2005 Jetta BEW crank no start, G28 code. "solved", picture

[ejb11235]

TLDR; Trying to figure out why my car won't start. ECU is throwing a code about the Engine Speed Sensor, which has been replaced twice. Put an oscilloscope on the G28 engine speed sensor and see something that I don't understand.

My car died on the highway last January and is now crank-no-start. There are a bunch of facts that I will omit for this post. I have a specific question about the G28 engine speed sensor. The car is had been throwing an "implausible signal" code on the camshaft position sensor, but now it is showing 16705 - Engine Speed Sensor (G28) P0321 - 000 - Implausible Signal - MIL ON. My mechanic had the car for a while and replaced the engine speed sensor twice, but it's still showing the code. He finally gave up and gave the car back to me and told me I need a new ECU and to go to the dealer.

I've completely unwrapped the wiring harness that leads from the 60 pin connector on the ECU. I inspected and found no issues. There is continuity from the connector to the lower connector that attaches to the pigtail on the sensor. I have not checked resistance. Unconnected, voltage to the sensor is around 2.5 volts (on the black wire, pin 2). I notice that the shield on the cable does not appear to be connected to ground.

When I put an oscilloscope on the output of the sensor (the brown wire), I see the following ... the signal in grey is the from the self-study guide ... the yellow trace is what the sensor is actually putting out. (2volts/div) ... output amplitude seems a little low but maybe normal ... the engine is cranking not running so the RPMs are a lot lower.

What are the extra glitches in the signal?

I found a youtube video that shows the output from another G28 ... it doesn't have the extra glitches. (note: injector signal and camshaft position sensor signals also shown)

 

[BobnOH]

No idea what any of that means. It sounds like wiring/connects.
See the Special Notes here.

 

[ejb11235]

No idea what any of that means. It sounds like wiring/connects.
See the Special Notes here.

Thanks. I've seen that note, and spent some time the other night trying to track down these TSBs.

When found in a 2004 VW Golf, Jetta, or New Beetle with 1.9l TDI-PD (BEW), check for TSB 2010461 (01-07-45 formerly 97-05-03) or 201463 (01-07-44 formerly 97-05-04) calling for a flash update to the ECU and a secondary ground installation.

Does anybody have them, or know where I can get them (other than the dealer ... which I will go to if I can't them online)

 

[oilhammer]

Someone posted that TSB here, because I gave him the lead on it (and he thanked me in the thread).

You may also look into checking the ground to the BLOCK itself, as sometimes the upper bellhousing bolt where the ground wire attaches is perfectly fine to the bolt, and to the transmission case, so the starter works fine... but the bolt to block ground is suspect (especially in salty regions), and this can cause some really strange happenings with voltage generator signals like the G28 produces.

 

[oilhammer]

Found it:

How to: solve the 04 BEW hard starts and 16705 G28 fault code.

This TSB will fix the 04 BEW hard start issue. I was getting a "16705 - Engine Speed Sensor (G28): Implausible Signal" code almost all the time. I would clear it and it would come right back within a day or so. I was also getting a large cloud of smoke when starting when cold along with a rough...

forums.tdiclub.com

 

[ejb11235]

Someone posted that TSB here, because I gave him the lead on it (and he thanked me in the thread).

You may also look into checking the ground to the BLOCK itself, as sometimes the upper bellhousing bolt where the ground wire attaches is perfectly fine to the bolt, and to the transmission case, so the starter works fine... but the bolt to block ground is suspect (especially in salty regions), and this can cause some really strange happenings with voltage generator signals like the G28 produces.

THANK YOU!!!! Yeah I have been reading about grounds and not quite giving that the attention it deserves. The grounds I have looked at (like near the ECU) have been fine, but I haven't looked at the engine block grounding. And my car is a snow car ... been driving it up to the ski area for years.

 

[ejb11235]

Found it:

How to: solve the 04 BEW hard starts and 16705 G28 fault code.

This TSB will fix the 04 BEW hard start issue. I was getting a "16705 - Engine Speed Sensor (G28): Implausible Signal" code almost all the time. I would clear it and it would come right back within a day or so. I was also getting a large cloud of smoke when starting when cold along with a rough...

forums.tdiclub.com

I think my 2005 is different. It has a shielded cable that has 3 wires in it ... all three go into pins on the ECU connector. But the shield on the cable isn't connected to anything.

 

[wonneber]

I read the how to fix it thread.
Something caught my eye.

"Now separate the wires by pulling them apart. They are just pressed together:"

How can they keep a good connection?
After years of hot & cold, humid & dry I can see how they would corrode and make a bad connection.

 

[turbocharged798]

They are ultrasonically welded together. I didn't know what ultrasonic welding was at the time. Pretty weird pulling apart a harness and seeing the wires just pressed together.

 

[ejb11235]

You may also look into checking the ground to the BLOCK itself, as sometimes the upper bellhousing bolt where the ground wire attaches is perfectly fine to the bolt, and to the transmission case, so the starter works fine... but the bolt to block ground is suspect (especially in salty regions), and this can cause some really strange happenings with voltage generator signals like the G28 produces.

@oilhammer: Are you talking about GND 2?

Screenshot 2020-09-30 103544.png

• ejb11235
• Sep 30, 2020

Ground locations, engine compartment, GND2 highlighted

 

[oilhammer]

Yes, The one coming from the negative battery cable.

 

[ejb11235]

UPDATE: yesterday I cleaned the ground connections under the battery and to the engine block. Everything looked clean to begin with and I didn't expect a change in the behavior. There was no change in the behavior.

I put the oscilloscope on the engine speed sensor again and checked the signals with the camshaft position sensor unplugged and a replacement ECU plugged in. I'm still seeing the glitch. The car still tries to start with the camshaft position sensor plugged in (I can smell combusted diesel) and doesn't try to start at all with the camshaft position sensor unplugged.

Closer inspection of the signal trace reveals that the glitch comes consistently in the same rotational location ... 7 teeth after the two-tooth gap.

So here's what I think is happening:

• the engine speed sensor is putting out a signal that the ECU cannot interpret (the glitch)
• Cranking
  • with the camshaft position sensor connected, the ECU uses that to get the engine started, but then tries to use the engine speed sensor, but the signal is bad
  • with the camshaft position sensor disconnected, the ECU has to rely on the engine speed sensor, but the signal is bad
  • this would explain why the engine tries to start with the camshaft position sensor connected, and does nothing when it's disconnected

Hypothesis: the reluctor is either damaged or there is something on it ... next step is to pull the sensor and inspect with a bore scope. Just like this https://www.picoauto.com/library/case-studies/skoda-octavia-crankshaft-reluctor

2020-10-01 after cleaning grounds, spare ECU, CPS disconnected, 020ms--anotated.png

• ejb11235
• Oct 2, 2020

 

[oilhammer]

And the surface of the block around the hole that the CKP rides in is clean and free of any debris and rust buildup? Meaning, the sensor is indeed spaced correctly from the crank trigger wheel?

I have seen a car (a gasser, but SFIs use the same basic setup as the BEW) that had a "romantic" encounter with a pothole that had resulted in a smashed pan AND a chunk of aluminum from said pan came up and just barely made contact with that wheel. We had put a new pan on, refilled with oil, and the car would not start, or would sometimes sputter, that was it. Found out eventually by doing what you are doing that the signal was not right, and was erratic.

 

[ejb11235]

I haven't pulled the CKP sensor yet, but I will inspect the mating surfaces and the sensor when I do. Fortunately my neighbor has a bore scope that I can borrow (not that they're that expensive), so if I can get my hands on it I should be able to inspect the reluctor wheel this weekend. I have one of those aluminum pans that protects the oil pan ... years ago my mechanic "insisted" I get one when he found out it was a ski car.

 

[ejb11235]

Well, here's the problem ... there's a tooth missing from the engine speed reluctor. This might be an interesting hobby project but I doubt I can justify paying a mechanic to fix it ... don't you have to pull the crankshaft to replace the reluctor?

missing tooth.jpg

• ejb11235
• Oct 3, 2020

tooth missing from the engine speed reluctor ring

 

[JETaah]

Yup... you'd have to pull the trans, and then the crank. I've swapped out crankshafts before without yanking the engine but had to make a fixture to hold the crank while it was supported on a lifting device. Also, I had a lift.
How are you at welding?
BUMMER!

 

[PakProtector]

That is insane. Good work on finding it. Have to say that if you are going to do away with the car, I would try finding a *REALLY GOOD* welder. Having seen some that make magicians and saints look like newborns, it might be worth a bit o' hunting.
cheers,
Douglas

 

[ejb11235]

Thanks guys. It's not the outcome I was hoping for, but at least now I can be in peace, knowing what broke my car.

Looking at the diagnosis path ... my mechanic replaced the G28 sensor twice. And I looked at a bunch of different things before I put the scope on it. In retrospect, the ECU was complaining about an implausible signal on the G28 ... I'm wondering why neither of us put a scope on the output to see if the ECU was actually telling us something important.

I'm actually thinking about an electronic solution ... I don't really need the tooth, I need the signal that the tooth causes. So I'm giving some thought to how I could make a box with some digital signal processing that would replace the glitchy part of the signal with a synthesized signal.

 

[PakProtector]

Oh my goodness...that sounds interesting. While you are contemplating that, try writing yourself the code to adjust your shower water to the right temperature. You are looking for the guy that wins the bet on welding an aluminum can back together...
cheers,
Douglas

 

[ejb11235]

Thanks Douglas. I have been wanting to get more into the hardware/software space after being away from embedded systems for decades, so this is an interesting opening. I have already started working on the basic signal processing algorithms. What a prize if I can figure this out! It's worth keeping the Jetta.

 

[oilhammer]

Yep, that was really the only thing left. Glad you found it, although I was thinking it would have been damaged from an oil pan tooefing, but it looks like it literally just broke a tooth off!

That is certainly a labor-intensive job to fix, unfortunately.

 

[ejb11235]

“When one door closes, another opens; but we often look so long and so regretfully upon the closed door that we do not see the one which has opened for us.” -- Alexander Graham Bell

I have already started working on an electronic fix. This will take a while, but I've been wanting to get some experience with signal processing, analog-to-digital conversion, etc. But this is taking me in exactly the direction I want to go, and what better way than to have a real problem to solve!

Very preliminary work ... just enough for to start understanding some of the basic signal processing problem. A few more sessions and then I'm going to start reaching out to some contacts. Processing rotational position sensor output has to be a well-worn problem.

dsp_2020-10-05.png

• ejb11235
• Oct 5, 2020

 

[DuraBioPwr]

That sucks bad!! I too have been having issues with crank position sensor in my car. P0727 over and over. Replaced sensor twice now. Mine runs fine, but chooses to not start randomly. The latest foray was the cam went out at 272K and sent metal into the crankcase. Didnt let it get that bad before replacing so engine is fine, but the shards of metal stuck to the magnet on the CPS and was enough to interupt the signal to cause no start. Wipe off metal flakes and good to go. Well doing it again so maybe i have metal still getting on it or I need to do the TSB fix.

Are u sure the 'missing tooth' is not supposed to be like that? I think there is a intentional gap on the reluctor ring in order for the ECU to know where the crank is. Maybe get another reluctor ring and mount to the harmonic balancer and do external crank sensing. Just an idea.

 

[DuraBioPwr]

Never mind they dont have 'missing teeth' looked at a pic online. Just a long 'flat' spot.

 

[ejb11235]

@DuraBioPwr nice idea about mounting a new ring externally ... I suppose one could take the harmonic balancer (I haven't actually looked at it so this might not be possible) and machine marks into the edge!

But yes, this is the reluctor ... there are two places where there are no two "missing" teeth. They come somewhat before top dead center, apparently. But my reluctor as an additional spot!

reluctor.jpg

• ejb11235
• Oct 6, 2020

The cool thing is I have developed my algorithm and now I can actually label the positions on the signal...this means I can develop the signal-injection routine and then switch to an engineering approach and start specing things out, fine-tuning the algorithms and the parameters, do timing tests on microcontrollers, design the electronics, etc.

output of my R&D software ... the colored marks are all generated by code using only information available before the event, illustrating the possibility of doing this in realtime

location_detection.png

• ejb11235
• Oct 6, 2020

 

[oilhammer]

I wish you were closer, I'd offer to buy the car and just fix it, LOL.

 

[DuraBioPwr]

Many race engines have external crank position sensing. Nothing new so maybe could be done.

 

[ejb11235]

I wish you were closer, I'd offer to buy the car and just fix it, LOL.

Well there is that ... actually I'm using this problem as a vehicle (no pun intended) to explore some areas of computing I want to learn more about.

 

[luxuryrules]

Someone posted that TSB here, because I gave him the lead on it (and he thanked me in the thread).

You may also look into checking the ground to the BLOCK itself, as sometimes the upper bellhousing bolt where the ground wire attaches is perfectly fine to the bolt, and to the transmission case, so the starter works fine... but the bolt to block ground is suspect (especially in salty regions), and this can cause some really strange happenings with voltage generator signals like the G28 produces.

I am curious about this connection, do you mind elaborating? My BEW 5 speed, it goes like this: a single long bolt, from the block, through the trans bell housing, to the 18mm head. Then there’s a smaller stud, in which the ground cable is looped over, then a smaller nut. What’s the block connection here? I assume it’s the bolt?

 

[oilhammer]

You are overthinking it. The bolt, the crankcase, the gearbox case, are all grounded together.