Hi all,

 

I could really use some help with a problem that has just appeared from nowhere on a 2003 110 TDI (ASV), 5-speed manual that was previously driving without fault. 

 

Put simply, Traction Control light on dash is on more or less permanently and car runs in Limp Mode while TC light is on. 

 

Symptoms in detail:

 

• At Ignition On, dash lights run through check sequence (as normal), then all but Handbrake and Battery lights go out (as normal), but then the Traction Control (TC) light comes back on approx. 5-10 secs later and stays on (I read this as Traction Control setting itself to OFF due to an issue 'somewhere'). 
• Starts on the button and idles fine, but has absolutely no boost whatsoever (no acceleration to point of holding up traffic!).  Anything other than very gentle acceleration results in detectable (but mild) misfire and fair bit of smoke (unburnt fuel type smoke - you can smell it!) - i.e. it's in Limp Mode good and proper!

The following DTCs have been read and repeatedly erased but just come straight back (read with Autel VAG405 Scan Tool - I don't have access to VCDS):

 

'Engine' Codes

• 17946:  Fuel Shutoff Solenoid (N109): Open or Short to Gnd: P1538
• 19561:  Valve for Intake Manifold Flap (N239): Open or Short to Gnd: P3105
• 18080:  Coolant Fan Control 1: Open or Short to Gnd: P1672
• 17957:  Boost Pressure Regulation Valve (N75): Open or Short to Gnd: P1549
• 17849:  EGR Valve (N18): Open or Short to Gnd: P1441
• 18027:  Glow Plug Relay (J52): Open or Short to Gnd: P1619

 

'ABS' Code

• 01314:  Engine Control Module Sporadic 013 - Check DTC Memory (Note no specific ABS or TC related DTCs logged)

Despite their persistence, the above Engine DTCs all seem questionable - e.g. the engine starts fine and switches off, (surely?) proving nothing amiss with the Fuel Shutoff Solenoid, nor its feed.  The Glow Plugs are getting power on initial cold start-up.  The cooling fans both run without any problems (tested by putting AC on).  And the feeds for the N239 and N18 valves check out (as detailed below).   So, not sure why these DTCs are constantly being logged, even when erased then immediately re-read again.  

 

Investigations so far have deduced the following: 

• Reason for no boost is Boost Solenoid (N75) valve is not letting vacuum through to VNT actuator (measured with Vac gauge), so vanes are remaining in 'rest' position.  Plenty of vacuum definitely exists at 'VAC'  inlet to N75, but no vacuum is getting to 'OUT' port to feed the VNT actuator. 
• I have swapped out N75 valve for a known good one taken off another identical car (also an ASV 110 TDI) with no faults - No change to fault symptom.  Moreover, the 'suspect' N75 valve works perfectly fine on the other car, confirming it to be good.
• The N75 valve is getting a 12V (nominal) feed from Main Relay 109 and the other pin on its connector has good continuity (and zero resistance) back to Pin 62 on the ECU, but the ECU doesn't seem to be 'driving' it at all.
• The EGR solenoid (N18) valve is definitely being 'driven' (its OUT port has measurable vacuum, which increases with revs), as is the intake manifold flap (N239) valve (Anti-shudder valve?).  Both of these are also powered from (and are getting) the same 12V feed as that for the N75 valve, which comes from Main Relay 109.
• Some forums suggest TC light on may be due to problem with the MAF sensor.  This has been discounted, as item has been tried on the other (good) ASV 110 TDI and works without fault (known good MAF sensor from other TDI also tried - No change to fault condition, so it's not that either).  12V feed on Pin 2 of MAF connector also comes from Main Relay 109 and is fine.  All other MAF plug pins have continuity back to ECU too.
• Some forums suggest Brake light switch may cause issue with TC light on.  Brake light switch also confirmed not to be causing problem (again, known good one tried - no change). 
• Some forums talk about TC light problem being down to issue with alternator wiring harness (due to iffy wiring at 4-pin connector next to Starter Motor).  However, this has also been checked and resistance measured, and found to be sound.  On this theme, I considered whether alternator itself might be possible cause (at DFM connection), but alternator seems to be working fine (i.e. no battery light on dash and stable 14.5V output).
• Every engine bay earth I can locate has been taken apart, cleaned, reassembled and resistance checked (with ref right back to battery -ve terminal), including the one up on the bulkhead, by the Engine Mngt ECU. 

So, after all that, this now has me stumped, as nothing is making sense.  Is the TC light indication an actual issue with the TC system (note, ABS light is NOT on, except for during initial Ignition On check), that in turn is causing the Engine ECU to not drive the N75 valve as part of the Limp Mode settings?  Or is the TC light a symptom (i.e. acting as a tell-tale) for an engine management issue?

 

Any help, tips, pointers gratefully received.

Regarding code from ABS module - it is just reminder that you have engine problems. That means you have no issues with breaking system. When you fix and remove engine problems those codes from ABS won`t appear again once you clear them.

Edited 14 October, 20169 yr by Arminator

Hi Arminator,

 

Yes, that is as I thought too, given that the ABS light itself does not come on and there are no actual ABS fault codes coming up. 

 

Have rechecked everything to do with the main relay again (I have even been inside it and checked solder joints, cleaned contact faces etc.) and concluded that it is definitely being driven correctly by the ECU and is switching as it should.

 

I am about to take a look at the Boost Sensor and its wiring, theory being that is the device that tells the ECU if there is too much boost and triggers Limp Mode in that situation.  No access to VCDS, so can't check live data.  Will see what can be gleaned from physical checks. 

 

Wish me luck!

Did you check ECU connector (under the windshield tray)?

Do you feel any issues while driving (slow acceleration and so)?

Yep, as stated in my initial post, I have checked/traced the quality of connections from sensors all the way back to the ECU connector (this was one of the first things I did).  Every path from each sensor back to their respective ECU pins has checked out as sound - i.e. good continuity and negligible resistance from end to end, no shorts to earth, etc. And the ECU connectors are completely clean and corrosion free too. 

 

Incidentally, everything with the Boost Sensor also checked out fine today too - i.e. sound connections from Boost Sensor back to ECU connector pins. I wasn't sure what readings to expect on the Boost Sensor itself due to lack of info, but randomly checking static resistance between its connector pins provide relatively sensible readings in the order of 1.5 - 5 kohms (i.e. as you would reasonably expect for any typical sensor).  I tried the Boost Sensor on my other working 110 TDI (just with Ignition ON, as I couldn't actually install in the Intercooler) and no faults showed - i.e. no Traction Control light on dash.  So I have concluded from this that the Boost Sensor is NOT causing the fault after all.

 

As for drivability issues, yes, again as stated in my initial post:

 

• Starts on the button and idles fine, but has absolutely no boost whatsoever (no acceleration to point of holding up traffic!).  Anything other than very gentle acceleration results in detectable (but mild) misfire and fair bit of smoke (unburnt fuel type smoke - you can smell it!) - i.e. it's in Limp Mode good and proper!

 

Just with the engine running (i.e. car static, not actually out on the road driving), if I raise the revs, when it gets to approx. 2000-2500 RPM it starts to misfire (slightly) and the engine visibly shakes (I can also feel this when I actually drive it).  And as previously stated, you can smell the unburnt diesel from the exhaust (i.e. the fuelling is completely up the creek!).  

 

So despite still starting perfectly and idling smoothly, she really isn't happy revving!  The shaking, misfiring and lack of power outwardly has all the hallmarks of fuel starvation, but I am putting this down to the ECU going straight to Limp Mode all of the time (as indicated by the Traction Control light coming on all of the time, even just with ignition on!) and backing off the fuelling.  On one short occasion when I managed to start it mid-journey without the Traction Control light coming on, the car felt/drove fine.  So, whatever the problem is, it is putting the car in Limp Mode and in doing so it is disabling (turning off) the Traction Control.

You need VCDS.

Fleabay has knock off versions which work perfectly. I'm not condoning piracy but £39 instead of £300 when it will help pinpoint a problem which will need more money throwing at it to fix just makes sense.

Yep, thanks for that mate.  TBH, the need for VCDS was recognised but hasn't been able to be realised yet.  Point taken all the same  :-)

 

With the best of intentions, a combination of working full time, trying to find actual meaningful tech info on this engine management system (even down basics like what relays it has, what they're for, where they are located and how to tell which one is which in the Relay Plate!), checking though all of the bl33din' obvious stuff and trying to simply make sense of the fault codes relative to components and wiring that all seem to be perfectly good - and fulfilling domestic duties between all of that ;-) - has all kinda stood in the way of time available to spend searching out reliable sources for VCDS.  Having never seen/used VCDS in the flesh before, there would be the small matter of sussing how to drive it too (not excuses, just the way it is).  The 'fleabay' tip is noted with thanks, albeit with slight caution as I've seen plenty of mixed comments about some sources on various internet forums.

 

I'm a time-served electronic and electro-mechanical engineer by trade, with a healthy wealth of experience in maintenance, fault investigation and rectification on a broad range of on and off-platform systems.  So in the absence of VCDS it is hopefully understandable to have approached this with fundamental fault-finding skills by checking for all the basics like failed components (sensors/actuators, etc) and supply/wiring/connection issues first ('cos let's face it, that where a good proportion of fault causes actually lie), subject to availability of meaningful/reliable tech info on the system, that is.  And this approach has served to identify a few engine bay cable/earth issues I wasn't even aware of along the way, some of which had to be addressed to eliminate them from enquiries as much as anything. 

 

'scuse if this sounds like anything approaching a slight rant; it's not intended that way, just current levels of frustration coming to the surface a little :-/.  Any feedback received about this issue is genuinely gratefully received. 

 

As far as latest investigations are concerned, I've just tried disconnecting the ABS pump/ECU to see what happens.  This was in an attempt to rule out the legendary ABS pump pressure sensor fault, as I've read plenty about that one triggering the TC light, so I wondered if that also might have been prompting the Engine related DTCs to appear.  Trying that puts the Traction Control light out (as expected), but what I wanted to do was see what happened if I then erased and re-read the DTCs logged in the Engine Management ECU with the ABS ECU/pump still disconnected.  And lo and behold, the same engine DTCs still come straight back every time (this backs up Arminator's earlier comment, which I agreed with but just needed to try this to satisfy a burning curiosity - Sorry!).  Anyway, unless I'm mistaken and/or my Scan Tool is failing to find something, this is starting to lean more and more towards an Engine ECU fault.  Question:  Would VCDS be able to confirm once and for all if the Engine ECU itself is faulty?

 

Thanks all.  Will see what's out there on the VCDS front.

I think VCDS lite may be all you need, for a 2003 Octy?   If so you just need a cheap cable and a free download. Worst case if that still doesn't do what you need, register the software for $99 and it probably will.  Other mk1 Octy owners should be able to verify or falsify this.  Trying and failing to remember if I've plugged my workmate's car (2004 ASV) into mine, I don't think I have.

 

I've got various documentation (copyrighted, via erWin Skoda) for your engine - cos of that workmate - so if there are any specific questions about wiring for example, I may be able to help.   

Hi Wino,

 

Thanks very much for your reply.  That's useful info.  I'm just starting to look into this now.  As the great saying goes, every day's a school day! 

 

I take it the cable you are referring will be broadly referred to as a 'VCDS cable'?  In other words, just any old OBD2-to-USB cable won't do the job due to the wrong interface being built into it(?), so I'm assuming it has to be a 'VCDS' cable, cheap or otherwise? Sorry, just making sure I go searching out and sourcing the right thing here.

 

Ref the VCDS-Lite free download you refer to, I'm just looking at the feature list and capability info online now and it looks like it should hopefully read DTCs on my car OK, although this is based on the fact that it says that "VCDS-Lite is...for older (pre-2006) VW-Audi Group cars that do not require a direct CAN diagnostic connection".  It says "Won't work with CAN-Bus vehicles", which appears to apply to all Golf platform cars Mk5 on.  Mine is a 2003 model (basically a Golf Mk4 platform under the skin) so it certainly fits the 'pre-2006' mould, even though it actually has CAN-bus paraphernalia spread around the car itself (e.g. I know for sure that it has CAN-bus modules in the doors to cover the Central Locking and Electric Window functions, for instance).

 

Oh, and thanks very much for the offer of help/info in the event of wiring related queries; I'll bear that one in mind for sure.  Ironically, despite starting with very little info just a few days back, I've now managed to build up a fairly meaningful picture of what is in the system, where it is on the car, and how it is all connected, purely by literally getting amongst it all with nothing more than a methodical, logical approach, a DMM to trace everything from end to end, and a lot of Tea and patience while I think it all through and manage to make sense of it!

 

As for the fault itself, I dug the ECU out from under the windscreen this evening (what a ridiculous place for it BTW!) to see if there was any evidence of water ingress.  Immediate impression once out was that the outer case has definitely suffered over the years for sure.  The ECU is mounted at an angle, and as such, the area by one of the cover screws at the lower end (where water would naturally tend to settle/hang) was suggesting that all was not too clever on the water ingress front.  Sure enough, that screw was the only one that wouldn't yield and despite being really careful (i.e. by gently applying force in both directions to 'encourage' it) it sheared off as it had fused itself firmly in with corrosion. Steel screws going into an aluminium case - Say no more!  Once inside, there is clear evidence of water ingress in that same corner - Not too much and not too severe, but it's there all the same, so it looks like I am probably dealing with an issue with the ECU itself after all. 

 

So back to my earlier question:  Would VCDS be able to confirm once and for all if the Engine ECU itself is definitely faulty, or would that be down to interpretation of any further DTCs it may manage to unearth?  Or am I going to have to get the ECU sent away for testing to get that level of diagnosis?  My gut suspects the latter, but only because I have no notion of the exact capabilities/limitations of VCDS on this front.  I intend to give VCDS-Lite a try, but figured it worth asking this question anyway.

 

Cheers once again.

I think VCDS would only indicate an ECU fault if it couldn't communicate with it. Not too sure on that though.

I think you've very likely found the problem. Can you put the casing back together with a silica gel sachet inside and see if things improve?

FwIW I will show you the lead I use in a bit, and check the diagnostic wiring of your car to see if it looks compatible. I think the issue with CAN is that it is the only diagnostic protocol on later cars, where earlier ones used K-line; which is what the 'cheap cable' can deal with.

Edited 17 October, 20169 yr by Wino

The wiring for your car does look like it includes the K line, on pin 7 of the diagnostic conn, so this (what I have) should work with it as far as I can tell. Cheaper versions of the same thing can be had, but this is a reputable company, so I bought with justified confidence.

 

Edit: Will work, just tried it on my workmate's 2004 car.

 

May not be necessary to solve this problem, but might be handy in future.

 

I think your diagnostics make a strong case for the problem being with the ECU, as you've tested just about every other possibility methodically.  I just hope that drying it out - and keeping it dry - resolves things without expense.

Edited 17 October, 20169 yr by Wino

Hi Wino,

 

Apologies for delayed response; I'm not long in from work and completion of domestic obligatories!.  Thanks for the replies, some good stuff you've come back with there, and that cable looks like a good'un for sure (and from a very reputable source too, which is a big bonus!).

 

Your feedback about what VCDS will/won't show regarding the ECU is in line with my gut suspicions - i.e. No comms with ECU would confirm a completely failed unit, but ability to communicate with it and get fault codes to display may require a deeper understanding of what is going on within the ECU (or not as the case may be!), to then interpret/decide whether it points to it having failed or not.    

 

Ref the evidence of water ingress, I'm afraid it is exactly that - just evidence of past misdemeanours.  The ECU was actually completely bone dry when I opened it up, but I could immediately see the tell-tale white, crusty deposits on the 'wrong' side of the seal, with a snail-trail where water has clearly been sitting in the well at one side of the lower aluminium cover and has lightly 'burnt' into its surface finish.  The visible deposits have leached onto the PCB and have fouled part of the earthing/bonding strip around the edge of the PCB, but don't visibly appear to have reached as far as any of the actual on-board components.  However, my view is that if there has been any water in there (no matter how small the amount), even if it does not appear to have physically reached components, the enclosed environment within the whole of the ECU will have been subject to a significant increase in humidity and probably condensation repeatedly forming throughout, especially on warm days with the sun beating down on the car.  And there's no telling how long this has been going on for, either.  One thing is for sure, though: it ain't going to have done it any good and given the symptoms I have, my electronic engineering head fears the damage is most likely already done on this basis.

 

It has crossed my mind that the 'dirty' area of bonding between the PCB and the case may be creating a 'noisy' RFI source within the ECU (RFI really can cause chaos in electronic systems sometimes!).  I'm tempted to try cleaning up the offending area, putting the ECU back together and seeing what happens.  Let's face it; I've nothing to lose by trying.

 

I still intend to get hold of VCDS-Lite in any event, as that just makes plain sense given recent events and what I now know about it (it's like everything; if you don't know what you're missing and have never needed it, you don't go looking for it and don't miss it either).

 

Thanks again for all your help mate - Much appreciated.

You're most welcome.

Hi All,

 

Slightly delayed update - I've been busy over the weekend very carefully drilling out the remains of sheared screws (1 in the ECU case and 2 in the top of the intercooler, for the Boost Sensor) and then (successfully) reinstating an original thread in the ECU case.  On top of that, I wanted to be absolutely sure of the success of latest developments before declaring (I didn't want to go counting my chickens too soon!).

 

In a word, it appears to be sorted!.  With reference to my last post, my gut feeling about the potential RFI source created by the water ingress related crud on the edge of the PCB and ECU case contact points clearly had some mileage in it. I carefully cleaned up all the deposits on the PCB, and reinstated a clean contact surface on the edge of the ECU lid to enable it to make good contact with the bonding strip on the PCB.  After examining the PCB under a monocle to confirm there were no obvious dry joints, I then put the ECU back together and tried it back in the car....

 

With battery reconnected, first stage was just with Ignition on......No Traction Control light, even after a good minute or so.  Mmm, promising!

 

So I then plugged the fault code reader in (without cranking, just with ignition on),.  Faults still logged since last power-up were erased...then codes were re-read.  No error codes stored.  This was even more promising (fingers were tightly crossed by this point, with mood starting to err towards a mild 'chufty'!)!

 

Then I went for a start-up, with vacuum gauge connected to the OUT port on the Boost Solenoid Valve (N75).  Hey presto, bags of vacuum!  Switched off and restarted a number of times, just to make sure it wasn't a fluke.  Bags of vacuum every time, varying as it should with revs - WINNER!

 

Switched off and then reconnected all pipework, vacuum hoses, et al, to get it into a state as close as possible to enable normal running (car was generally still in a state of disarray due to previous fault-finding related dismantling).  Revs were raised/lowered as appropriate with the aid of my resident helper, while I watched for Turbo vane actuator action, which behaved impeccably. No misfiring/shaking above 2000 RPM, and Turbo was even producing its usual light whistle when releasing the revs.  All good so far...

 

So after then spending the weekend removing the intercooler, sorting its sheared Boost Sensor screws, then refitting it and generally getting everything finally back together late last night, the car was used for work today and has driven perfectly there and back.    

 

Back to the cause of the fault itself, I am working on the following theory: The (missing) signal required from the ECU to drive the N75 Boost Solenoid Valve is a PWM signal.  If that PWM signal was being interfered with by a sufficiently 'noisy' RFI source (definition of 'noise' being any unwanted signal) being generated within the ECU itself, this would account for the issue I was having.  The ECU case provides screening from RFI sources, and it clearly does this by electrically bonding with the PCB itself.  As the bonding between the case and the PCB was not 'clean', it must have been creating an internal RFI source within the ECU itself.  My background is in RF comms, so I am familiar with RFI issues that can be caused by nothing more than poor bonding.

 

And for those wondering how I re-sealed the ECU, the original beads of adhesive sealant were left in place, then with lid fitted but not fully tightened down, the ECU was gently heated with a hair drier to soften the adhesive/sealant and put it into a soft 'tacky' state.  While still warmed up, the lid was then screwed firmly down to spread the adhesive/sealant in sympathy to achieve a sound bond/seal. Job done.

 

So thanks for all the hints and tips along the way.  This one was certainly not a run-of-the-mill fault, but I've certainly learnt a lot about this system in the course of getting to the cause of it.  Every cloud, and all that...

I know this is quite an old thread but I have just been chasing my tail with the same set of issues on a 2004 1.9 TDI 90bhp Octavia. Fortunately I found this thread & have followed Lampman's lead & stripped out the ECU to look for what appeared more & more to be "ghost" faults. There were tiny signs that there may have been some water ingress but really not that much, so I cleaned up the casing & edges of the PCB, put it all back together & hey, presto! the faults are no longer showing. I needed the N75 valve anyway as it was groaning when I switched off the engine, but the rest of the faults have all gone. I wonder how many of us owners have replaced an ECU for this in the past, when all it needed was a little TLC.