The tank in wagon is under the cargo floor. I am not sure there is access plate from the top, but pull the carpet and check.

 

Mine is '98 e320 wagon. what exactly needs to be removed in the back to get to the sending unit? it looks like more than just pulling out the soft panel under the 3rd row, but i don't want to dismantle more than necessary to replace the sending unit. note: i was told the fuel pressure sensor wasn't working and that was likely causing my p0446 code.

thanks
Scott

 

https://www.ilcats.ru/mercedes/
Here is the latest, free EPC.
Start with posting VIN and follow to your tank picture. That should answer main question.
OP never come back to report how he access the tank.

 

It should be on the floor in the middle. More or less lined up with the fuel filler neck. Fuel "TANK" pressure sensor. Prior to removal the whole assembly, check the connector and power supply pin. The sensor may not be getting power for some reason.

 

thanks mr boca, can you be more specific on confirming that the fuel tank pressure sensor is bad?

there are 2 connectors (white rounded and black rectangle) on the fuel tank sending unit.....is the power coming thru the white rounded one one with (i think) 3 pins? across which pins do i measure the voltage (or which pin and ground)? ignition in the 2nd position? voltage should read between 0 and 5v, but not exactly 0 or 5v?

if power is getting to the sender, how do i confirm that the tank pressure sensor in the sender isn't working?

 

I assume you found what you were looking for. Yes it is the 3 pin one. Not sure if you have the same color code but red/green is 5V and brown/grey is the electronic ground, and the grey/violet should be the signal back to the ECU. With the ignition ON you should have 4.7 to 5.3 Volts. The signal output is between 4.7 and 01.V depending on the pressure or vacuum. I do not think you can test the sensor operation itself without precise vacuum / pressure source that you can adjust in millibars and check the output level.

The only thing you can do is to check that the sensor gets the power and the signal wire is continuous all the way to ECU connector D pin 5.

I am assuming P0450 is the only code you are having for the EVAP system.

The other thing to consider is if your charcoal canister is plugged, this may lead to out-of-spec tank sensor fluctuation (more than 100 millivolts) which will give rise to P0450. This means that a replacement of the sensor assembly may not solve your problem with a plugged canister. Hopefully this is not the case.

 

For what it's worth fuel tank pressure sensor does go bad. I replaced the fuel level sender on the 1997 E320 (E36 Frankenstein) because of the P0450. I had 5 volt reference with the sensor unplugged, which led me to believe it wasn't a ECU power supply issue. The P0450 was a hard code too, came back right away after being cleared. Luckily a forum member had a spare level sender, so I plugged that one in, hard code went away.

Then I installed the "new" sender and was able to pass smog.

 

a little more detail: i have a p0446 code not p0450. i replaced the fuel cap, tested voltage at the purge valve and the charcoal canister vent valve, and bench tested both valves. i finally paid someone to diagnose. a smoke test showed no leaks, but they said the fuel tank pressure sensor wasn't working (i'm not sure how he determined this). i've ordered a new fuel level sending unit, but i'm hoping to confirm the diagnosis before i install the new part.

i will test voltage at the harness connection to the sender unit. how do i test continuity of the signal wire back to the ecu connector d pin 5?

 

P0443 - > purge valve
P0446 -> canister shut off valve
P0450 -> fuel tank pressure sensor

Related to electrical issues when they come up individually and immediately reported.

Have a look at the attached.

Do you have P0446 immediately, or after several drive cycles ?

I have no idea how the technician had determined the fuel tank sensor issue, while a solitary P0446 points to a shut off valve issue.

As depicted in the attached, if the fuel tank pressure sensor output is stuck at a plausible value, it would be very likely you have a large leak test failure, the dreaded P0455, and blinking low fuel indicator.

In any case you can test the continuity at the ECU connector side by unplugging the connector D, and checking continuity at the connector between pins 4 and 5. It should read high resistance. Then at the fuel sender end put a jumper wire between the brown (ground) wire and the grey (signal) wire to form a loop. Then you should read few ohms (low resistance.

Can you confirm again under what conditions P0446 comes up ? immediately or after few cycles of driving ?

 

I was worried about the tank pressure sensor diagnosis for a p0446, but the dash board and my (cheap) scanner say "evap circuit malfunction" so i was hopeful a failure of this sensor could induce that. also, i found this as one of the steps to diagnose: "How does a mechanic diagnose the P0446 code? Uses a scan tool to monitor tank pressure data and make sure it is reading properly".

after the p0446 code popped up last summer, i tried some things and reset the CEL. the light came back on after 2 relatively short drives, while sitting at a stop light (this sequence happened twice). after another reset, the light took a lot longer to come back on, but eventually did (while sitting in my driveway after a drive). the guy who did the recent diagnosis reset the code and the CEL is still off after only a little driving with many readiness codes not set and no pending codes.

 

If this is the case, probably the electrical connections are fine, but the tank pressure sensor is somewhat stuck with a plausible value, and the ECU blames something else, like the canister shut-off valve. Since you ordered the part, you might as well replace it, and hopefully it will fix your problem. I am sure the MB SDS or equivalent can read the change in value of the pressure, with the shut-off and purge valves are operated by the scanner. The parameter is Fuel tank pressure difference, and the ECU reads in millibars. When the major leak test is done (with shut-off valve closed and purge valve open), The tank pressure should show a difference of 4 millibars in 10 seconds. If not, you get the P0455, and you are not getting this code, which puzzles me.

The smoke test is useful in most cases, but I am not sure whether it will detect all leaks in dynamic (vacuum or positive pressure) conditions, or leaks where leak is "hidden" (fuel tank pressure/ sender assembly seal for example).

In any case, you checked the purge valve, and shut-off valve, and no hose / connection leaks, the tank pressure sensor is the one that is remaining, and replacing that will hopefully solve your problem.

 

update: i confirmed power to the fuel level sending unit (4.93v), replaced the part, filled the tank to about 1/2 full (gauge appears to be working correctly), and started driving to get the readiness monitors set. note: briefly i had pending p0450 and p0460 after starting the car with the sending unit not connected. those codes went away and haven't returned.

after some driving, the low fuel light started flashing with a pending p0455 large evap leak code. i had just bought gas and the cap wasn't fully tightened. i tightened and let the car idle for 20 minutes to set the evap readiness monitor (the light wasn't flashing). after 20 minutes, the low fuel light was flashing again (current p0455 code) and all but secondary air readiness were set. l loosened and re-tightened the sending unit (probably made it a little tighter than before), reset the DTC code and idled for 20 minutes again. the low fuel light started flashing after about 18 minutes with pending p0455. 3 of 9 readiness codes are set.

any thoughts on this....clearly something i created in the process of replacing the part, right? i can't imagine its related, but i haven't replace the plastic cap that covers the sending unit yet.

 

It appears that, you are one step further. If the ECU detects an electrical fault in Purge valve / shut-off valve / tank pressure valve, it will not pursue doing the tank pressure leak test. Now that you have a large leak test failure, you either have a genuine leak in the system (which includes all the hoses in the system, tank itself (not just the fuel cap), or the fuel tank pressure sensor is stuck, not registering the expected pressure change limit. As you changed the sensor, I would double check the hoses, and hose connections.

 

The fuel level sender has a gasket to seal it. If the gasket is broken, it will register as large vacuum leak.

 

Thanks for the responses.

Deplore, the sending unit came with a new gasket. the old one was softer and worn. i didn't lube the new gasket for fear of contaminating the fuel and it was very tight and squeaked a bit when turning it in place. is that normal or maybe i damaged the new gasket?

Mr Boca, Can you give some guidance on checking hoses and hose connections? Is this a visual inspection? Where are the points of inspection? Am I looking for visibly leaking fuel, fuel odor, or something else? Wouldn’t the smoke test have done this?

Scott

 

Look at the picture in post 10 which shows the hoses between components. Have you replaced the shut-off valve? Electrically it may be fine but mechanically it may not fully shut off, so the air may be leaking through it. Hoses to/ from the canister, to/ from the purge valve (especially connection to the intake manifold of the engine).

You should not get any fuel in the evap system, only fuel fumes.

Some leaks occur when vacuum is applied and present, and I do not believe smoke test will detect those.

Check for cracks on the components like the fittings of the valves, canister / purge valve enclosures. Unfortunately you cannot activate the valves manually to test the system without the use of a scanner.

 

Well, if the gasket was no good, you'll find out soon enough when you fill up the gas tank...gas will seep out.

 

I have less than 1/2 tank so no leaking at the gasket now. Note: without the plastic cap over the sending unit and the floor access panel to the tank still open, i don't smell any gas inside the car. any way to confirm that gasket is good without filling up? i'll visually inspect it again, but should i lube it to keep it from drying out?

the indy who told me the tank pressure sensor wasn't working said there is a small fuel leak around the fuel pump. that's not part of the evap system, but could it be a problem causing p0455? my pressure at the rail was good last time i checked in august.

i bench tested the vent and purge valves (in august), but i will replace the vent valve/hose with the used one i bought on ebay. i'll check the hoses and the charcoal canister again. it concerns me that i don't hear the ticking of the purge valve like i used to (i know its getting power, opens when voltage is applied, and is clean), but i can't imagine that is related to my code.

 

Fuel leak around the pump should not affect the large leak test which is for vapor.

Did you have the purge valve clicking after you had P0446?

The leak tests are not done if you have a fault with purge valve or shut-off valve or tank pressure sensor. I am not sure if the purging is activated with a fault code in the evap system

 

perhaps the p0446 was the tank pressure sensor and i had the large evap leak all along and only now i'm seeing that. i don't remember hearing the purge valve ticking in years....long before the p0446.

i started today with pending p0455 and no flashing low fuel. i swapped the old shut off valve for a newer one, tightened the fuel level sending unit (it was already pretty tight) and inspected the hoses around the charcoal canister (see photos). nothing looked obviously bad except for two of the pics of the charcoal canister. i idled the car for 18 minutes and got the flashing low fuel light and still pending p0455.

would you take a look at the pics and let me know if you see anything? i noted 2 suspicious nicks on the canister itself. the hoses look ok, but there is a little corrosion on the connections of the lines from the tank and to the purge valve (not at the canister).

Photos (you may need to cut/paste the link; i used "insert link" for the same link below):
https://drive.google.com/open?id=1-4i9V6jkgCXuzZ3IZ9xWL3VbX9r_JKwQ

https://drive.google.com/open?id=1-4i9V6jkgCXuzZ3IZ9xWL3VbX9r_JKwQ

 

Hard to judge from the pics, but a hole as small as 2 hundredths of an inch can generate a large leak. You may clean the suspected areas and try to "patch" and see if helps.

Also a clogged canister can raise a P0455, and this is usually caused by overfilling (especially by the pump attendants who are unaware that they damage your car by overfilling).

Make sure that all hose clamps are properly tightened and in good condition.

 

i spoke to the indy mechanic who diagnosed the fuel tank pressure sensor. he said he tested from the fuel tank side of the purge valve to the gas tank and vent valve. he was confident that there were no leaks in the system and was surprised by the p0455. Note: after his test, i replaced the fuel level sending unit and got the p0455. then i replaced the vent valve and i'm still getting the p0455.

he suggested testing the purge valve and making sure the fuel level sending unit is good/sealed (note: i don't see/smell any gas around the gasket, but i'm not sure what else to check regarding the part).

to test the purge valve, i plan to remove the tank side hose from the valve and feel for vacuum and listen for the solenoid engaging (i've read it sounds like a popping sound). does the engine need to be at operating temperature for the ECU to apply voltage to the vale (ie, how will i know if its a successful test)? any suggestions?

another note: as soon as i got the p0446 i replaced the gas cap, but bought a cheap one from autozone. i've ordered an oem cap.

 

update: i disconnected the tank side hose to the purge valve and started the car. no ticking, pulsing or popping sounds. i idled until operating temp....still no action from the valve. with the key in the on position (not running), i see 12.55v from 1 pin to ground and 5.56v on the other pin to ground. with the car running, i see a constant 14v and 7.6v to those pins versus ground. with the car running i see a constant 7.6v pin to pin. i removed the purge valve completely and applied 12v to the pins and heard the open/close click.

does this mean that the valve itself is working properly? is it possible the valve isn't getting electronically pulsed by the ECU (or whoever tells it to open periodically)? how do i verify this and could it cause a p0455?

recap: at this point i have replaced the 1) fuel level sending unit (after market part. this eliminated the original p0446 and now i have p0455, so i tried removing and re-tightening it), 2) the vent valve (used, but tested OEM) and 3) the gas cap (oem). the hoses look fine to me and a smoke test (indy mechanic) revealed no leaks between the south side of the purge valve and the tank/vent valve (note: the smoke test was done with the old sending unit in place which is why i suspected that my install of the new part created the p0455 code).

any suggestions for next steps?

 

The purge valve electrical issues (open, short-to-ground and short-to- power) generate a purge valve specific fault code (P0443). You can generate this by unplugging the purge valve connector while the engine is running. You can then reset the code after plugging it back.

The pulses for the purge valve are quite narrow, and it is possible you may not be able to hear it with the engine running, but you should feel it with your finger on it. Just because the solenoid valve is activated it does not mean it is properly opening or closing the vacuum line. The way it is functionally tested involves in connecting a vacuum gauge to the tank side and observing the oscillations in the vacuum pressure.

In any case, the code is P0455, meaning the purge valve has passed the electrical test at least, but this does not mean the valve is not leaking. Also the hose and connection between the purge valve and the engine intake may be leaking, but I am not absolutely sure whether it will cause a P0455, as this is isolated by the closed purge valve when the leak test is performed.

Another way to check the leak is to apply vacuum with a vacuum pump from the hose connected to the purge valve (to the tank) and make sure that the canister shut-off valve is closed (actuated by the scanner), and then observe the change in vacuum reading over time. This will not pinpoint where the leak is, but at least confirm that the EVAP system still has a leak somewhere.

Also note that if the canister is plugged, this will cause the tank pressure sensor to oscillate beyond an acceptable range, or possible purge valve leakage will cause a P0455, even though there is no leakage to the external world.