[FNChaos]

Where did you get the low pressure fuel pump? Did it have an Audi part number or did it cross to something else? Who is the part's manufacturer?

 

[dr.electron]

I bought the pump from ECS Tuning, and it's made by Continental. They're a legitimate supplier, so I figured I'd save a few hundred $. If replacing the filter again doesn't do it (cheap to try), then I may pay the price and go get an actual Audi pump, to see if that does it. But it does seem unlikely that a pump from Continental simply wouldn't work properly out of the box.

 

[FNChaos]

I agree, Continental and ECS are both legit so I would expect the pump is good.
I saw a note saying an adaptation needs to be performed if you replace the pump controller. Maybe worth trying if you haven't already?

 

[dr.electron]

Oh - I didn’t realize that. I’ll investigate first thing in the morning!

 

[TT'sRevenge]

Well....I celebrated a bit too early. I drove the TTRS around a bunch today, and eventually the CEL came back on. Low pressure again. Now, when I look at pressure in real-time, it runs around 3.5bar at idle. Which is a little low. I cleared the CEL and it hasn't come back on (it used to come on almost immediately), but I'm certain it will at some point.

This is a new low pressure pump, a new sensor, and a new filter. High pressure still runs dead on where it should. Although it's much better than before, the low pressure side still appears to be on the edge.

Well I guess the only thing left in the equation is the controller?

The only thing I can think of is that in the process of changing out the low pressure pump, I got some dirt and crap in the gas tank. I vacuumed around the pump and cleaned up as best I could, but when I pulled the pump up there was a bunch of dirt under its edge, and I'm sure some of that made it into the tank. But it wouldn't have been much. So I suppose my next step is to replace the filter AGAIN and see if that does it. I can't think of another reason the pressure would be low.

Pump is running 68% duty cycle at idle - which is awfully high. What reason would there be for the pump to be working so hard to maintain pressure, even at idle?

I doubt you could have possibly put enough dirt into the tank to have clogged a fuel filter, at least not accidentally so. However you may have a restriction else where in one of the lines/hoses somehow. I don't know how that be given nothing that large should ever pass the fuel filter but stranger things have happened. But this also doesn't preclude the line prior to the filter (from the tank) being clogged. Possibly something to investigate there.

 

[dr.electron]

Per the Audi repair manual, the pump adaptation - which it describes in detail - isn't necessary for TFSI engines with a timing chain. Odd distinction to make.

Per your point about a clog/restriction prior to the filter, it has occurred to me that maybe I twisted/pinched/kinked one of the lines built into the pump when I pushed it back down into the tank. So I will likely go open it back up and look. The controller is a possibility as well, but seems like such an unlikely point of failure. I may get there, though.

The thing that gets me most is that after I replaced the pump, it worked fine for my initial test drive - which was spirited. It wasn't until I'd been driving awhile the next day that the pressure dropped. That's what originally made me consider the filter, since if I did get junk in the tank it could've taken a bit for it to accumulate in the filter. But I agree that it seems unlikely there was enough to cause a problem. I may replace the filter today anyway since it's a $30 part and a very simple process, since I have a lift.

 

[dr.electron]

So....... I went back out just now to start up the TTRS and check my low pressure fuel numbers again before replacing the filter, and they look normal again. 5bar at idle running 41.6% duty cycle.

I suppose this suggests that the problem happens once the engine heats up. I'll go for another drive later and log some values to see what happens...

 

[dr.electron]

Here are some plots of low-side fuel pressure, pump duty cycle, and engine temp. The top plot was me driving around for a bit this morning. As you can see, the pump was actually just fine for several minutes. You can see where I put the hammer down in the duty cycle plots that go nuts. Duty cycle starts to go up until it rails just under 70%, and then pressure starts to drop. This plot led me to suspect a temperature dependence on pump performance.

The second plot was this evening, after the car cooled down. I just started it (you can see where duty jumps from 9% right at the start) and let it idle until temp got to 94C. This time the duty cycle railed sooner, and at a lower engine temperature. Which suggests that if it's temp related, it isn't engine block temp, per se.

I considered if it might be the pressure sensor is responding to temperature, but it's on the fuel rail right next to the engine block, so I can't imagine the little bit of moving air from driving would help much. It's also a new part. I put a hair dryer on the sensor before I started that idle test and let it get almost too hot to touch. No change in pressure reading.

This does seem to exonerate the fuel filter, since everything worked fine after a cold start. The fuel filter wouldn't unclog itself for a cold start.

The only part I haven't replaced is the pump controller - which is rather spendy. I suppose it's possible that it was damaged when the old pump died. It would have to have failed in a way that it would be commanding 70% duty cycle, but not actually driving the pump appropriately. Tomorrow I'll try measuring the voltage applied to the pump to see if it's truly close to 70% of battery voltage. If it is, then that suggests that the controller is doing what it's supposed to, and either the pump is still bad or something else in the system is occluding flow. If the voltage is off, then the controller is the problem.

 

[dr.electron]

For what it's worth - this has stumped my local shop so far. They're really good about providing advice when I fix stuff on my own since I've given them plenty of business fixing things that I (a) won't do, or (b) get into and can't finish. They're down to replacing the controller at this point.

I'll continue to keep a diary of this little adventure here for future generations...

 

[SwissJetPilot]

Since your data is being collected with a Ross Tech VCDS, you can always post on the Ross Tech Forum and see if they can provide any insight. When you put up your first post, you'll need to upload your entire fault scan as proof you're a legit owner -

And if you haven't already, be sure you've registered your VCDS. Link here -

https://www.ttforum.co.uk/posts/9661048/

 

[dr.electron]

Interesting - I don't think I've ever registered my HEX-V2, though I've used 9 of the 10 allowable VINs. I'll go start a conversation over there if I don't solve this soon. Regardless, I'll register the HEX-V2 and I'll keep this post updated as well.

 

[TT'sRevenge]

I suppose this suggests that the problem happens once the engine heats up. I'll go for another drive later and log some values to see what happens...

...Or the controller. They are known for overheating and having issues! Maybe not super common but seems to happen enough.

Provided this isn't a 3-wire pump like in the B8 cars (I don't think it is but I don't know about the RS pump), then should be able to just bypass the controller run the pump direct so it's always at 100% and see if it stalls/cuts out again...but then wait I don't think the car is stalling, right? It's just giving the code for low fuel pressure and MIL and/or EPC on?

 

[dr.electron]

That's an interesting point about just bypassing the controller to see what happens. From the wiring diagrams, it looks like the pump is just two wires, and the other pins in that connector are for the fuel level sensor. As I understand it, the controller is just using PWM to drive the pump motor, which means it's applying battery voltage at a variable duty cycle. My next step is to look at applied pump voltage - with a multimeter and an oscilloscope - to see if the controller is really applying the power it thinks that it is. But to your point, I could also apply power to the pump directly on my own with a variable power supply and see what happens to pressure. I'll try that if my measurements don't help. I really hope this is the controller - otherwise I'm not sure where to look next.

 

[dr.electron]

Popped the rear seat off and set in to monitor signals to the fuel pump. It turns out that there are 6 wires to the pump. I'm confident that three are for the fuel level sensor - power, ground, and signal out (it's a potentiometer, I believe). Two are presumably power as they're larger gauge wires for battery and ground (ie. battery negative). The last is the duty cycle signal from the controller.

I assumed the controller was modulating power to the pump by duty cycling the applied battery voltage - but that is not the case. I looked at the duty signal with an oscilloscope, and it appears to match what VCDS is telling me for the pump duty cycle. I once again let the car run until the duty cycle rose and railed - it took a good long time today, since the car was completely cold (still looks temp related...). I looked at the signal when the car wasn't running - 9.8%:

After I started it when it was still cold - 41%:

And once it railed - 68%:

These duty cycles all seem to look right, so the controller was doing it's job and increasing duty cycle as pressure dropped.

This means that either the pressure is actually dropping because the pump can't keep up, or the pressure measurement is wrong and it's actually higher than it appears. For the pressure measurement to be wrong, either the pressure sensor is sending the wrong signal as things warm up, or the controller is reading the signal incorrectly as things warm up. If the pressure is correct, then the controller is doing the right thing but the pump can't keep up. That would either be because the pump starts to fail as it warms up, or because there is a restriction to fluid flow that increases with temperature (or time).

Not sure where this leaves me. If the controller weren't so expensive and difficult to get, I'd just replace it to see what happens. But I think I'll swap back to the original sensor next to see if that makes a difference (I'm betting it won't), and then I may try specifically testing the pump per the Audi methodology. It'll be a pain, but I can do it. If I rule out the pump, the controller is all that's left, really, since I can't see how there would be a time or temperature dependent flow restriction.

Someone please correct any of my assumptions that seem questionable.

 

[dr.electron]

Upon closer inspection, I've been tricked.... I was looking at the connector under the rear seat, on top of the plastic cover for the fuel pump - where there are clearly six wires. However - if I pull the cover up, I see that the harness on top passes through, goes off to the controller, and comes back with a FIVE pin connector to the pump itself.

So I need to repeat my test above and look at the actual voltage applied to the pump. What I was measuring was the voltage provided to the controller, and the duty cycle the controller was reporting back to the ECU. I was NOT looking at the actual voltage applied to the pump. I shall try again tomorrow. Stay tuned.

 

[SwissJetPilot]

We have a wiring diagram if you need a reference -

2.2 Wiring Diagrams & Component Locations - A005TT20021

https://www.ttforum.co.uk/threads/faq-audi-tt-8j-workshop-manuals-self-study-programs.1833829/?post_id=8991881#post-8991881

or here -

http://www.lz7w.com/audi/mk2ttwiki/pdf/erwin/

 

[Vassilis]

Is the car remapped ?

 

[dr.electron]

Is the car remapped ?

Yes - it is, now that you mention it. The previous owner put a United Motorsport tune on it. I have the boost set to the minimum (you can set it with VCDS) - which is 20psi. But I've been driving the car for a year, and this is the first time I've had fuel issues.

 

[dr.electron]

We have a wiring diagram if you need a reference

Thanks! I actually downloaded all of the documentation from Erwin when I bought the car - before I realized you had all of it here, as well. Excellent community resource.

And looking at the wiring diagram again led me to a revelation: the low fuel pressure sensor is connected to the ECU, mnot the pump controller. So the duty cycle signal I was measuring comes from the ECU and goes to the pump controller. All the pump controller is doing is converting the duty cycle command from the ECU into an appropriate drive voltage for the pump. And it's also somehow translating fuel level sensor info for the ECU.

With this in mind, I wonder if the pump controller can truly be replaced with 3C0906093A. The 3XX part is compatible with my TTRS according to the Audi USA parts website, and that would explain why I can't get the 8J0906071A anywhere in the US. If this is the case, it would be great because the 3XX part is far less expensive and readily available. Right now, the wiring from the ECU connects to a harness that passes through the cover over the low pressure pump, goes down around the gas tank to the controller, then comes back from the controller to plug into the pump. It's #7 in the diagram. Connector 8 plugs into the harness back to the ECU:

Part 3C0906093A just connects in-line between the ECU harness and the pump, and fits into a large indent in the pump cover. Given the relatively simple function it's performing, it may do the job. My only concern is that the TTRS fuel pump is a higher-power pump than the one in other cars that use the 3XX controller, so the 3XX may not be able to drive it sufficiently. However, if that's the case then why does Audi USA say it's compatible with the TTRS?

 

[SwissJetPilot]

Be a little wary of the 3rd party illustrations as they're not always 100% correct and sometimes not even for a Mk2 TT.