An enduring riddle of fuel and air. - PeachParts Mercedes-Benz Forum

Shern · #1 05-03-2020, 07:52 PM

Come with me on a journey, through the banal and into the bizarre...
I promise, it's better than simple "air incursion."

What happens is this:

I shut the car down. I get what looks like drain back in the clear line, back into the IP.
Note: This does not form by air bubbles coming up. This forms by fuel draining down.

There are two methods by which this occurs. Either at once, as in a burp, which produces about an inch of air in the line, or gradually over a longer period (or both) which reaches a final length of 3.5 inches. Never goes beyond this regardless of time.

From tank strainer to cigar line, I have replaced nearly every component in the
fuel delivery system. I have rebuild the lift pump -except for the large spring. I have installed one of Greazers Performance overflow valves.
I have tried clamping off various hoses after shutdown. If I clamp the inlet before the primary fuel filter, it still disappears.

QUESTION: Can internal IP fuel pressure overcome lift pump valves only while the car is not running? OR... with a stronger OFV, can a lift pump over-power itself?

Thinking is this... lift pump creates pressure. Pressure exists in a circuit between the exit valve in the lift pump and the overflow valve in the Injector pump. When motor is running, lift pump pressure overcomes the OFV and exits through the return. When motor is not running, pressure within the lift pump overcomes the exit valve in the lift pump and drains back to the tank OR through the tiny o-ring in the bore and into the crank case oil.

If not... where in the hell is the fuel going???

I've been dealing with this issue for a couple of years now. Prize of fame, glory and canonization to be conferred upon the man who solves this enduring riddle.

Diseasel300 · #2 05-03-2020, 08:55 PM

Which fuel line is getting air in it? There should be a bleeder hole in the secondary fuel filter housing that conveys directly into the return line to the tank. It's entire purpose is to burp any air that may be present in the secondary fuel filter back to the tank. Have you checked to make sure that tiny little hole isn't plugged up with muck?

Alec300SD · #3 05-03-2020, 08:58 PM

You probably have an air incursion at the banjo fitting at the fuel filter head for your return line.
You probably need to refresh both crush washers on the banjo fitting.

Get the bolt tight to where it wants to turn the fitting, then secure the hard line with one hand while you rotate the bolt another 30 degrees to get it 'gudentite'.

What I believe happens is that there is an air bubble trapped in the hollow bolt.
When the car runs the bubble moves from the clear line into the hollow bolt.
The return flow is not enough to wash the bubble out of the hollow bolt.

The bubble magically reappears in the clear return line (highest point of the fuel system) when you shut down, as there is no fuel flow to keep the bubble in the hollow bolt.

Shern · #4 05-03-2020, 09:14 PM

Quote:

Originally Posted by Diseasel300

Which fuel line is getting air in it? There should be a bleeder hole in the secondary fuel filter housing that conveys directly into the return line to the tank. It's entire purpose is to burp any air that may be present in the secondary fuel filter back to the tank. Have you checked to make sure that tiny little hole isn't plugged up with muck?

The clear return line from the IP.
First I've heard of this hole... interesting. This might support the hypothesis below.

Quote:

Originally Posted by Alec300SD

You probably have an air incursion at the banjo fitting at the fuel filter head for your return line.
You probably need to refresh both crush washers on the banjo fitting.

Get the bolt tight to where it wants to turn the fitting, then secure the hard line with one hand while you rotate the bolt another 30 degrees to get it 'gudentite'.

What I believe happens is that there is an air bubble trapped in the hollow bolt.
When the car runs the bubble moves from the clear line into the hollow bolt.
The return flow is not enough to wash the bubble out of the hollow bolt.

The bubble magically reappears in the clear return line (highest point of the fuel system) when you shut down, as there is no fuel flow to keep the bubble in the hollow bolt.

I have replaced every crush washer in creation. That is an interesting theory though... Running with this, I'd have to purge the bubble? There's no indication a new bubble is being created each time. I've replaced the cigar hose with a clear 1/4" hose and no bubble forms. Taking this a little further, why would the expansion of the bubble (assuming the air was very compressed) occur over such a long period of time? The fuel still has to be getting out of the way and going somewhere, right?

Diseasel300 · #5 05-03-2020, 09:29 PM

For giggles, have you tried blowing compressed air down the return line to the tank? Do it with the tank cap off obviously. Maybe there's a partial restriction in there you're working against?

Alec300SD · #6 05-03-2020, 10:12 PM

Quote:

Originally Posted by Shern

The clear return line from the IP.
First I've heard of this hole... interesting. This might support the hypothesis below.

I have replaced every crush washer in creation. That is an interesting theory though... Running with this, I'd have to purge the bubble? There's no indication a new bubble is being created each time. I've replaced the cigar hose with a clear 1/4" hose and no bubble forms. Taking this a little further, why would the expansion of the bubble (assuming the air was very compressed) occur over such a long period of time? The fuel still has to be getting out of the way and going somewhere, right?

Bsaed on this info most (if not all) of your air leak is at the cigar hose.
You should not be able to twist the cigar hose on either metal line.

Venturi effect: whille the engine is running, air is being sucked in at the cigar hose,
traveling up to the hollow bolt, and lying in wait for the chance to get in your clear hard return line.

The venturi effect can also account for the formation of an air bubble in the hollow bolt from a leak at the crush washers.

engatwork · #7 05-04-2020, 06:52 AM

If it is on a 98 E300 then, as long as the cranks/runs, don't worry about it.

Shern · #8 05-04-2020, 01:45 PM

Quote:

Originally Posted by Diseasel300

For giggles, have you tried blowing compressed air down the return line to the tank? Do it with the tank cap off obviously. Maybe there's a partial restriction in there you're working against?

Gave this a shot -no restriction. Interesting Idea though.

Shern · #9 05-04-2020, 01:52 PM

Quote:

Originally Posted by Alec300SD

Bsaed on this info most (if not all) of your air leak is at the cigar hose.
You should not be able to twist the cigar hose on either metal line.

Venturi effect: whille the engine is running, air is being sucked in at the cigar hose,
traveling up to the hollow bolt, and lying in wait for the chance to get in your clear hard return line.

The venturi effect can also account for the formation of an air bubble in the hollow bolt from a leak at the crush washers.

Venturi effect is a compelling idea, but there are two issues here.

1.) If I take the cigar hose off completely (total open circuit), the clear line won't drain back immediately. I can still have a solid column of fuel, visible in the return line. Fwiw, it's on as tightly as possible without the factory clamps cutting into the rubber (does not twist whatsoever)

2.) I believe the narrower, non cigar-esque diameter of the temporary 1/4" clear vinyl hose creates a special condition. It doesn't empty the way the cigar hose does on shutdown. So as the fuel level in the return line drops, it is able to pull fuel backward from the 1/4" clear line via siphon

The FSM lists a method for purging the spin filter of air involving loosening the banjo fitting at the line going into the IP. You're meant to prime until all air comes out. I've never seen the need to do this but if I have a trapped bubble, perhaps this is the only way?

Any other ideas?

dieselbenz1 · #10 05-04-2020, 02:17 PM

Your lift pump has check valves and down stream is under pressure. Pinch off the return line then pressure the inlet of the lift pump to 10 psi and see how long it takes to bleed down maybe even find a pin hole leak with soapy water. One member found the fuel filter housing had a hair line crack.

Shern · #11 05-04-2020, 02:30 PM

Quote:

Originally Posted by dieselbenz1

Your lift pump has check valves and down stream is under pressure. Pinch off the return line then pressure the inlet of the lift pump to 10 psi and see how long it takes to bleed down maybe even find a pin hole leak with soapy water. One member found the fuel filter housing had a hair line crack.

Sorry, I'm not quite following you here.

I need to understand to what "downstream" is relative. Suction between tank and lift pump, pressure between lift pump and return -I assume this is the downstream you're referring to?

Does this require some kind of pressure gauge or is there an amount of time that should allow 10psi to build?

dieselbenz1 · #12 05-04-2020, 03:05 PM

Yes that is downstream. I have used a hand pump with a pressure gauge. If it is holding the 10 psi for 12 hours then it's safe to say you have no leaks.

Diseasel300 · #13 05-04-2020, 03:18 PM

Quote:

Originally Posted by Shern

Venturi effect is a compelling idea, but there are two issues here.

1.) If I take the cigar hose off completely (total open circuit), the clear line won't drain back immediately. I can still have a solid column of fuel, visible in the return line. Fwiw, it's on as tightly as possible without the factory clamps cutting into the rubber (does not twist whatsoever)

2.) I believe the narrower, non cigar-esque diameter of the temporary 1/4" clear vinyl hose creates a special condition. It doesn't empty the way the cigar hose does on shutdown. So as the fuel level in the return line drops, it is able to pull fuel backward from the 1/4" clear line via siphon

The FSM lists a method for purging the spin filter of air involving loosening the banjo fitting at the line going into the IP. You're meant to prime until all air comes out. I've never seen the need to do this but if I have a trapped bubble, perhaps this is the only way?

Any other ideas?

Let's back up. If you aren't getting the bubble with the plain vinyl hose, but you are with the cigar hose, you have one of two issues:
1: The cigar hose is leaking air back.
2: There is no problem. The cigar hose being larger in diameter than the rest of the tubing in the system will naturally hold air, just like the prefilter does. It will take weeks or even months of operation to burp the bubble out by absorption.

Are you trying to solve a starting and/or running problem? Or are you just obsessing about air in the line?

Shern · #14 05-05-2020, 04:02 PM

Quote:

Originally Posted by Diseasel300

Let's back up. If you aren't getting the bubble with the plain vinyl hose, but you are with the cigar hose, you have one of two issues:
1: The cigar hose is leaking air back.
2: There is no problem. The cigar hose being larger in diameter than the rest of the tubing in the system will naturally hold air, just like the prefilter does. It will take weeks or even months of operation to burp the bubble out by absorption.

Are you trying to solve a starting and/or running problem? Or are you just obsessing about air in the line?

I'm having about a minute or so of nailing on the first start of the day. Occasionally a miss until it all sorts itself out.

Okay.

So I went out yesterday and bled the system the FSM recommended way.

As soon as I loosened the banjo bolt, the level in the return line rose to normal. Instantly. I primed a few times until fuel was squirting out and then tightened it all up. Didn't help.

Your second point is what I was leaning toward. However, a.) I'm not getting a single chime that this is normal from the countless w123 drivers on this board, and b.) any siphon effect would leave more fuel in the line. As you can see below, the final resting point is far below this point.

I can prime and fill the hose up, but the level will gradually drop until it hits one of the points in photos above.
I can see it dropping as though it is being sucked back.

That fuel has to be going somewhere... there are some great minds on the case now. This time it's getting solved. What next?

dieselbenz1 · #15 05-05-2020, 06:23 PM

Pressure the line and find the leak.