80 valve.. okay.. but still 11 inches of mercury is a pretty unhealthy engine. mine was at 15 inches of mercury when i jumped 2 teeth on my timing belt..

[Audiophyle]

The car is running great, there is no way the timing could be off & the car run the way it does. Not to mention there would be far more codes being thrown than just an evap leak code. Before the leak the car pulled -18" at idle consistently, after leak DTC its pulling -11". Its not an unhealthy motor, it is just a vac leak. Have you never had a vac leak on a car before?

[bigern45]

lol.. hold on big fella.. no need to be a smart ass. i wasnt saying that your car had jumped teeth, i was just saying that THAT IS AN ENORMOUS leak. i dont see a n80 valve being bad or evap leak causing that much vacuum loss. and yes.. it is running unhealthy.. so right now it is unhealthy. even if it runs good on the road. not saying timing belt jumped, or that you have a bent valve . me saying that my car was running 15 inches of mercury when my belt jumped was just me saying that my car was really unhealthy, and it had better vacuum at idle than your car does now.

that being said. i hope you find your problem. and i hope its something that ive never heard of before... so then i learn, and again win for using this forum.

.. i ran a 12.9 @ 112 with a jumped belt, so i couldnt tell my car was off two teeth till i removed the cover and checked..

[Audiophyle]

Not being a smart-ass, just saying a vac leak doesnt make an engine unhealthy.
Timing off will def make an engine unhealthy, as your lack of vac had to do with the misaligned valve operation not a leaky hose somewhere, there is a huge difference. & in a situation like that, yeah -11" at idle would be serious bad news. Do you not have a VAG cable? Did it not throw any error codes?
That is a major malfunction, and should have thrown warning lights at you. (how many teeth does it take before a valve contacts a piston? Or possibly another valve?)

A vac leak from an external source (ie hoses, evap system, etc) is not as big of an issue, and it doesnt take much of a crack to completely loose all vacuum. Tiny leaks (such as this one) are practically inaudible when pressure/leak testing lines, which makes hunting them down that much harder.

& a 12.9 is pretty quick, what did ya run after fixing the timing issue?
My Honda ran a 13.18 @ 110, but it was quite healthy.

& we are all here to learn, so we can agree on that atleast. lol

[bigern45]

so we disagree on what is a healthy engine as well. i guess thats what being a tech for 15 years does for you.

.. you can be an extremely healthy person. you catch a bad virus and you get the shits. even though its temporary, for the time being , it is very unhealthy, would you agree?? your car is compensating for the vacuum leak, is it not?? so your short term fuel trim levels are probably at a very high percentage during idle and before boost. then as you go into boost, your fuel trims probably drop and go negative. drive this car long enough and im sure the long term fuel trims will be out of wack as well, depending on how easy/hard you drive your car.

[Blue_S4]

You said it's a tip. If you still have your secondary air injection pump hooked up a failed pump can cause a leak like that

[The_Jerbel]

If he would boost leak test the car he might just find a few surprises, but apparently he's not interested in doing that...

[AddICTioN]

Audiophyle- did you figure this out? I'm having a similar issue.

[Audiophyle]

Havent had a chance to touch anything since my last posting, busy with finals. Plan to have the issue found & resolved by the end of the week though.

Jerbel, you said that before & I had planned to do so but in my searching for methods of building my own tester I found a few threads saying they work great for finding leaks in the charge piping, but not for leaks after the TB (where actual vac leaks occur). After reading your last post I was going to share the threads I found, but cant find them anymore. Instead I found more backing up your claim, so I will be building a boost leak tester this week. I am pretty sure I have narrowed down the leak, but the tester will be handy to have anyway.

[The_Jerbel]

Sounds good. A boost leak tester will test everything from the Y-pipe all the way to the IM, and even the cylinder chambers. I've discovered more than a few leaky fuel injector seals using a boost leak tester...

[Audiophyle]

Well today I looped the 2 lines together, effectively eliminating the charcoal canister & gas tank from the vac system. (or so it should have anyway). Leak is still present, and hadn't changed at all. I am back to being officially stumped.

Jerbel, now I remember why the boost leak tester wouldnt work. All of the vac system is protected from boost via check valves, so no pressure reaches outside of the IM. Boost Leak Tester (BLT) connected and pressurized I didnt hear any immediate leaks anywhere (no hissing), but might try soapy water tomorrow in the daylight. One thing I DID notice was I could clearly hear bubbling/gurgling in the oil from the removed oil cap. Is that normal??

[bigern45]

..man.. it would drive me nuts to not have my car running good for this long.. you have much more paitence than me..

hope you find it soon!!

[mholme]

The bubbling is normal.

[Audiophyle]

The bubbling is normal.

HUGE sigh of relief, thank you.

[The_Jerbel]

Well today I looped the 2 lines together, effectively eliminating the charcoal canister & gas tank from the vac system. (or so it should have anyway). Leak is still present, and hadn't changed at all. I am back to being officially stumped.

Jerbel, now I remember why the boost leak tester wouldnt work. All of the vac system is protected from boost via check valves, so no pressure reaches outside of the IM. Boost Leak Tester (BLT) connected and pressurized I didnt hear any immediate leaks anywhere (no hissing), but might try soapy water tomorrow in the daylight. One thing I DID notice was I could clearly hear bubbling/gurgling in the oil from the removed oil cap. Is that normal??

Yeah, the bubbling is normal. It's common for the check valves to wear out, which can cause vacuum leaks. Try the boost leak test with soapy water, and you might consider replacing the check valves.

[Audiophyle]

I had thought about that, and physically tested every check valve I could find under the hood. All hold -25" vac flawlessly.
I have also used about a million hose clamps on this car already, damn Audi & they're extensive love for vac lines. lol

I am removing the SAI system right now, which held vac fine when tested, but was throwing DTCs for inefficiency. Removed all of the vac lines associated & capped the connection to the main vac line near the DV solenoid (I forget the part number). Fingers crossed that does it, but not gonna hold my breath.

I am also going to remove the Spider hose assembly, which I have already completely replaced anyway, and install a catch can system. With the pancake valve unhooked from the Y pipe I noticed vac dropped to normal if I plugged the valve, but the car would start smoking. (sucking oil into IM, the purpose for the pancake valve) I think if I remove the valve, and cap the vac line to the PCV & run a Catchcan without a vac source, it should fix the vac leak DTC. I know the PCV is supposed to be under vac, but anytime there moderate throttle applied the PCV system reaches ATM pressure anyway, so no vac at idle (ie minimum blow-by) should'nt be an issue. Then I can also upgrade the hose diameters & increase flow capability above the OEM system, so under heavy load (boost) the PCV system will allow greater ventilation over OEM, which WILL be beneficial.

& Jerbel, I know we have butted heads in other threads, & I gotta say I appreciate your help here.

[The_Jerbel]

No worries, man. I'm happy to help.

You've obviously been thinking about this a LOT, but I still have to insist- a boost leak test will really help you to diagnose the problem. If your pancake valve is worn out, a boost leak test will show it. If you've got any leaks, a boost leak test will show it. If you've got a leaking N249 valve (the DV solenoid you're talking about), the test will show it.

BTW, your catch can setup does not require a vac source. You're right about that.

[Audiophyle]

If your pancake valve is worn out, a boost leak test will show it. If you've got a leaking N249 valve (the DV solenoid you're talking about), the test will show it.

How? By design the valve is supposed to allow unrestricted flow outward from the crank case.
Although I do remember some air coming out of it while testing previously, which could mean that check valve isnt working right.
(but I am pulling it all out for the catch can anyway)

Every vac line exiting the IM is protected by a check valve, which prevents any positive pressure from entering any of the vac system & all of the control solenoids. Because of the check valves, the entire vac system "should" stay in vac even when the car is off (Pull your brake booster line sometime as an example). The fact that the fuel tank doesnt hiss after removing the cap on an empty tank & the line I am talking about above doesnt hold vac but should, screams the leak is outside of the IM (behind the protection of the check valves. I dont see a boost leak tester affecting anything outside of the IM & charge piping.

With a bad check valve there should still be vac at idle, but anytime the throttle is applied the LDP would detect the vac dropping inside the check-valved system when it shouldnt be dropping. Because the leak is constantly there, & behind the check valves, there has to be a physical leak in the system somewhere.

I plan to attack the car with soapy water & the tester anyway, but not very hopeful that will find the problem.

[The_Jerbel]

Well, not exactly. EDIT: The pancake valve is designed to allow unrestricted airflow toward the crankcase from the Y-pipe, and a restricted amount of air flow into the Y-pipe from the crank. Just grab a brand new spider hose sometime, and see for yourself. You can blow a limited amount of air out the pancake. A failed valve will allow unrestricted air flow from the crank into the Y-pipe.

I understand what you're saying about the vacuum system containing one-way check valves after the IM, but I can definitely tell you that reality trumps theory every time. The theory is nice to cuss and discuss, but I know from personal experience that the entire vacuum system is pressurized when doing a boost-leak test. Try it out. I'm only suggesting that you do so for your benefit. I don't make any royalties off of people using a boost-leak tester, or anything I would also pressurize the fuel tank to see if you can isolate any leaks in that area.

[Audiophyle]

Such a liar, we all know you make royalties of people using boost leak testers.
& I hear ya on the "theory" of check valves.

On a side note, today I clear cornered my oem headlights & got started on my catch can.




Just about finished with it, and will have more pics tomorrow.
I am curious about the line from the N75? (boost control solenoid) that attaches to the spider hose. According to an S4 service manual I found online, that line is just a reference & uses ATM pressure. So deleting the spider hose setup I can just leave that hose open to the air, right?

[The_Jerbel]

Yup.

BTW, I just realized that I wrote something in my previous post that made no sense whatsoever. I changed it, so it makes more sense now.

[Audiophyle]

You are backward about the pancake valve. It allows unrestricted flow FROM the crankcase TO the Y pipe, so under boost all crank pressure can easily be vented into the Ypipe & not blow out crank seals. The pancake valve restricts air flow FROM the Y pipe to limit the amount of vacuum inside the crankcase to prevent oil from being sucked into the IM.

[The_Jerbel]

Correct you are, I did have it backwards.

[a4turbo_2002]

Its not a seperate system. The n80 uses vacuum from the motor to draw vapors from the tank and reintroduce them into the combustion cycle. If this valve fails then there will either be no vac to the ldp/tank or there will be constant vac...either way it should throw a code for the n80 and a large vac leak(unless its stuck open). Basically everything that runs off of vacuum is related on this car. There is a continuous loop going from the n80 to the charcoal canister, filler neck, LDP, and then back to the vac source on the drivers side of the motor. If any of these arent completely sealed then youll get a code.

Audiophyle-I worked on a buddies car that had a similar problem a while back and it ended up being the small line that makes a 90deg turn out of the LDP and goes to the drivers side of the engine bay. The little connector thats next to the vc got brittle from so many heat cycles and cracked. I replacd it and everything went away. But it sounds like you already know of one leak at the filler neck. Like i said before if there are any leaks in the Evap vac loop then you will get codes. Replace that hose and see what happens. Also, be sure to clear all your codes after doing this. If you have the ability to log you can try blocks 70(n80 pulse rate) and 71(LDP test system reed switch). The n80 should stay around 15-20% when the car is warmed up and LDP reed switched should be closed when there is vac present.

You cant really cap the LDP and n80 lines or else youll get more codes.

Ding ding ding ABSOLUTELY CORRECT! Do you guys seriously think the evap just gets its vacuum fromm magic? Lol