EVAP System Operation Q

Status
Not open for further replies.
Joined
Feb 6, 2010
Messages
4,836
Location
Central Texas
Here's a previous thread on my efforts to extinguish the P0455 Large EVAP Leak code: http://www.bobistheoilguy.com/forums/ubbthreads.php/topics/2755714/2

I'm currently revisiting this as inspection is due. P0455 is once again illuminated + a few others related to the SAS system/blower. I was able to eliminate the later using the "diode mod". Now I can focus solely on P0455. I suspect either the canistor shut-off valve or a cracked hose on top of the gas tank. As the sled is now ~ 17yrs old and this valve has never been replaced, I'm going to do so. Pierburg makes a replacement that is about half the cost of the OEM part.

I found this detailed EVAP System description very useful:

Quote:
Onboard Refueling Vapor Recovery (ORVR) Components
-1. Canister purge (CP) valve
-2. EVAP canister
-3. Shut-off valve
-4. Air cleaner (ACL)
-5. Fuel tank pressure sensor
-6. ORVR valve
-7. Roll-over valve
-8. Control module Motronic 4.4

EVAP canister
Hydrocarbon vapor pressed out of the fuel tank is dealt with by the EVAP canister. To minimize pressure drop in the lines between the fuel tank and EVAP canister, the EVAP canister is located close to the fuel tank. The EVAP canister volume is 2.1 liters. A new type of carbon reduces pressure drop in the gas flowing through the EVAP canister.

ORVR valve
There is a combined float roll-over valve by the connector for the large EVAP line to the fuel tank (called the ORVR valve). As its name implies it has two functions.

-a roll-over function which closes the large EVAP line connector so that fuel cannot leak out if the car is on its side or roof.
-a float valve function which closes the large EVAP line connector when the fuel level reaches the valve when the fuel tank is full

Roll-over valve
A small EVAP line runs from the ORVR valve to a roll-over valve located as high as possible on the fuel tank. The roll-over valve ensures that the fuel tank can breath if the ORVR valve is closed when the car has been refueled.

Pressure sensor
The pressure sensor provides information about pressure in the fuel tank. The sensor is connected to a Y pipe from the ORVR valve.
The pressure sensor also obtains the fuel tank pressure via the roll-over valve if the ORVR valve is closed.

EVAP canister shut-off valve
The valve is located close to the EVAP canister. The valve is part of the leak diagnostic system and temporarily closes the EVAP canister fresh air intake during the diagnostic test.

Filter
The filter is located by the fuel filler pipe and is connected to the EVAP canister shut-off valve via a hose.

Canister purge (CP) valve
This valve controls the flow of fuel vapor and air from the EVAP canister to the engine. The valve is controlled by the engine control module (ECM).

EVAP line
The line between the EVAP canister and fuel tank has been replaced by a large 13.5 mm line.
This allows fuel vapor to move easily between the fuel tank and EVAP canister when the car is refueled.

Fuel tank filler cap
The fuel tank filler cap has a overpressure/vacuum valve.

Check valve
The check valve at the bottom of the filler pipe on cars with ORVR is the same as that on AWD cars.

System description:

When refueling starts the check valve in the fuel tank opens. The fuel flow creates negative pressure at the top of the fuel filler pipe. This results in an air flow in the fuel filler pipe. It is important to have low resistance between the fuel tank and atmospheric pressure (via the EVAP canister shut-off valve filter) to stop fuel vapor escaping from the fuel filler pipe. The principle of the system is called a dynamic seal.

When the fuel tank is full the ORVR valve closes and the fuel tank pressure increases rapidly. This closes the check valve in the fuel tank, preventing fuel coming out of the filler pipe when the fuel pump pistol is shut off.

Fuel vapor in the fuel tank (together with air) reach the EVAP canister via the ORVR valve. Fuel vapor is absorbed by the carbon in the EVAP canister. If the ORVR valve is closed the fuel tank is vented via the roll-over valve.

When the canister purge (CP) valve is closed the fuel vapor remains in the EVAP canister.

When the engine control module (ECM) transmits an opening signal to the valve, a vacuum from the intake manifold or from the fresh air intake creates a flow through the EVAP canister. The EVAP canister is connected to the outside atmosphere via the canister shut-off filter which allows the flow to evacuate the EVAP canister.

The air flow through the EVAP canister removes stored fuel vapor which flow to the engine and enter the combustion process. The canister purge (CP) valve pulses rapidly when the car is being driven and is idling.


Diagnostics are carried out as follows:

The EVAP canister shut-off valve and canister purge (CP) valve are closed and the fuel tank system is now completely closed.

Pressure in the fuel tank should be stable. If pressure drops the canister purge (CP) valve is leaking.

The EVAP canister shut-off valve opens, the tank system is open. The canister purge (CP) valve is pulsed and because of the negative pressure in the intake manifold the engine starts to suck air through the EVAP canister. Because the EVAP canister shut-off valve is open the pressure in the fuel tank drops slowly. If the pressure in the fuel tank drops rapidly the EVAP canister shut-off valve is clogged.

The EVAP canister shut-off valve is closed, the canister purge (CP) valve continued to be pulsed, pressure in the fuel tank should drop rapidly. If pressure does not drop enough there is a large leakage in the fuel tank system. A diagnostic trouble code (DTC) will be stored.

The canister purge (CP) valve is closed, the EVAP canister shut-off valve is still closed, there is negative pressure in the fuel tank.

The negative pressure (A) in the fuel tank has stabilized, if pressure starts to increase too quickly, this indicates a small leakage in the system. A diagnostic trouble code (DTC) is stored.

The EVAP canister shut-off valve opens, the EVAP function is enabled and the diagnostic test is finished.

To carry out the diagnostic:

-engine coolant temperature (ECT) is between -10OC and +107OC.
-no permanent faults stored.
-engine idling.
-speed must be 0 km/h.
-the concentration of fuel vapor in the canister must not be too high.
-there must be more than 11 liters of fuel in the tank
-the pressure in the fuel tank must not be too high.
-fuel trim must be active.
-the car must be below 2700 meters above sea level.

The diagnostic test starts at the earliest 17 minutes after the engine has been started when all conditions have been met. It takes approximately 30 seconds.


Volvo%20ORVR%20Components_zpsbponrx7k.jpg


Questions:

1. Is the shut-off valve pulsed like the purge valve during driving?

2. If the shut off valve is stuck open or unable to seal, wouldn't this be an unmetered vacuum leak, and thus lower gas mileage?


Quote:
The EVAP canister shut-off valve opens, the tank system is open. The canister purge (CP) valve is pulsed and because of the negative pressure in the intake manifold the engine starts to suck air through the EVAP canister. Because the EVAP canister shut-off valve is open the pressure in the fuel tank drops slowly. If the pressure in the fuel tank drops rapidly the EVAP canister shut-off valve is clogged.


3. Why would the fuel tank pressure drop slowly if the EVAP canister shut-off valve is open? I would think it would quickly equalize to atmospheric pressure?

4. How could the pressure in the fuel tank drop rapidly if the EVAP canister shut-off valve is clogged?

5. In a properly functioning system (ie no vacuum leaks), should air be drawn into the fuel tank when the gas cap is loosened (WHOOOSH) before filling up? No air is drawn into an empty tank when I do this. If this system is designed to be sealed, and engine vacuum is pulled constantly on the tank vent system when the engine is on, shouldn't there be a WHOOOSH when the cap is loosened when the engine is turned off?

OR

Is this tank vacuum released when the engine is turned off because the ECU opens the shut-off valve, thus air is drawn into the air filter, through the valve, through the carbon canistor and into the tank, immediately upon ignition turn-off?
 
The professional way to find EVAP leaks is with a smoke tester. The car is put on a lift and a purpose-built machine used to blow synthetic "smoke" into the gas fill pipe. Then a visual inspection is made to see where the "smoky" air leaks out.

---
1. From the description, the shutoff valve is closed only to execute certain parts of the self-test, and open at all other times. However it may close while the car is off. The description is not clear on that. (your question 5).

2. No. The purge valve is connected to the manifold. When the PCM opens the purge valve it accounts for the extra air flow to maintain proper mixture.

The shutoff valve opening (and it will be open most of the time the car is on) only lets air in or out of the canister and the tank.

3. The shutoff valve must have a smaller orifice than the purge valve, so when both are open the purging will dominate and pull a slight vacuum in the tank and canister-- even with the shutoff valve open. This is normal purge mode, and the first part of the self test. In the second part of the self-test, the shutoff valve closes and a larger vacuum should rapidly form in the tank (if there are no leaks). Then both valves are closed to see if this vacuum holds. If the pressure leaks up toward atmospheric, the PCM concludes there is a leak and throws the code.

4. Because (provided the rest of the system does not leak) there is no way for air to get in with the shutoff valve closed. So if the PCM sees the pressure dropping too far into vacuum during the first part of the self-test, (when the shutoff valve should be open and letting air in), it concludes that the shutoff valve is faulty.

5. The description doesn't clarify whether the shutoff valve is open or closed when the car is off.
 
Last edited:
^^^

To elaborate on what he said, I had an old Chevy Caprice that did just that on tank cap release. I got WHOOOOSH and could see & smell the fuel vapor. Being it was OBD1 I don't think the ECU cared enough to set a code because I never got one.

To me that means the shut-off valve failed. It was not letting metered air into the fuel tank when running to let the fumes be purged towards the intake. So when my half-full tank sat in the sun the pressure built up in the tank thankfully not to the point of blowing the seams. It could have also been the purge valve but again being OBD1 who knows.

You can't have the engine pulling vacuum on the tank without fresh air entering via the canister. The purge valve is likely metered via voltage and orifice to control the amount of fuel vapor burned. The shutoff valve is more than likely just open.

Like was said above you never want to actually pull a significant vacuum on the tank. Just enough to get the fuel vapor out or to run a leak test.
 
The gas cap has a pressure / vacuum release valve as a safety mechanism to prevent ever blowing the tank seams. It does allow safe levels of pressure or vacuum to exist in the tank but it is never supposed to open in normal operation. Releasing HC vapors through the cap would cause pollution and defeat the whole purpose of the EVAP system.
 
Re: Smoke tester. Agreed. I bought a new empty gallon paint can to rig one using an old large aquarium pump. First, however, I'm going to replace the canister shut-off valve. My indy ordered one for me today at significant cost savings to me and it'll be here tomorrow early afternoon. (I'll have to think of some nice way to say thanks.)

1. The document didn't say, and I don't know, if the ECU uses both the purge & shut-off valves to keep a certain pressure or vacuum on the tank by regularly reading the fuel tank pressure sensor. OR if the shut-off valve is ONLY used to Test the EVAP system and the rest of the time, it remains unenergized (OPEN). There is no chk valve on it so excess tank pressure could be vented (car sitting in hot Tx sun for hours)to atmosphere or air can flow into the charcoal canister, then out to the purge valve up front so vapors can be burned.

__I left it alone since I figured it was only energized to test the system and it WOULD NOT cause an unmetered vacuum leak when the PURGE valve is open/pulsed. I'm rethinking the later.....

2. I wasn't sure the PCM would account for this extra airflow as it enters the intake manifold downstream of the MAF. So far research has revealed this might be the case.

__I think if the shut-off valve (SOV) is open, it allows air to flow all the way to the purge valve(PV) up front. I did prove this with an earlier test. This is why a faulty gas cap & any rotten tank/canister/valve tubing will also trigger a P0455.

3. The SOV has much larger dia. tubing than the PV, to minimize pressure drop and flow resistance.

SOV:
9486016.jpg



PV:
8627005.JPG


4. OK (momentary lapse of reason.....)

5. I think the SOV is OPEN when the car is OFF, as this is its unenergized state.

One thing I learned is that I'll have to temp block off the charcoal canister air filter or pinch the line closed BEFORE pressure_testing/smoking the system. Also I'll have to remove the gas cap until I see smoke coming out, then reinstall it.

Much thanks for the great answers here MK378! I'll have to OPEN a bottle of champagne when I FINALLY extinguish the dreaded CEL that's been in my face for far too long. What a royal PITA!
 
Since I posted that I read your other thread, at least at that time you were able to build and keep pressure and/or vacuum in the tank, so perhaps there is actually no leak.

Have you tested the sensor like you were talking about there? The check chart suggests that a bad sensor could result in the "large leak" code.

There is also a mention in the description, but not shown on the schematic, of a passive system that acts when filling the tank. Vapor laden air that is displaced from above the gas in the tank as the level rises goes through the canister to get de-vapored then through the shutoff valve line and to an outlet near the top of the fill pipe (where you put the pump nozzle in). This would require a relatively large air flow and may explain the larger hoses on the shutoff valve. I think the restriction is built into the air filter that vents to the atmosphere (part 4) not the shutoff valve itself.
 
Last edited:
Yes, I could indeed "expand" the tank with the little airbrush compressor. Had to be careful not to over do it.

No, I haven't tested the pressure sensor. When using my indy's Volvo tester, I can see the tank pressure change while it does the diagnostic test. Problem is it quits before completing the test. I think this is due to it sensing the large leak, so it exits.

Yes, you are correct re: larger hoses on SOV. Not sure of any restriction on the air filter as I've never removed it.

I'll know tomorrow if the new SOV solves this problem. This has been a burr in my saddle for quite sometime. To the point of wanting to fix it myself and not give up!
 
Update:

Installed new SOV, CEL lit on way to indy's shop. P0455. Deleted it, then ran EVAP lkg diagnostics. It pulled the tank down to about -463Pa, paused, stopped the test and set P0455. I cleared it, retested, tank dropped to -780Pa this time, paused, stopped the test then set P0455 once again. Ran it a third time and it pulled -463Pa.

Ran a few errands on the way home. No CEL yet. Rather aggravating to find NO JOY on spending $70 on this SOV only to find it wasn't the problem. Of course, there might be more than one leak.

Time to rig my DIY smoker, block off the canister air filter, and fill the whole system with smoke from the PURGE valve up front. This is one frustrating &*%$#!
 
Update II:

After more research, I went out and removed the rear floor covers and hatch covers to have another look at whatever I could see. The fuel pump collar is the only one visible. I could detect a faint whif of gas. I noticed that the gummy gasket material where the fuel pump wiring enters has some cracks in it. I pried this up to discover the wiring goes straight down, then across to the right, then through more gummy stuff. The fuel return connector came off easily by hand, the fuel send not so easily.

I'm not sure what's under the other cover. All I see is a metal can or cover. I'm unable to pry it out of there due to its being wedged between the tank and underside of the metal deck.

I stuck an inspection mirror at a 45° angle next to the fuel pump hatch, pointed towards the left, rear tire and saw more tubing, evidently going to the fuel tank pressure sensor, ORVR & roll-over valves. There is no access hatch for this area. The tank currently has about 12ga. in it. Had to add more today for the evap leakage diagnostic test, even though I'll now have to drain it before lowering it.

Anyone know what the "L" hose is? Is this the tank vent line? I know what the "J" hose is.

Tomorrow, I smoke...heavily. Could be multiple leaks from the front to back lines running on the passenger side, near the tank, etc. What a PITA! No wonder the "pro's" do this with smoke. Otherwise, it's just a deep black hole. And the solution to mine appears to be far down the list. Not just the gas cap.....
 
Update III:

Built a simple smoker using a new metal, gallon paint can, incense, some brass fittings and an old aquarium pump. I used this to feed smoke into the vac line that connects to the PURGE valve located on the radiator fan shroud.

The first place I detected smoke was out of the air filter located on the fuel filler hose (pass side rear). This is normal as the SOV is OPEN when the ignition is OFF. I stuffed in a foam ear plug to block it, then let it go on pressurizing. Every now and then I could smell gas fumes outside of the car, but wasn't sure where they were coming from due to the wind.

Crawled under the car, no sign of smoke. Decided to remove the PURGE vac line where it connects to the charcoal canister. As I moved it back & forth, I heard it SQUEAK loudly. Didn't see any smoke though. I finally was able to remove the connector and smoke exited from this line as well as the canister.

This is a prime area for a leak + this connector is now 17 years old! I stuck a red vinyl cap inside to block it off and it still smoked. So I pushed it in even further, and it still smoked! So I cut it off and used some red silicone tubing to fit over the end of the existing tubing and over the connector on the canister, then secured with a couple of zip ties. Peering down inside the cut-off connector, I saw two O-rings and maybe a third. I could easily manipulate this while still attached, it SQUEAKED and it seemed loose, but I saw no smoke. Decided eliminating it as a leak source was best.

This line is actually metal with a plastic covering over it! I plugged it up with a gloved finger and waited. Didn't take long before I heard the metal lid on the paint can POP (It didn't pop OFF, just expanded enough to press out a kink in the lid). So no leaks on this line between my finger and the other end where it connects to the PURGE valve.

I hooked this back up to the canister and waited. No can POP. Smelled a whiff of gas.....discovered it was coming from the gas cap area. I removed it and found a small tear. I put some silicone grease on it and it still leaked. So I cut the torn seal off, added two layers of a plastic trash bag between the cap & nozzle and tightened it down. I could still smell gas! WTH? So I unscrewed it, moved to an area of clean plastic and tightened it down again. This time it held better, but I could still get a faint whiff of gas.

This was a cheap O'reillys cap I paid $10 for. Looks like I'll be buying an OEM replacement.

I was able to get the metal paint can to POP with pressure this time. Which leads me to believe it was the SOV valve, loose canister vac fitting + torn gas cap seal.

I now decided to pull a vac instead of inject smoke from the same point. At first, I used a hand-held vac pump: No joy. So out came the air brush compressor. I pulled a vac for about 3 min and began to smell gas fumes from the compressor. So I pulled the line off and the PURGE line was sucking air. This was a good sign. Unfortunately, I don't have a sensitive enough vac gauge to measure this. I rinsed & repeated with the same results.

So it's now holding a good vac and pressure with the canister air filter outlet plugged, trash-bag seal on a cheap gas cap, and the odd, rigid 90° angle PURGE connector replaced with large dia. red silicone hose, held in place with zip ties.

Tomorrow I'll pick up a new gas cap from a WorldPac warehouse, lube the seal with silicone grease and repeat the test with pressure & vac.

With the smoker, I plan on looking for any possible vac leaks in the engine compartment. Two possibilities are:

1. Large vac hose to brake booster is hard as a rock stiff. I've replaced the lower section of it as it was easier to get to.

2. Fuel pressure regulator is on the firewall side of the engine, right above the steering rack, on the passenger side. I see a small rubber elbow, and two other vac lines that look original. Time to replace these with silicone.
 
Update IV:

Turns out the local WorldPac warehouse doesn't carry the OEM Volvo gas cap. Instead, it's made by Gates. My indy convinced me to buy the OEM cap from him so I forked over $32. Even he admitted troubleshooting EVAP codes are quite difficult for him and time consuming, after the usual suspects are checked. It DOES have a superior seal and lanyard. I lubed the seal with silicone grease and tightened it down.

Retested using his tester and this time the tank vacuum dropped to -1200Pa, held then exited. I've never seen it drop that far. It PASSED! Amazed...I ran it again. This time it dropped to about -400Pa and FAILED! Go figure that....

On to other things needing my attention.....
 
Some leaks leak under vacuum but not under pressure (while others are vice versa). You could pull a vacuum on the tank then pour some sort of liquid on the seals to see if it leaks in. Ordinarily, water or alcohol are used for that purpose, but you may not want them in your fuel system so consider stuff like WD-40 or kerosene.

Use the pressure sensor to get accurate readings of tank pressure. If you see the pressure improve when you douse a suspected spot with liquid, that is because the liquid has temporarily slowed down a leak.
 
Originally Posted By: sleddriver
I lubed the seal with silicone grease and tightened it down.


I always thought getting silicone anywhere near the fuel/intake system was an almost guaranteed way to poison the O2 sensors, thus the need for specific "O2 sensor safe" sealants. How do you go with Silicone grease?
 
Re: Pressure vs. Vac...yep. Agreed. I've ruled out the tank for now, especially after getting a good look at the top of one in a Volvo junk yard. All the vac lines on top are hard, unlike the 850, so no worry about the old rubber line that was prone to rot.

Pressure sensor is on far left, top side of tank. Inaccessible as installed. Must drop tank. As noted above, I did see a deeper vacuum after the new gas cap. That's progress.
 
How do I go = Why did I go?

It is a very, very thin film. Gas will never contact the rim. Besides, I've never heard of this 'poisoning'. Not saying it's not true. I always lube dry rubber/seals though as they last longer.
 
I was suggesting you could use your scan tool or voltmeter in conjunction with that hopefully precise built-in sensor to monitor the exact vacuum in the tank while you pump on it with the airbrush pump connected to the purge line. This should be a lot more exact on small vacuums than a dial gauge.
 
Last edited:
Agreed. I'd planned on monitoring that pressure sensor with a DVM to do just that. It's a three-wire sensor, so I'm guessing power, ground and signal. Might either be current out or voltage out. Not sure.

Don't currently have a scan tool, but a good one, allowing real time measurements, sure would be handy, in both this situation and others. Do you have a recommendation?
 
Update V:

It's now been 8 days since I passed (finally) the evap test using a Volvo diagnostics tester. No CEL!! No P0455 Gross Evap Leak Code!! This is an encouraging sign! Time will tell however. Previous experience revealed it was 'cured' for a few months, then reappeared! So for now, I'm "cautiously optimistic"....
 
Status
Not open for further replies.
Back
Top