Cold start white smoke

I have only had this car since the summer, and lately we have been having some nights that fall below about 60 deg F. On these mornings it smokes profusely and dies after running for a few seconds, and will not restart, tho it kind of coughs along at idle once in a while for a few seconds. If I give it any accelerator pedal it quits. When I get home from work later in the day, when it has warmed up, it will start fine and run great.

It is throwing a P1562, a P0380 and a P1268 after these events (three times so far). It has had a couple of other ongoing issues, that I don't think are related, one, a P0605 that keeps coming on at low revs, and a P1550 that has stopped recurring, but which came on at high revs when I would have expected the wastegate to open, and which caused a sharp power drop.

The car looked brand new with 136,000 miles on it when I bought it, so I suspect the previous owner rarely got revs high enough to open the wastegate. I think I have exercized it enough that the 1550 is history, as I have not seen it in a month or two, and its likely that the P0605 is the ECM vacuum line.

I'm waiting for my VAG COM gear to arrive, and so only have my OBD scanner, but I have a business trip this weekend and really need some sort of workaround so that I can get it to start in the morning. Does the ECM rack the timing based on temp, either coolant or air ?

My manual talks about a "cold start injector" but what I have read here suggests that it is only a timing advance, is there anything else going on during cold start? I live in So Calif, so when I say cold start I do not mean COLD start. My older VW diesels have a manual advance knob that I never use as it is only necessary at sub zero temps.

Thanks for all the advise you all have posted. This place is an amazing resource that has already taught me a ton.

THANK YOU!

For starters, can we get model/year info?

You only have a wastegate if it is an AHU engine (<99.5 Jetta/Golf). Otherwise it is a VNT actuator that controls boost levels.

White smoke and rough startup indicates incorrect IP timing. Do you have any record of when the timing belt was done or who did it?

Welcome to Fred's, stick around and read all you can. There's alot of info here, you've come to the right place.

Car is a 98 Jetta

Check the glow plugs visually and with a multimeter (one code)
Check the fuel supply to the injection pump (another). Clear fuel hose is really handy. How old is the fuel filter and does the pump appear to be wet anywhere?

P1268 I can't find right now

For future reference, please translate your codes, so the people trying to help don't have to.

P1562 quantity adjuster upper limit exceeded
P0380 glow plugs

I also cannot find a translation to P1268 aside for some Ford stuff that cannot apply here.

White smoke often means retarded timing and unburned fuel.

Given that it has trouble running in the cold, I suspect that the pump cannot advance the timing enough when the engine is cold.

The QA code may be because the pump is trying to compensate for low idle by injecting more fuel. If the car wants more fuel than it can deliver, it will throw a QA code.

Please check the timing with VCDS and report back.

The N108 'cold start valve' - a misnomer is the solenoid that controls case pressure to the hydraulic timing advance mechanism. It is active at all operating temps, working to provide requested injection timing.

1998 Ahu

Sorry about that, it is a 1998 AHU Jetta.

If the timing was off, it should not only occur at low temps. It runs fine all the time, with no smoke and quick starts, this has only occurred on three cold mornings and each time it has started fine later in the day. Given how warm cold is here, I'm talking like 55 deg, I doubt that inadequate advance is the problem.

When this first happened a couple of weeks ago, I threw a set of glowplugs in it, so that shouldn't be an issue. I must have mis-remembered the 1268 code, it was the first time I had seen it. I'll recheck when i get home and post the correct one.

As to the other codes, I have read so many posts about the 0605 and the 1562, I assumed those with enough knowledge to be helpful would be familiar with them. I will list more completely in the future.

If the timing was off, it should not only occur at low temps.

Click to expand...

Actually, it will ONLY be noticable at low temperatures. That's because the computer will compensate once the engine starts. Before that it is at the mercy of the physical position, or 'basic' timing, of the pump.

Try one other thing. when it is cold, disconnect the coolant temp sensor and see if it starts easier.

I will try the temp sender disconnect, great idea. I do think tho that if the pumps basic timing was that far off, it would be a gradient, so that at 65 deg it would be a little funky, barely run at 60 and at 50 deg not start. Probably those would not even actually occur until much lower temps. This is so acute, with no sign of anything until it seems to cross a threshold, I can't help thinking it is a specific cold start enrichment or advance malfunction.

But my manual claims that a P0605, which reads as a Read Only Memory fault, requires ECM replacement, so it may all be related to a faulty computer after all.

I've noticed other members who have taken all the steps, replaced the k75 and all the hoses and even the ECM and are still having p0605 problems. I guess that's why Jose at Euro Tech has decided to stop working on TDIs. One of the things I have loved about my older VW diesels is how utterly simple and bullet proof they are. When I was wrenching for a living, the last thing I wanted to do was work on my own car, and these things would run for years with nothing but oil and filters. But my TDI is soooo much quicker than my Caddy and Golf, its a trade off I guess. Hopefully when I'm Vagged I'll feel les discouraged.

I've gathered that the hose inside the ECU or bad MAP sensor can cause p0605.

check out http://forums.tdiclub.com/showthread.php?t=88039

I happen to have the translation of P0605 memorized - ECU memory fault. Others do not. Again, please translate any codes mentioned.

It helps me, and people like me, who understand the finer details of how these cars work, because we don't have to look up the codes. It helps the OP, because if I don't have to spend a bunch of time deciphering the codes, I'm more likely to help with the problem. Finally, it helps 3rd parties in the future, because they can search either the code, or the symptom, determine the meaning, read the trouble shooting process, and fix their cars themselves.

Onto other things:

The timing of the VE cars is electrically controlled, but hydraulically actuated; and requires case pressure to achieve advance. The starter cranking over the car over at ~250 RPM isn't producing enough case pressure to actuate the advance mechanism much, even if the ECU is requesting full advance. That means that the starting conditions of the car are mostly based on the pump's position relative to the crankshaft. (the static timing)

As a for instance; lets say that when cold (~30F), the injection needs to be at least 1* BTDC to start; warm (~70F) at least -1* BTDC to start. Lets also say the pump is set such that with no case pressure, (fully extended timing spring) the injection event occurs at -3* BTDC. With the ECU requesting full advance, but case pressure only as high as the starter can develop, (~3*of advance) the injection event occurs at 0* BTDC.

This possible case results in the car starting when engine temp is ~70F, but not at ~30F.

Why does it run fine then after it starts? Because once the engine is turning at sufficient speed ~2000 RPM case pressure is sufficient to produce up to ~12* advance when requested. 12-3 = 9, so that means the car can achieve up to 9* advance no problem. Under these conditions EGTs at WOT are a little high, but everything else works fine.

Regarding gradients; the amount of advance required to initiate ignition follows a gradient. The flash POINT of the fuel is not.

So, to reiterate my original post on this thread;
Please translate any codes refreneced while posting,
and check the timing in basic settings with VCDS and report back.

P1248

P1248 was the other code I got that I misreported as a 1268. It is faulty cold start injector N108, or needle lift sensor G80.

I reported in my initial post that it "dies after running for a few seconds, and will not restart, tho it kind of coughs along at idle once in a while for a few seconds. If I give it any accelerator pedal it quits."

Therefore the case pressure (is that oil or fuel?) will be adequate to operate the advance once it starts, which means it should then run ok, which, as I said it does not.

The flash point of the fuel is a gradient proportional to temperature which is exactly why this cold start mechanism is in place. I stand by the idea that for a the car to start and belch smoke, then die at 60 deg, but start and run perfect at 65 deg, it cannot be a basic timing issue.

There will be some gradient where one goes from hard starting to no starting across some temperature gradient, at any given timing (maybe, but very unlikely to be between 60 and 65 deg, more like 30-40 or 34-37 or whatever) but there are enough variables, including four cylinders with slightly different compression, four chambers with differing deposits and four injectors with slightly different patterns, that there cannot not be a bright line of a flash-don't flash at a single precise temperature point for the whole engine.

I have been running biodiesel almost exclusively since I got it, and in my manual I do notice o-rings on either side of a "strainer" in front of the cold start injector. I wonder if these might be neoprene and deteriorated. What does this N108 device do and how does it use a strainer? Other posts have commented that it is really just an advance actuation solenoid, but there is some fuel quantity adjustment going on in this pump as well.

Today, with a coolant temp reported as 70 deg, it started right up and ran great. I cleared the codes and drove my 30 mile combined mountian road, stop and go and freeway commute with none of them recurring. In the past I have had that ECM fault come back evey few miles during a low load, low revs period. Hmm. Most of my problems seem to be fixing themselves.

Can I still check timing in these pumps with my dial indicator?

nesdon said:

P1248 was the other code

Click to expand...

Here is the translation for P1248. Start of Injection Timing Regulation: Control Deviation.

The car is saying "Hi - I'm a 98 Jetta; and I can't time myself"

Nesdon - since you know FOR SURE that it's not the timing, I guess that makes finding what the problem is a lot more difficult. Why are you so sure of that again? Oh right...you decided it couldn't be the timing.

Nesdon, everyone else that has responded to the thread has alluded to the same possible cause that I have. I admire your perseverance. You are truly a MAVERICK! I'll leave you to it now, since I clearly have nothing constructive to add to this conversation.

Could it be the timing?

I suspect it is the timing

I just don't think it is a problem with a mistimed pump or the basic timing, I think it is a problem with the cold start timing control. You made a rational and resonable argument to back up your diagnosis of a mistimed pump and an explanation of why my argument about gradients was false.

I made a rational assessment of the flaws in that argument, mostly based on the internal contradictions in it due to the fact that my obtuse and disjointed writing style prevented you from noticing what I had said. That is that the car started and ran, before it began to smoke and die, then ran roughly a few times before fully failing to start. And finally that it started easily and ran perfectly at an ambient temperature 5 degrees above the one where it failed to continue running.

I am a scientist, I am not sure of anything until it is proven. I am here to engage in a diagnostic process with other mechanics more familiar with these particular systems than myself (as you clearly are). In fact I could sure be wrong as I don't know much about these pumps, but isn't this a picture of you holding that very N108 "cold start injector"?



I really am sorry if I gave the impression of being disrespectful or arrogant. It was not my intention. I truly appreciate your input and can understand, having read many posts here, how frustrating it must be when these know-nothing newbies come in here and ignor everyone's obviously sage advice.

Please accept my apologies if I was rude or disrespectful. Attitudes, sarchasm and such are difficult to read online. I am a huge proponent of dousing even any suggestion of flames, hope I have succeeded in that.

Once I get my Vag Com I will check the timing profile and the temp sender values and report back. I noticed that the manual mentions fuel and air and coolant temp senders in the diagnositc path for this cold start system. I'll have to look more closely to see where they are and how they are involved.

Thanks for the input.

It's best to start with the basics. It would help tremendously if you were to get the basic timing information because it helps when trying to trace the pathways. It's even harder at a distance.

QA limit stop attained in my experience is a sign of fuel starvation. The QA will cover the dump port more and more trying to inject fuel all the way to full stroke until physical limitations are reached. If you don't get airlocked injectors then your fuel supply might be ok but that only means you aren't getting air in the system. It can also happen if the QA is misaligned after dissasembly from the main body or a "hammer mod" in the wrong direction/too far, putting the quantity adjust solenoid range out of physical limits, it is stuck, or if the engine can't start (searching for idle fuel quantity).

The engine won't start if these conditions are met:

Insufficient cylinder temperature
- low compression
- insufficient glow in cold (<40F)
Insufficient fuel quantity
- Blocked fuel filter
- Air leak in fuel system
- Fuel too cold (cloud/gel)
Incorrect injection timing
- Basic setting incorrect
Insufficient air mass
- Blocked intake pipe/manifold

Symptoms of basic timing too late are hard starts, especially in low temperature, excessive smoke, and reduced performance.

Causes of timing control not attained are
- Basic timing out of range
- Physical blockage in advance mechanism

I've seen the N108 sieve get metal shavings on it. You want that to not be your problem. If you get to that point, it's time for a pump rebuild (starts at $600).

All being said, based upon the codes seen, the first step is to verify basic timing, and then get more invasive. I see no way to get N108 removed without getting the pump off of the engine, unless the mounting is radically different on older engines.

update - vag com data

I finally got my HEX CAN VAG COM working today. It is not a very intuitve interface, but I imagine I will figure it out. I went into an advanced menu in the engine module and found a long check list with many repetitive check boxes. A couple of them returned BTDC numbers.

I'm not sure if the 0.0 BTDC numbers reported before I started the car were actually the inital timing or an empty register, but once it started it was running around 1-2 deg BTDC and returned a number as low as .7 deg BTDC. The duplicate data being reported seems slightly different, maybe from processing time, as they move together and stay very close.

I'm fairly certain that the initial timing is fine. I struggled for the last couple of weeks trying to deal with the loss of prime, as per my other posts, and in the course of that replaced almost all the hoses. The little one in the ECU had an obvious crack. I didn't do the red and blue down to the wastegate cause they are so cute, (anybody have a source for colored hose?) and cause the ends on the turbo are impossible to see from above.

So far, a 100 miles of driving with no DTC. But as to my cold start, most of you have probably noticed that the weather in so cal has been very warm (thankfully the brush-filled canyon I live in was spared), and I haven't seen an coolant temp below 60 since these previous posts, so I'm still unsure of the cold start failure. I'm going to start logging the prodigious data the VAG reports and hope I can zero in on the fault.

For now, I am giddy to be back in my lil TDI. I've had to suffer through couple of weeks in my bouncy, AC-less, CD-less 17mpg 97 Nissan 4x4 ranchtruck cause my caddy has a blown half shaft and I haven't wanted to face a couple hours on my back in the dirt up to my elbows in moly, and anyway wanted to spend my descretionary wrench time on the TDI.

I'll report back.

to view the initial timing with VCDS;

open the 'engine controller'
-> measuring blocks -> basic settings;
select group 000 -> go;
a TDI timing button will appear
-> TDI timing

a window will appear with a graph. Using the drop down window, select the engine you have in your car. It'll be ~7th down from the top. Technically, you're on spec if the apex of the horizontal and vertical yellow lines is between the red and green diagonal lines.

Best mileage will be around the blue, perhaps slightly below.

If you're modded (chipped and nozzled) it's best to be just below the green, as this will provide the most advance and lowest EGTs at WOT.

Adjusting the timing is accomplished by rotating the pump by loosening it's 4 mounting fasteners. Looking from the front of the motor (passenger's side of the car) rotating the pump counter-clockwise is more advanced.

The timing is squirrely on the A3/B4 chassis, and will be pegged off the graph advanced when the timing is actually quite retarded.