AWR-fix: Siphon pump repair
 

EDIT: another member did an amazing more detailed step-by-step to do this process. I recommend reading my thread for additional information then use his thread add the recipe to follow. The thread is here:

https://r.tapatalk.com/shareLink?sha...0&share_type=t

So tonight when I was about to take a 3 hour drive up north, I go to a pit stop on the way and my car stalls with 18 liters of gas.

Being the self proclaimed master of the BMW X5 fuel system, I figured it was the universe giving me my chance to prove myself.

After three or four start attempts that went nowhere I double checked that the fuel pump which is less than 1 month old was working.

I could hear the hum of the pump from under the car for 3 seconds I knew it was getting power I wasn't sure if I was making pressure.

I pulled up test 6 from the hidden menu and sure enough it had about 17 liter of gas on the left and 0.0 liters of gas on the right.

This points to an obvious problem in the siphon jet.

I put in about 4 gallons of gas so I could get home. I took out my rear seat, open up the left side of the tank and holding the tank open a little bit turn the key to the on position. I was recording with a camera with video so I could share with the group. I will find out soon enough if liquid proof includes gasoline as my iPhone got sprayed with gas.

https://uploads.tapatalk-cdn.com/201...f60bacccb6.jpg

Here is the main reason the O-ring just popped right out of the groove just like my wife's car.

It was my primary suspect and I was planning on redoing that o-ring.

I replaced that O-ring and did another test. Not at all surprising to me the useless seal in the other connection had failed as expected. It was only leaking maybe 5% as much as the huge leak from the O-ring it was bad enough that I wanted to fix it and figure out a good way to show other people they can fix it easily.

https://uploads.tapatalk-cdn.com/201...f63c40f876.jpg
Here's the new o-ring in place

https://uploads.tapatalk-cdn.com/201...3733e0698e.jpg
Here is the useless seal taken apart. Be careful taking this apart I use a piece of wire to hold open the one clip at a time as I pried it apart. If you take it apart carefully enough you will not have to create a new way to hold it together. If you break it taking it apart don't worry you can use a hose clamp to hold it together when you're done.

Be very careful about the spring and the one-way valve I almost lost the part as the spring shot it across my car.

https://uploads.tapatalk-cdn.com/201...49a9b8bfc8.jpg

https://uploads.tapatalk-cdn.com/201...731571365d.jpg

The female part is .2 millimeters bigger than the male part so of course it leaks under pressure.

Rather than build an O-ring Groove which is a great solution I wanted a quick and easy solution

https://uploads.tapatalk-cdn.com/201...2dabec903a.jpg
4 wraps of fastape Teflon tape folded in half so really eight layers gave me a nice tight fit that does not leak a drop.

https://uploads.tapatalk-cdn.com/201...f345ee6ae6.jpg

https://uploads.tapatalk-cdn.com/201...139d72b0d7.jpg
Here is the "useless seal" reassembled. Notice that it's a perfectly tight fit. Gooder (sic) than new.

https://uploads.tapatalk-cdn.com/201...64e5e21bf5.jpg

https://uploads.tapatalk-cdn.com/201...b1b36918e1.jpg

And I think part of why these all fail is because of the way it is held together. The O-ring seal is being held way offset. I added this little zip tie to help hold the O-ring tighter. Normally there will be quite an angle and the side with the O-ring will be about 1.5 mm lower. Zip tie acts like lever on the fulcrum in the middle and holds the fitting around the O-ring tighter.

https://uploads.tapatalk-cdn.com/201...77d84666c1.jpg

Yes that's the hose clamp down in the bottom of the gas tank! Be careful not to drop crap into the gas tank.

https://uploads.tapatalk-cdn.com/201...2f24c7bb24.jpg

While I was rebuilding the old ring I sealed that the tank with a piece of an old glove stretched over the opening.

In the morning when I go to do my three-hour trip up north, I will bring a gas can a long and drive down to single digits to make sure my siphon pump is working. Since about a half a gallon a minute less is spraying out the side, I am quite confident in the repair!

Siphon pump DIY repair
 

http://www.youtube.com/watch?v=jsgWAnmvKeg

Don't worry it looks like an amazing amount of gas sprayed everywhere but it's exaggerated by the phone camera being about 4cm from the leak.

It was a holy crap amount of leak though and absolutely why my car starved of fuel with 15L or something remaining.

I just replaced the electric pump a couple weeks ago and I'm curious if the new pump with more pressure blew or the o-ring or the old pump actually had more pressure to overcome the leak because I could drive to zero DTE just a month or two ago, I perform that test about 3-4 times a year.

This job could be done in about 90 minutes especially if you have a helper. As I was photographing and inventing it probably took me about 3 hours.

About the sixth time I've done the remove a fuel pump procedure on an x5, this was the first of all the future times where I removed the back seat bottom.

It is captive by the center position seat belt. It takes a 16mm socket and several hundred ft·lb of torque to remove the rusted in bolt but with the right power tool it was simple. (was a bit of a challenge to get the rusted bolt started to put back in but still quite worth it, fighting the damn seat bottom the entire time really gets on your nerves).

Putting the seat bottom back in; push the center male buckle through the seat and latch it: since it is floppy, it will fight you trying to put it up through the hole in the seat.

The other two are pretty easy: pretty much ignore them, get the seat lined up and start to push the back down into place, the left and right female buckle should be lined up with the holes but flat forward, there is enough give in the leather to reach a skinny finger in and pull the front edge vertical into the hole.

At that point just pound the seat backwards with your fist then down again with fist it should snap right into place.

Very Interesting, had I known you were taking the fuel sender out, I would've begged for a picture of the float and arm with some form of reference of angle...

Luckily this seems to work on mine but I'll check the o ring next time as a precaution.

What you show as the one way valve, is this what keeps fuel in the fuel rail when the engine is off or something completely different?

Siphon pump DIY repair
 

The check valve you describe is either in the electric pump or the FPR (fuel pressure regulator) built into the fuel filter I forget which, but I'm leaning toward the pump since I get a lot of spray when disconnecting the supply hose.

This check valve does not seem equipped to hold pressure just to prevent back flow. I suspect part of the "oh shit the siphon foot is plugged, let the pressure out" system, in the case pressure builds in the siphon feet, the pressure would prevent flow backwards and let the pressure relief valve dump the excess fuel directly info right tank.

I don't seem to have any pics of the sender with the arm to measure the angle but I have discovered that the floats on both sides hit the top of the tank at a total of 82L of fuel (on my car at least).

https://uploads.tapatalk-cdn.com/201...fa8c724869.jpg

55 and 27L.

The tops of both sides are symmetrical, so if they both stop climbing at the same time and at 55/27 the arm bend is correct (or the foot is properly engaged into the dent in the tank).

You should be able to put in about 10L once the test 6 numbers max out as the floats smack into the top of the tank. They should both stop climbing at the same time.

(You can also observe them coming off the top, that should be a lot easier come to think of it). At about 60-70 miles for me, I finally saw the numbers start to wiggle as I hit bumps or took curves.

If one of the sides starts to show a drop off the top sooner than the other i would investigate.

How the hell the computer guesses a DTE before the first 10L is consumed I haven't a clue. I think it assumes a full tank any time you are within 10L from the top; maybe explains why the DTE can swing pretty hard once the floats can actually measure how much fuel is in the tank.

I'm going to check my left fuel sender as it doesn't report 55 when full (event though its new from bmw) when the right one now reports 27...

I'll check the oring at the same time. I'll also remove de middle seat belt, I've got no use for it in the foreseable future and it truly is a big annoyance.

To make the sender report a higher number, you need to straighten the elbow on the float apparently about 1-2cm = 5L.


Sent from my iPhone using Tapatalk Pro

I've got a gas station about 2 miles from the workshop so I'll fill up and make it see full.

But that's gonna have to wait like everything else, its way too cold, windy and wet right now for me to do anything that's not urgent or stranding me.

I just filled mine up and yep - exact same numbers (as always) - 55L and 27L. And like you, I get about 120km before the numbers start to "wiggle".

And before I filled it (it took over 71L to fill) I had confirmed that the right tank was reading 1.4L. Left was reading about 20L. So siphon should be good. :thumbup:

The key is the right side should read the same (1.4 ish) down to 0.0 left


Sent from my iPhone using Tapatalk Pro

Understand the theory but I'm not likely to try often it in practice!

1. Water is an BIG problem for CRDI engines (or more correctly, the mechanical fuel pumps in CRDI engines!) - and I don't want to find out there's a few hundred millilitres of water sitting in there!! Pretty certain there isn't - I have had the car down to a few litres before but I also don't trust all service stations - it's not unknown the get a batch of contaminated diesel - rare but not unknown.

2. I just paid well north of $100 for the 70L of diesel I just filled up with - the "sticker shock" isn't insignificant! :yikes:

I'm just happy to see the right side stop at 1.4L while the left more or less matches what the gauge says. :thumbup:

Oh don't do it often, but once or twice a year run down to single digits with a fuel can in the trunk. The only way to know if the siphon is working. That said less than an hour your can pull the left side sending unit to verify the seals are good if they are still in good shape you could drill a couple small holes an put some stainless screws to prevent the o-ring from popping out. I will show that the next time I have the siphon side out.
I can't speak for diesel but with gas in most of America, one of the few few few benefits of polluted gas (E10) for non corn farmers is that the alcohol absorbs the water and removes it.

The x5 fuel system circulates the fuel so fast that no fuel it all gets left behind it's totally mixed homogeneous. If there are any impurities they will get sucked away and into the filter. Due to the rapid mixing I'm sure that any water will also get hold in slowly a little bit at a time


Sent from my iPhone using Tapatalk Pro

Yeah, water and petrol (and esp. E10 etc.) will "mix" but not so with diesel - there's no way to get them to mix and if water gets into the mechanical high pressure (HP) fuel pump, it will destroy the HP pump pretty quickly, galling the HP pump faces due to the low lubricity of water vs diesel. :wow:

The E53 3.0D has at two fuel filters and two electric fuel pumps before the HP pump so water getting into the system can render a vehicle useless (not cost-effective to repair).

For some reason, I feel like I need to qualify this statement. :dunno:

The bit that is important here is not that the right side stops are 1.4L, it's that the left side continues to drop AFTER the right side stops at 1.4L. And that's the key, as andrewwynn has been saying. And has also been stated, you can be doubly sure the siphon is working well by letting the left side run down until virtually empty, as verified by the BC readings.

Andrewwynn - thanks for your work on this. My son's car has approximately 150K miles, and it seems as if it would be worth a bit of preventative maintenance to both 1) replace the pump in the fuel sender and 2) replace the o-ring/washer in the siphon unit. If I understand things properly, you also placed a zip tie so as to prevent the plastic connections from getting skewed? I think I understand about the teflon tape to make a slip connection tighter, but my understanding is that teflon tape will be dissolved in gasoline. My understanding is that it may be OK when used in pipe threads where there will be essentially no contact with gasoline, but not OK when exposed to gasoline. A quote from another site: "inside the seal, it should never touch gas, in which case teflon is your best bet becuase it is clean and seals well. but gasoline is made up of 2 perticularly teflon breaking down components, benzene and cycloalkenes. benzene will dissolve the coherent bond of teflon and aluminum, and the alkenes will break down the chemical bond,y ou will be left with a white goo."

So if the teflon tape is going to dissolve, it seems as if a gas resistant o-ring is a better solution. However, I'm not sure I want to try to carve out a groove for an o-ring. What about a stainless clamp to give a bit of a squeeze to the joint?

- My 2006 X5 M54 6spmT with 128K runs fine, zero issues.

- Should I do this "siphon pump" fix on a preventive basis?

I recommend preventative repair of the siphon pump "seals" at about 100,000 miles or more accurately 3000-4000 hours. Or: when the electric pump fails or any excuse to remove the back seat.

Teflon is INERT. I'm not aware of a common chemical that dissolves Teflon.

From the oatey website:

I have to remove my jet pump sender soon and will disassemble the joint I fixed with Teflon tape specifically "FASTape" brand by Oatey. I will report back of course if that was a really bad idea or as predicted perfectly fine idea.

I will also get a sense of the nylon zip tie trick worked to keep the o-ring joint square and functioning.

I actually recommend FIRST check your jet pump "seals" (quotes on purpose as shade on the piss poor design) because I probably replaced a perfectly ok pump with 20,000 miles of life because the jet pump was spraying like a garden hose

Update: I dug into my fuel pump and gauge readings today; the knock off sender unit has a rectangular vs square float and was hitting the back of the tank.

https://uploads.tapatalk-cdn.com/201...88d79b7c96.jpg

I promised I would look at the Teflon seal when I dug into this problem and so I did.

As expected the Teflon which is INERT did not react with the gas at all. I actually un-wound and rewound the same piece I put in five months ago. It looked and felt exactly as it did the day I installed it.

https://uploads.tapatalk-cdn.com/201...e53c18aad6.jpg

The o-ring fix also worked great https://uploads.tapatalk-cdn.com/201...6adf331734.jpg

I will post update for the right side pump and sender unit since I've had to re-enginneer the float that was hitting the back of the tank.

Apparently the OE sender unit has a lower limit of 1.4 L at the flood stage of the surge tank. The knock off was technically more accurate as it will drop down to 0L much closer to the bottom of the tank but was not accurate at the flood stage but was coincidentally closer than the OE BMW part. I tweaked the float arm to get an accurate reading at the flood stage so I will get a hyper accurate DTE (and as a bonus the gauge actually displays the actual fuel level vs "off by a gallon" the factory builds into the design

Thanks for the update on the teflon withstanding gasoline. I guess you can't believe everything you read on the internet! Great to hear the repair is standing up to the test of time.

Like mentioned before, it is likely that the low quality thread seal tape may be plastic and not Teflon.

Do you have the dimension of O-ring and where to get it?

Sadly I just rummaged for parts in my basement and never measured the o-ring. If you look at my master thread on fuel system diagnosis I may have taken a pic with my calipers to show the size.

The size is certainly metric and a standard size you could buy at a hardware store. I had some flashlight o-rings like 18-19-20mm in my basement I thought it was one of those but the other seal is about 27 it could be in that range as well. If nobody beats me to it I'll get a measurement the next time I take apart one of these.

I already did both mine and wife's but I will do the siphon refurbish on any BMW before about 4000 hours


Sent from my iPhone using Tapatalk Pro

So to confirm, the Teflon Tape was the way to go and you haven't had to go back and mess with it after the repair? ALso... I couldn’t tell where you put the zip tie?

https://uploads.tapatalk-cdn.com/201...e0ef8b8e0a.jpg
• Yellow points to the zip tie. (It's wrapped around the post inside)
• Red shows the direction the zip tie pushes
• Blue shows the force of the fulcrum
• Green shows the resulting force up on the o-ring side to prevent the design flaw to repeat

Teflon tape was to repair the non o-ring seal (other design flaw: a flare made of plastic with no way to add compressive force that gets a D- guys who the hell approved that design!


Sent from my iPhone using Tapatalk Pro

https://www.realoem.com/bmw/enUS/sho...diagId=16_0460

Looking at realoem tells me the O-ring is PN 16116761645, about $2.50 at dealer.
Size 22x2 mm.

Thanks. I think It was posted on the fuel pump thread the o-ring size but illogically not here.


Sent from my iPhone using Tapatalk Pro

I have a theory as to the root cause...

Somehow the O-ring shrinks a bit (? Ethanol, ? Other chemical, ? Normal aging of rubber BUNA etc.)...then the pressure from the return fuel (probably around 30 psi in the return line) pushes the O-ring out of its groove ---> reducing the flow necessary for the Venturi effect to work.

Siphon pump DIY repair
 

Negative. The plastic twists over time.

https://uploads.tapatalk-cdn.com/201...e0ef8b8e0a.jpg

Refer to this photo, without the zip tie, the bottom piece will rotate clockwise since the force down is only on the right (opposite of Green arrow)

It can twist far enough that the o-ring will push out of its seat.

All that is needed to prevent this is some counter force to the force twisting clockwise, the zip tie does the trick and is a permanent solution.

So I have another question... in this screen cap I circled in red where the zip tie goes. But does it just loosely sit around the odd shaped part? Does it need to be tight? Seems like it would act as a buffer for the connection not to move back and forth if that’s the case, is that why you used the zip tie because it’s such an obscure shape and there isn’t anyway to create an spacer for it? Also in the green circle, am I missing a picture step? Please advise.

http://i284.photobucket.com/albums/l...stekrkq1g.jpeg

I think it's the latch (head) on the zip tie that is doing the job here. It stops the lower section moving up (pivoting) and sitting wonky....

That’s what I was thinking. So it needs to be tight enough to keep the head in the gap. Wouldn’t it be better to move it to the other side?

The further the "stopper" is from the pivot point, the more effective it is. ;)

I did position the head opposite from the fulcrum for exactly that reason. I wasn't sure if the main body of the tie was thick enough to not fit into the socket.

https://uploads.tapatalk-cdn.com/201...d58d098737.png

https://uploads.tapatalk-cdn.com/201...d47271431f.png

You can see that it actually holds tighter than straight. I am ok with that as once operating there will be torque from the gas pressure pushing to straighten it out.

SCIENCE!

Lol. Thank you guys for clarification. Sorry to ask so many questions just wanted to make sure I was on the same page.

Thanks for this thread.
I changed the O ring and was testing to see if the siphon pump worked properly now. As it turns out, it was. I ran out of gas on an off ramp today near work. Between the 3+ gallons from the red gas can a guy from work brought and the 22 gallons at the station down the road, I proved it runs all the way to empty. I just need to figure out why my level indicator stops and sticks just a shade under 1/2 tank so this doesn't happen when it's in the middle of a snow storm. I had the transfer pump side pulled up for a bit when I fixed the O ring and possibly got the level sensor arm twisted or caught on something in the process of putting it back in?

My fuel gauge got stuck at about 3/8 of a tank. It was from the sending unit on the left side reading high. I have a how to on XO to fix it yourself. The spring part of the wiper is not pushing hard enough on the resistive surface.

Search awr-fix in the title and you should find it. It's not a difficult fix. I found a BMW service bulletin on how to tweak the resistor wiper. I only saw it once and couldn't find via Google so it was all from memory but it worked like a charm my sending unit snapped into action and has been working perfectly since.

I found the other post. Nice write up.
I'll check the sending unit. My first guess is that I bent something or that the float arm popped out of the wiper bracket rather than the wiper itself. I do have 1000 grit wet or dry paper. Personal experience here on a similar repair to an e46.
In that case the float arm popped out and was pivoting at will. I'll wait just a bit because when I filled the tank after the event, I topped it off to see how far down I was when I ran out of gas. Bone dry I'm pretty sure.

You can tell when you are about to run out of gas by watching test six. The right side will stick at 014 or so until both tanks are depleted. Once that starts to drop you have about 3.6L of gas and can go maybe 10 miles before a walk with a gas can is in your future.

I didn't put teflon tape on the non O ring seal. Is that why the test 6 results show 020 to 000 on the right side as the fuel sloshes around? 020 on left hand turns and 000 on long sweeping rights. The left side is reading 027 +/-

Once below 28L or so total the right side effectively is empty. It should register about 1.4L on level ground.

When you accelerate or turn all hell breaks loose with the floats thats why the dash uses a buffered value.

A firm right turn will spill more than normal fuel from right to left and can eaisky register 000L for a bit until the siphon jet sucks some fuel from left to right.

If on level ground the right side drops lower than the 1.4 ish, then you have siphon problems and need to address it.