[jsunma]

Recently, I discovered that my left (driver's) side fuel level sensor had developed the common cracking in the plastic cap which produces fuel odors outside the left side of the car. Actually replacing the part - which seems to be a common task - wasn't difficult and the new part seems to have solved the problem of fuel odor. Unfortunately, it's caused another problem...inaccurate fuel level readings.

First, the operation of fuel level reading in this car is a bit unclear to me. It appears that there's a float in BOTH sides of the tank, and so therefore if there's a problem with the float, it could be on either side? Recall that you need to pull out the right side mechanism (which is the fuel pump itself plus what apparently is a float mechanism) in order to change the left side sender. So, if I now have a float problem, could it be on either side?

I checked both floats to make sure that nothing was hanging them up...but it might be necessary for me to open everything up again and check again. But before I do, I wanted to see if anyone else had encountered this on a regular 328i (N52). Meaning, it has just the one fuel pump and hasn't been modified as I understand some of the N54 people do (none of that seems to apply to me).

I had the battery disconnected during this work. I've checked and there are no codes related to the fuel system. No SES light (meaning that the DTME pump, etc. is working correctly).

Now, when I fill up the tank, when I go to start the car, the fuel gauge shoots up to full, but THEN starts to drop within just a few (like 2) miles of driving. It drops to about 7/8 of full and from that point SEEMS to work normally except that there's some amount of "hunting" on the gauge...meaning that it sort of goes up and down as I drive. It does generally drop over time and goes pretty much to full empty...although since I can tell it's no longer accurate I am keeping track of miles and not trusting the gauge. Which is a pain, since my gas gauge was totally accurate for 230K miles until I messed with it by having to open things up.

Any tips from anyone who's had this problem so that I can be most efficient in going in there to solve this problem?

[gbalthrop]

Originally Posted by jsunma

... First, the operation of fuel level reading in this car is a bit unclear to me. It appears that there's a float in BOTH sides of the tank, and so therefore if there's a problem with the float, it could be on either side?... I checked both floats to make sure that nothing was hanging them up...but it might be necessary for me to open everything up again and check again. But before I do, I wanted to see if anyone else had encountered this on a regular 328i (N52)... I've checked and there are no codes related to the fuel system. No SES light (meaning that the DTME pump, etc. is working correctly).
Now, when I fill up the tank, when I go to start the car, the fuel gauge shoots up to full, but THEN starts to drop within just a few (like 2) miles of driving. It drops to about 7/8 of full and from that point SEEMS to work normally except that there's some amount of "hunting" on the gauge...meaning that it sort of goes up and down as I drive...

Please see NEXT Post for attachments, so as NOT to ruin margins here.
I've never opened my tank, but here is what Bentley describes, and what INPA shows happens on my 3/2007 build 328xi N52K:
1) The Fuel Tank is "Saddle-shaped", with a Left Tank and Right Tank, built around drive shaft for lower Center-of-gravity.

2) Bentley diagram of NON-Turbo fuel tank is attached. Turbo tank is similar. There is a Fuel Level Sender with float arm on each side of tank. Fuel Pump is located on Vehicle Right tank and Suction Jet Pump pumps fuel from Left Tank to Right where Pump Pickup for delivery to fuel rail is located. The Fuel Level Gauge on Instrument Cluster displays TOTAL (buffered/muted) level of BOTH tank Senders.

3) When Tank is full (BOTH sides), both level senders should provide signal to JBE Module (to which they are wired directly) of approximately equal fuel level. When TOTAL fuel level per gauge approaches 1/4 Full, there should be essentially NO fuel in Left Tank (due to Suction Jet Pump operation), and Right Tank ~ 1/2 full. JBE Module sends signals to KOMBI (Instrument Cluster) Module via K-CAN Bus, so you can test BOTH signals as received by JBE, and by KOMBI.

4) INPA & ISTA both show signals received from the TWO tank Senders, both when connected to JBE Module (hardwired signals from senders), and when connected to the KOMBI Module (K-CAN Bus from JBE).

5) ScreenPrints from BOTH INPA & ISTA for my 328xi are attached, showing values when TOTAL tank Level was ~ 1/4 per gauge, and all/ nearly all fuel was in Right Tank due to Jet Pump operation. INPA Menu Path for each screen is:
INPA > JBE > F5 Status > F3 Status Sensors
INPA > KOMBI > F5 Status > F1 Analog 1
ISTA Live Data can be viewed when connected to each module and selecting "Diagnosis Scan" Tab. To connect to a module, select Module in "Control Unit Tree" or "Control Unit List", and then click "Call up ECU functions" button. Click Diagnosis Scan Tab, and select ECU function you want to view INPUTS or Live Data for. Ignition ON, Engine On/Off.

If you have a Windows OS Laptop, you can download INPA (BMW Standard Tools), & ISTA from MEGA sites linked in attached pdf: 1 E9x References.

Knowing HOW the system works, and having INPA, ISTA, or equivalent Live Data readout with capable Scan Tool, allows diagnosis without opening tank. Please let us know what you find,
George

[gbalthrop]

Attached are ScreenPrints referenced in Prior post, INPA & ISTA, showing tank Sensor signals as received by (1) JBE Module to
which tank sensors are hard-wired, and (2) KOMBI which receives those signals from JBE via K-CAN Bus.
Also attached is Bentley Fuel System Schematic.
George

[jsunma]

Thanks George! I have ISTA and INPA and thus can follow your logic and test this out.

I'll repost with what i find.

[jsunma]

Well, after filling the tank completely, I find from ISTA that the left side reads 599 ohms and the right side 996 ohms. So I'm assuming that the left side is the one that's hung up somehow, as it should be roughly equivalent to the right side...but isn't.

So when I get a chance I'll dig into that and remove the left side regulator, free up the float again, and see what I see.

Thanks again George!

(My copy of INPA is screwed up, so I couldn't see the better reading that converts the resistance into a measure of liters...but I think I get the point)

[KClemente]

Your level sender float is upside down. Look at how the rod is positioned on the sensor and compare to photos of the regulator on google. Chances are its out of alignment

[jsunma]

Originally Posted by KClemente

Your level sender float is upside down. Look at how the rod is positioned on the sensor and compare to photos of the regulator on google. Chances are its out of alignment

Thanks for that input! Using George's trouble-shooting procedure above, I think I've figured out that the problem is in the left side of the tank...that would be the float/mechanism on the new regulator I installed. But I've had to go out of town for a few days, so won't get back to opening it up and checking things out until next week.

When you say "upside down", can you be a bit more specific by posting a photo or a link to one of the images on google? I really don't see how it could be upside down. I was thinking maybe it was just hung up on a hose...although since the hoses all run below the float (once they're installed and pulled over to the right side), I'm not sure how that could be either.

[jsunma]

So, the saga continues.

Today I finally got around to opening up the left side of the tank, since that - based on my ISTA readings of the ohm values of each sender - was indicated as the side that was reading low when it should have been pretty close to the same as the right side (since I was completely full of gas). I had expected to see either the float binding on something or something installed "upside down" as a previous commenter had suggested.

Neither was the case. The sender float was positioned correctly. Nothing was binding or catching the float arm. I put the whole thing back together and went and filled up.

Same thing happened. For about the first half mile, the gauge read completely full and all looked good. Then it dropped about 1/16 (so it was one bar on the gauge below 100%) and that's where it is now. Attached is a screen print from INPA, which you can see shows the left side having less fuel in it than the right side...when both should be pretty much the same (since I had totally filled the tank). I have the same data from ISTA (no screen print), where the Right Side sensor is reading 995.9 ohms and the left only 526.6 ohms. I believe that would translate roughly to the 35 vs. 21 as shown in the attached INPA screen shot.

What gives? This is really frustrating.

[sabino56]

I just replaced the FPR on my E90 this weekend. I still have the old one and measured the resistance of float min to max = 51 ohms to 994 ohms. Only confirms what you already know but I was curious.

Did you measure the resistance on the pins on top of the FPR directly or just via INPA? Pretty easy to get to the top of FPR and measure there - I think that the first step if not already done. If different from INPA then it’s a wiring or connector problem (maybe clean pins and connector?).
If it’s low resistance on a full tank - then it seems next step is taking the ring off and pull up FPR and you can probably get space to get something in to move float up/down and see if problem is in float sender?

What a pain, after just doing this it wouldn’t be something I’d want to get into multiple times…. Good luck

[Artmasterx]

I had a similar issue, and I think this solved my problem. Maybe it can help other as well.

After replacing my fuel pressure regulator, I had an issue where the left side level sensor would not read full. Typically the fuel gauge would read about 7/8 after filling up, and the resistance of the left sensor would be significantly below its max value of ~990 ohms.

I just took out the FPR again to try something. Not unhooking it completely, I left the tubes connected.

I tested the resistance of the float sensor from bottom to top, and it goes from ~50 to ~990 ohm, which I think is the right range. It also reads this same range in the hidden menu (so nothing weird with the wiring in between). How to get to the hidden menu, if you don't know: https://www.e90post.com/forums/showthread.php?t=4320. Menu 06.02 (I think) shows the resistance value of the fuel level sensors.

My working theory was this...
The FPR has that spring arm that basically goes straight down into the tank (not talking about the level sensor arm, but the plastic extender thing with the spring). And the level sensor is mounted to it. When I had the FPR out, there was a decent amount of play in that spring arm when fully extended. I think I bought the VDO version of the part. If it is "bent" in the fully extended state, then I think it could effectively make the level sensor not be able to float to the top of its range (because the sensor housing is now at an angle). I think this is especially true if the arm is bent toward the left side of the car, which is think is more likely with the way you have to insert the FPR.
If this is true, then the sensor may not read the full resistance and therefore the car won't think the left side tank is full, even though it is.

Now, if you look down in the tank below the FPR, you will a shallow depression in the bottom of the tank (straight down the opening, slightly forward on the bottom of the tank). This depression is meant to hold the tip of spring arm in the same place during bumps, etc. And I am sure it is also to make sure the level sensor is positioned correctly as well, since the sensor is attached to spring arm.

This depression also has a small (smooth) raised lip around it. My theory is that the when I put the FPR back in originally... and also the second time after I took it out again to check it... that I wasn't careful to make sure the spring arm tip was in the depression. It may have been caught on the lip of the depression, and then would be tilted at an angle when I close up the lid which could make the sensor read incorrectly. And because if the lip, it will not work it way back into the right place.

I tested it out today and it now seems to be working correctly. So I hope this may help some others as well.