Fuel injector leak or HPFP?

[Smac770]

As for the Bosch LSU 4.9 broadband lambda sensor, search on google for "Bosch Motorsport Lambda Sensor LSU 4.9", should find a datahseet PDF, the second page has a chart and footnote of interest.

IDE00558 is the broadband sensor voltage value. IDE00559 is the broadband sensor current value and related lambda value. IDE00560 is the jump sensor voltage value.

[morris39]

Morris, as 1874 just say test off, I assumed the alternative value would be test on or test running or something, not an actual numerical value. I'll have to log that value too and see what comes up.
It could be then the 1873 is the integration sum computing the current cat OSC, and 1874 is the result of the division by the OSC value of a "marginal cat" from the lookup table based on modeled EGT and measured air mass integral during the test.

I start with the Quick Reference Specification Book for the model and year. I find these from google searches, often as documents at NHTSB. These provide a decent scope of what Audi actually uses on that engine as a trigger for that P code. Though sometimes, like with the P0420, you look at it and think "saaaay whaaat?"
You can also review the Audi and VW EPA certification applications on the EPA site. The Audi docs tend to just say "see confidential section", but the VW docs tend to have the OBD summary chart of P codes.
There's also the Bosch ME7 engine strategy document on ERWIN; it might be out there on the net too. As it's ME7 and not MED17, it's a nice starting point, but you can't take it as authoritative.
Then there's the various Audi and VW SSPs out there.

A US 2015 S3 would be engine code 215kW CYFB. Is that what you have? A CYFB is a Gen3 engine, so something we never got in the NAR market B8 vehicles. I can't find a quick ref spec book for a Mk3 A3 on google. Maybe someone with an active ERWIN access and pull it.
As such, I can't say what the P0420 trip point would be for that car, though division result of <1.0 seems logical assuming cat efficiency is testing in the same manner still.

Thanks for taking time. Yes it is a CYFB motor. I ask here b/c the S3 crowd are still mostly into cosmetics. I enjoy reading your posts, always with specifics. My interest is mainly to establish a data base which I can use for trouble shooting when the time comes. I am too old/stiff to enjoy most of the physical wrenching particularly without the pro equipment.

[silver_tt]

Just my opinion but I am not that surprised, aftermarket cats are a pain in the ass when it comes to jerking around with the readiness states (and based on past experience it also can take driving many miles to reset them even after things are fixed). So it is not surprising to me that a non-OE cat is not working perfectly with your car even if it isn't "defective" per se. That's why I was saying just go OE, it's not that much more expensive and then you can eliminate this being part of the problem. Also being a CA car won't make any difference.

smac is very smart with great posts but I also don't 100% follow on the ECU making up the voltages. The oxygen sensor is exactly that -- a oxygen sensor. It measures the amount of oxygen in the exhaust and produces a voltage which it sends to the ECU -- not the other way around. Looking at your voltage is the same thing as looking at the lambda since they are a function of one another. A wide band just measures the amount of oxygen much more precisely rather than switching back and forth between lean and rich like is happening in that YouTube video you posted. That guy in the video also says only use OE cats FWIW, I didn't watch the entire 30 mins but I watched clips and heard him specifically say that which I obviously agree with.

[Smac770]

That's how the wideband sensor works. The pump current is the measured value. It can be translated to a voltage, as it's a "simple" circuit. www.youtube.com/watch?v=Dwij_fuceUo

As I already mentioned, this is already pointed out by Bosch themselves.
https://www.bosch-motorsport.com/con...or_LSU_4.9.pdf
Page 2 shows a chart; read the foot note. Our voltage seems to follow the v=8 scaling. The formula for voltage value from IDE00558 is simply 1.5v + 8*61.9*(current value from IDE00559/1000). But that voltage is a computation based on the actual sensor measurement, pump current. The 1.5v base voltage seems to reflect that our ECU must use a CJ125 controller rather than a AWS controller. See Bosch tech doc Y 258 E00 015e. Bosch doc 0 258 017 025 can also be useful, but note that the Block Diagram pin numbers are actually for LSU 4.2, not LSU 4.9.

http://circuitszoo.altervista.org/Bosch_LSU_WB02.html
That might be useful in understanding the circuit layout.
Keep in mind that while LSU4.2 and LSU 4.9 sensors use the same 6 wire connection, the wires are in reverse order between the two. Ip, pump current, red wire, is pin 6 on LSU 4.2 and pin 1 on LSU 4.9. We use an LSU 4.9 sensor.

https://www.scannerdanner.com/media/...sistorchip.jpg
It's kind of interesting. It basically uses a voltage output old style O2 cell to produce a control voltage, the Nernst voltage, which is compared to the 450mV standard, to determine if the content is lean or rich. Then the control drives the necessary current in the pump cell to restore balance to the Nernst cell, 450mV, and the amount of current used reflects the lambda. Or something to that effect.

A proper cat with sufficient OSC is very expensive to build. If you have to care about emissions, you pretty much can't bother with anything but auto manufacturer OE. They are the only ones going to bother with the cost, since the EPA gives them no choice.

The 2013 A4 quick ref spec book says CAEB P0420 = Measured OSC / OSC of borderline catalyst value for main catalyst , <0.90. The 2009 makes reference to <1.00 and <0.90. The EPA docs are all "confidential info", but later EPA docs say <1.00. Probably more of that VAG emissions gray area. For example, there is a TSB for the original Q7 with 3.6 FSI engine for P0420/P0430 (same thing but bank 2). An ECM flash is necessary. Maybe an errant computation, errant lookup table data, or they were just knocking the result bar down a tenth or two.

[Smac770]

https://www.audizine.com/forum/showt...1#post12214451

I knew there was a P0420 TSB closer to home, but it's the 3.0T engine.

[silver_tt]

I don't see any footnotes if I have the right document but I think you are talking about the please note section which I saw previously yesterday:

Please note: UA is not an output signal of the lambda sensor, but
the output of the evaluation circuit. Only IP correlates with the
oxygen content of the exhaust gas. Amplification factor v=17 is
typically used for lean applications (lambda>1), amplification
factor v=8 is typically used for rich applications (lambda<1).

IP is a function of the oxygen content of the exhaust and UA is a function of IP

[silver_tt]

For the TSB you posted it looks like on the 3.0T they narrowed the spec (ie. "The diagnostic limitations were too narrow."). That's pretty funny.....

For "bad" cats usually you can tell pretty easily. Either the honeycomb is broken inside or you can see that it is otherwise compromised when you look inside. For both of my cats when they went out it was very obvious once I got them off of the car.....

Thanks for the info smac, I love it.......

[Smac770]

Yeah, that's the note. Says only the current value correlates with the oxygen content; Ua (German doc, so U is V ie voltage) is the output of a computation. And the pretty chart there with measured Ip vs lookup lambda vs computed voltage.

[wilmar13]

For "bad" cats usually you can tell pretty easily. Either the honeycomb is broken inside or you can see that it is otherwise compromised when you look inside. For both of my cats when they went out it was very obvious once I got them off of the car.....

My OEM cat with 100k mi looks fine, and just like the new $400 49 state aftermarket one did new, except the honeycomb was more tan than white... after 150mi and like 3 resets the aftermarket looked the same as OEM. I should have taken picture but I sent it back to rock auto, and they gave me full refund anyway.

A refurbished OEM one is on the way. It does say new in the listings but if you dig deeper, all of the Audi part numbers (8K0-254-252-KX for +'14 or 8K0-254-252-HX for '09-'13) in the $1300 range (which included $275 core charge) are actually refurbished. I doubt they have a way to code for this in their inventory system since how many things does an audi dealership sell that is refurbished? An actual new one lists for $2100 and has no core charge.

The NA E46 M3 had two cats placed right next to engine in headers... I was able to weld in 100cell race cats in the midpipe, extending O2 sensors, with fiberglass wrap on supersprint stepped headers and not trigger a CEL/MIL. It did trigger a P0420 until I wrapped the headers and midpipe over cats and sprayed silicone sealer on them to retain heat to the cats, but it worked great. Its comparing naturally aspirated 20yr old emissions controls vs 10yr old DI turbo emissions, but man I never thought a Federal aftermarket A4/A5/Q5 specific cat would not pass immediately. I was more worried about its lack of flow compared to OEM (cat was smaller diameter and the tubes had generic tubing bender grip marks). haha

[silver_tt]

hmm, the OE cat that was throwing the P0420 looks totally fine inside (your picture is just the outside)? If the "new" (didn't know it's refurbished but if you're buying it directly from an audi dealer, not a 3rd party, should be just as good) one looks the exact same as your original I really have no explanation for that.

The P0420 has come on a bunch of times since you cleared it right? In other words this is not just a one-off code .... you said in your first post you already replaced the O2's. If you look at that error code, you already replaced the O2's so scratch those. You also checked for a leak and said you used a pressure gauge which based on your reading was at least not grossly off. Manometer would be better but honestly if you had a leak that large it should be showing up in other codes further upstream from the cat in the engine itself. So the only thing left is "Catalyst faulty".

Possible Causes
Leakage in Intake and/or Exhaust System
Catalyst faulty
Oxygen Sensor(s) faulty
Oxygen Sensor(s) Control faulty

Possible Solutions
Check Intake and Exhaust System for Leaks
Check Catalyst
Check Oxygen Sensor(s)
Check Oxygen Sensor(s) Control
Perform Oxygen Sensor(s) Aging Check

http://wiki.ross-tech.com/wiki/index...4/P0420/001056

[wilmar13]

The car never had P0420 that I know of until I installed the new turbo. I had an intermittent P0299 for a long time, but it would stay off for long periods of time, then my girlfreind would say "the check engine light came on again" and I would generally just clear it after seeing it was P0299... its possible that one of those times it registered a P0420 as well and I didn't notice, I don't generally do stuff like that, but I am allowing the possibility it happened. That's just me trying to be thorough regarding the P0420 code as to whether it ever occurred before new turbo.

Yeah the pic is OE 100k mi cat on right, new aftermarket cat on left. When I say the OE cat looked fine, I mean the catalyst honey comb. No visible cracks or blockages on either end (you can see backside pretty well with light and O2 sensor removed)... but you can't see if the substrate is worn off. I reread a bunch on how to identify if your cat is bad (clogged, melted, cracked, etc) when it initially happened which was why I really looked hard at the OE cat.

When the aftermarket cat produced a P0420 right away, I looked for exhaust leaks... then intake, then I put old O2 sensors back in just to see... then got further into intake leakage diagnosis at which point I started this thread wondering if it was fuel injector related as the rail doesn't build pressure with heat soak as it should. There is definitely no major vacuum leaks as pressurizing intake at turbo even up to 30psi produces nothing (had to tighten an intake clamp at 25psi, but that should not be an issue), only leaking into crankcase at rate that jives with what is slipping past throttle valve. I have not even driven car 600mi since new turbo installed, so its possible that adaptations are still taking place to the new normal. At this point, I am just going to wait for new OE cat and see if I still get P0420. If so then I have no choice but to remove intake manifold to look at carbon buildup and see if there are any small leaks due to gasket or injectors. I am just spinning my wheels until then. Any fuel leak on HP side is minuscule as the fuel pressure doesn't drop either after shutting engine off, so if all is good with new cat, I will probably just leave it alone.

[wilmar13]

The only other thing is perhaps the turbo has a problem. It was a brand new IHI OEM turbo but it did seem like the impeller had more play when new than the one I removed. It does not have any issue building boost however. I do hear a whistle under boost which made me think I had leak somewhere, but I can't find anything, and I had gotten accustomed to driving it with a loose wastegate, so it is just as likely I am hearing normal turbo noise. Its also possible there is issue with hosing on turbo that does not show up on static pressure test. IDK,

[Smac770]

Morris,

Logged the IDE01874 this time and it just goes Test off -> Test on -> CatB1 Ok for my engine. IDE01873 was 0 until the CatB1 Ok, at which time it reported 1.4146. So a little better than last run. But I'm still unsure if that's the measured OSC value, or the ratio of the measured OSC value to a marginal OSC value (the marginal value being pulled from a lookup table based on the speculated EGT and the measured air mass). It's the ratio, that if below 1.00 (or .90 depending on the doc), trips the P0420, at least for our engine. I don't know the actual metric for P0420 on your engine. Here's a chart of the event:

Cat OSC test chart.jpg

red and orange are on the left axis, green and blue are on the right axis. The rpm was a steady 2050+-50. You can see the lambda doing its cycles around 1.0 and the rear voltage holding at 650mV. I was at a steady 55mph or so cruise. I added some throttle, probably an uphill in the road, at 289 and you can see the lambda go down to .95 (rich) and the rear voltage jump up towards 700mV (richer). As the throttle is now steady, the lambda settles back to 1.0 cycles and the rear voltage comes back down.

Then the ECM invokes the cat test starting with a rich mix. The lambda result goes rich and is held there, the rear voltage goes rich. The O2 is being pulled from the cat. Then the test flips the to the lean mix. We see the lambda result reflects the transition from rich mix to lean mix. The rear voltage drops off. I assume the test is about the time it takes for the rear voltage to swing back to 450mV neutral reading.

My IDE01873 result was 1.4146, and that is about the time in seconds from when the lambda swung from rich to lean to when the rear voltage approached 4xxmV and the test switched from On to CatB1 Ok. But I don't have any information to support that speculation. The numbers simply align in that manner, this time.

[morris39]

Morris,

Logged the IDE01874 this time and it just goes Test off -> Test on -> CatB1 Ok for my engine. IDE01873 was 0 until the CatB1 Ok, at which time it reported 1.4146. So a little better than last run. But I'm still unsure if that's the measured OSC value, or the ratio of the measured OSC value to a marginal OSC value (the marginal value being pulled from a lookup table based on the speculated EGT and the measured air mass). It's the ratio, that if below 1.00 (or .90 depending on the doc), trips the P0420, at least for our engine. I don't know the actual metric for P0420 on your engine. Here's a chart of the event:

Cat OSC test chart.jpg

red and orange are on the left axis, green and blue are on the right axis. The rpm was a steady 2050+-50. You can see the lambda doing its cycles around 1.0 and the rear voltage holding at 650mV. I was at a steady 55mph or so cruise. I added some throttle, probably an uphill in the road, at 289 and you can see the lambda go down to .95 (rich) and the rear voltage jump up towards 700mV (richer). As the throttle is now steady, the lambda settles back to 1.0 cycles and the rear voltage comes back down.

Then the ECM invokes the cat test starting with a rich mix. The lambda result goes rich and is held there, the rear voltage goes rich. The O2 is being pulled from the cat. Then the test flips the to the lean mix. We see the lambda result reflects the transition from rich mix to lean mix. The rear voltage drops off. I assume the test is about the time it takes for the rear voltage to swing back to 450mV neutral reading.

My IDE01873 result was 1.4146, and that is about the time in seconds from when the lambda swung from rich to lean to when the rear voltage approached 4xxmV and the test switched from On to CatB1 Ok. But I don't have any information to support that speculation. The numbers simply align in that manner, this time.

Thanks, that's interesting. From your experience it may be that my test was not long enough, just 1.5 min and so a new test did not initiate. OBDII log (partial) says " Conditions completion count / Specified conditions encountered count
Catalyst monitor bank 1 : 743 / 1693". So in the past, the test has been run automatically and successfully, car passes readiness test. I will redo test. I'm just curious as there are no problems not even pending DTCs/.

[morris39]

Smac
Forgot to mention that the number of suitable conditions for the auto cat test was about 0.3 X No of total drive cycles and of those 0.4 completed per Category 8. That is for about 36K miles driven. So the test drive is not random.
Ross-Tech did not have specific suggestions re values obtained and lack of test initiation except to try Basic Settings. I am a little hesitant as a typo might screw things royally

[Smac770]

Yeah, it takes about 5 minutes from a cold start for the test to kick off, little under 3 minutes for a warm start. And though the test is attempted every engine run, it doesn't seem to complete every time.
The return trip for that one I created the chart from (which was a cold start run to get lunch), test on occurred about 2.5 minutes into the drive. It held that status for 2.5 seconds, and then when to test off, not Cat Ok. And my result measurement stayed 0. I see it drove the lambda down to .95 for 2 seconds, but then it never bothered with the second half of the test. For what reason, I have no idea. It tried test on again 7 seconds later, but that only lasted maybe 1 second. It didn't try the test again for the rest of the 15 minute drive. So it doesn't necessarily accomplish a result every drive, apparently.

I assume you're not getting the catalytic converter readiness bit to activate? I don't know what activities compile into that bit status.

https://www.ross-tech.com/vcds/tour/readiness.php
By Basic Settings, I assume they are referring to the last bullet item on that page. Where you can run from Basic Settings a test sequence for the various readiness tests. Though it won't do you any good for the ones that require a long steady 45-55 mph cruise, etc. I think that's usually the EVAP one.

[wilmar13]

FWIW, My IDE01873 (test of catalytic converter: measured value) is .8818 with the 100k mi cat. I got new OEM cat delivered today and will monitor that when I install it.

[silver_tt]

Hey wilmar -- just curious if there was any update on this? I'm assuming you got the new OE cat, installed it, and life is peachy again? I know the aftermarket cat you had was still coding so just curious where you landed with this after now trying the OE cat instead.....

[wilmar13]

Hey wilmar -- just curious if there was any update on this? I'm assuming you got the new OE cat, installed it, and life is peachy again? I know the aftermarket cat you had was still coding so just curious where you landed with this after now trying the OE cat instead.....

Yeah, all good. I had IDE01873 value of 1.45 when it was installed and only a test drive... I intended to log pre and post cat O2 sensors to compare to before, but haven't gotten around to it... the squeaky wheel gets the grease, and other things were squeaking. My upper cam timing cover was leaking, and so was the orientation sensor. I found a pin hole in the PCV return to turbo line that was formed by rubbing against the heat shield. So replaced that, new PCV valve just because, and line to intake manifold from PCV. I paid closer attention and there were lots of dings in the mating surface of PCV on aluminum block, I filed them down. And helicoiled the remaining holes the tardbot who worked on the car didn't bother to do when I complained about the warranty work and stripping all the PCV screw holes. Don't think that was a problem as there was no oil leaks. But man there were some major deep gouges in the sealing surface for upper timing cover. Audi Burbank has (or had) a real issue with techs that just don't GAF. I don't even know how they hit it so hard, or with what (the engine was out to replace pistons under warranty too), but I just can't believe they didn't file down the mushroomed material. There were three different places it was leaking and it was where there was major damage to the aluminum mating surface. That may have been part of the problem. I repaired it, replaced all gaskets and used silicone where the problem areas were... no oil leaks, runs fine, and mpg improved a bit, but that is just as likely due to no longer using winter gas and warmer temps.

[silver_tt]

Awesome, thank you for following up with the solution. No worries about the squeaky wheel gets the grease, totally understand and I was mostly just wanting to be sure I didn't guide you to a $1,000 mistake :) I wouldn't recommend to someone anything I wouldn't do myself though and I was pretty sure your cat was shot. Good to know the 48 state thing is bullshit and it will still code even if they say it won't.

That's annoying af about the PCV on the aluminum head, honestly I can't believe half the stuff I read and see on youTube -- I would have helicoiled it just like you did if I had that problem.

My upper timing cover was leaking a bit on my car. That gasket is rubber as you know and hardens over a few years. I replaced it and all gaskets when I did the timing chain and now it's bone dry. You can buy a bottle of Gunk for $6 on Amazon and I used this to clean it all off. You spray it on, wait 15 minutes, then gently hose it off. Took off all the oil and dirt and now it looks like new (or close enough) and will help me see any future leaking right away.

[Smac770]

1.45 is back in the range I see on my car.

I just noticed your comments about part numbers and reman. Audi's part number system uses revision letters from A through T, except I and O. And they can be doubled up, like AK and such. U and X have specific meanings, U = used and X = remanufactured. They do not sell new cats for the B8 2.0T anymore, only remans. Thus you see the X at the end of the original part number. Audi actually has reman parts for a lot of stuff, particularly cylinder heads and cats. An example of a used part, https://parts.audiusa.com/p/Audi__/C...0254250PU.html I have no clue what that cat is for.