Warm-start problem of a different kind (SOLVED)

Hello fellow enthusiasts,

Besides the various and well-known WSPs, my A4 (AFN, exact details see signature) has a very special warm start behavior .
This quirk has existed for years, but it tends to be getting worse.
The cold start is, as always, excellent: turn the key and release it after what feels like two or three compression strokes -> the engine starts.

When starting the engine: Turn the key, then release it after two or three compression strokes -> The engine starts, but runs like a 'mess,' making terrible noises and shaking violently.
This condition can be repeatedly triggered when the material is heated (but it's painful to witness).
If you hold the key down for a slightly longer time (about five to six seconds), it works as if nothing happened.

This behavior worsens with increasing engine/engine compartment temperature. After a break during a high-speed highway drive, it almost takes two to three seconds of cranking before the engine starts, otherwise, you might experience the behavior described above.

I haven't found a reference source (for OT or similar) in my documentation. Apparently, the only available speed sensor is the G28, which is supposed to have a reference mark.

Can the G28 cause this problem when it's in a warm state?
Is there an oscilloscope waveform diagram showing what the signal of G28 should look like? The repair manual only provides a resistance value for verification.

Since I own a second MSG and the ESP was replaced three years ago (the fault was already present then), I can rule out these parts as potential sources of the problem.

The error memory shows nothing, even after several of the start-up attempts described above.

Otherwise, my car runs like a dream and has absolutely no technical issues!

Servus Oldi!

If you write "...like a bag of nuts...", does it then maintain a speed of approximately 500 RPM and not respond to the accelerator pedal?
If so, this is a widespread and well-known problem with the engine control unit due to a software error (almost standard equipment).

I wouldn't repeat or provoke that often in your place, because the ZMS will break down from .

The factory solution is: Press the ignition switch for a longer duration!
(No joke!)

Best regards from Munich!

EDIT: Check the idle speed using VCDS. If it can't consistently exceed 250 RPM, it will be easier and more frequent to trigger the issue.

Hi Dieselbear30x,

das war ja eine turboschnelle Antwort. Thank you for this!

Yes, the behavior you described (approximately 500 RPM and no gas intake) matches exactly!

It would be interesting to know why this problem gets worse when the engine is warm!

I will soon disconnect the coolant sensor with the engine warm (but definitely not in a supermarket parking lot , that would look like a broken-down Audi...), and check if the starting behavior is then similar to when the engine is cold.
Or have you already gained experience with this?

You write 'a widespread and well-known problem'. I have never read or heard anything about this problem before , even though I have been active in the forum for several years.

So, my replacement MSG 028 906 021 JC also has the software problem too...

I've also been thinking about the starter motor.
There are already new parts available for around €80. Are they any good?
I would almost prefer to rebuild my old starter motor myself (replacing the brushes, machining the commutator, and installing new bearings).

Best regards from Baden-Württemberg.

Yes, as far as I know, all AFN engines and even some others with VP37/Bosch-EDC are affected by this problem. It only occurs in BlitzStarters and is extremely rare even there. It doesn't happen in standard Otto engines because they tend to overcrank instead.
I only know that it somehow has something to do with the software related to the starting speed (or activation speed).
With the temperature sensor disconnected, the problem will not occur, just like it wouldn't with a genuinely cold engine (so you can save yourself the trouble of trying that).
The VAT exchange system brings nothing but unnecessary costs.

Therefore, when the engine is warm, the starting fuel amount is reduced to prevent black smoke from coming out the back.
Therefore, the amount of fuel is already tight for takeoff, and an insufficient starting speed and duration will cause the rest of the problems.

I would definitely not buy a starter motor from Eblöd, but rather, as you mentioned, have your existing one refurbished or purchase a decent replacement part from Bosch/Valeo. Make sure to check the power rating on the type plate and don't install anything with lower specifications!
"...first check the idle speed using VCDS before you replace it unnecessarily..."

Best regards from Munich!

Hi DieselBär30x,

Once again, thank you for the quick response and explanation of the cause.

I will definitely measure the idle speed using VCDS (I still can't quite get used to the other abbreviation for VAG-COM ).

Then, if necessary, find a good, used starter motor of the same type (1.8 kW) and overhaul it yourself.
I haven't yet inquired about the all-terrain tire price from Bosch, but it will certainly be above the €200 mark.

Due to the absence of a problem with the engine when it's cold, I previously described a circuit in another post where the magnetic field of the energized starter cable activates a reed relay (a glass tube with internal contacts), which then disconnects the coolant sensor during the starting process.
If you can find a suitable reed relay, that would be quite an elegant solution.

...und provoziert dann Fehlereinträge für "Signal Kühlmitteltemperaturgeber" - ist schon bißchen mehr Pfusch als Problemlösung, oder?

You would just need to keep the starter engaged for a little longer (maybe even an extra 0.5 seconds!), and that problem would be solved.
Audi/VW, on the other hand, do not consider this phenomenon to be a defect, but rather a result of improper use.

But first, measure the idle speed with the engine warmed up(!), and let me know what you find.

Best regards from Munich!

Oops.

No, I definitely don't want to create any error entries ...

For purely technical interest: Are both coolant temperature sensors compared, or is it sufficient to feed a plausible value, say 0°C, to the Motor Steuergerät (MSG)?

I will definitely measure and post the idle RPM of the warmed-up engine.

Good night!

Knockout.

Theoretically, it is possible to tamper with the temperature sensor of the VAT (Value Added Tax) system.
As I said, I see this kind of thing more as (completely unnecessary!) sloppy work.

My search found something here in the forum:
/viewtopic.php?t=1241&highlight=500
You have it written in black and white .

Best regards from Munich!

Hi there,
"@ DBär, a question for clarification: In the state where the engine speed is too low and the electromagnetic valve is stuck, does the MSG (Multi-Stage Injection System) engage, or does it cause the quantity control to open and close repeatedly (jumping back and forth across the start authorization threshold) due to insufficient, fluctuating engine speed? The shut-off valve is probably open."
I'm considering adapting the starting mixture or adjusting the ignition timing to get more torque at startup.
"@ Oldewurtel: Repairing a used starter motor is almost not worth it. I've done it myself. A repair kit, including the brush holder, the freewheel mechanism, and some small parts, costs around €150. The necessary high-temperature grease is only obtainable with a significant time and mileage investment. You can't get it from VW at all, and you can only get it from Bosch if someone still remembers the good old days when starters were repaired in workshops and knows where the tube of grease is located."
If you're going to buy one, then get an original ET.
hg
Herbert.

Moing HerBärt!

To be honest, I have no idea.
That might be something that one of the "freaks," such as Mulle, Wiesel, or Rainer, would know .

It feels more like the former, as the engine speed is constantly maintained at exactly 500 RPM (without any fluctuations!). I suspect that the "Start-Modus" of the Motronic system isn't disengaging, because in that mode, the fuel injection amount is also fixed and independent of the accelerator pedal position. This is why it doesn't respond to driver input until restarted.
ELAB will certainly be open .

Best regards from Munich!

I've added something else about it; please see above.
Sure, here is the translation of the text from German to English:

"lg" translates to "best regards" or "kindly regards" in English. It is a common abbreviation used at the end of emails and letters.

Always those notorious editors .
Increasing the initial mass or shifting the molten slag window (MSW) won't achieve anything in this case, except possibly producing black smoke – I've already tried it.
I believe that the engine speed must simply exceed 500 RPM for the system, either through the engine control unit (ECU) or its software, to switch from "start" mode to "idle/operation" mode, including all associated controls.
And these U500 values are achieved using the starter motor plus a starting amount and a specific time, after which it runs properly.

It's not as common or critical as it might seem from this description.
You really have to do something stupid or deliberately mess up the starting process to end up in this 500 RPM state, and even then, it only works reliably about once out of twenty attempts.
So, just hold the starter for half a second longer and everything will be fine again .

Best regards from Munich!

Even with my '98 ALH, it happened; I had the feeling that the engine was running in reverse .

Here's something from the troubleshooting corner: Does a temperature sensor, switched to a fixed value for the starting process, trigger more than just an error code? Does the MIL (Malfunction Indicator Lamp) turn on, and could it potentially switch into limp mode? If not, especially in older vehicles, I would disregard the error code and, while the starter is energized, switch the temperature sensor to a fixed value determined by the software.

It might even be possible to find a substitute value experimentally that doesn't trigger an error memory entry. Setting up and creating such a workaround seems justifiable in terms of effort. In the case of an older used vehicle where the manufacturer refuses any support and blames the driver for errors, I would be quite unscrupulous in devising and implementing a solution to fix the problem.

Hutfahrer wrote:

Let's try something from the "Frickelecke" (a playful term for a niche area): Does the temperature sensor, which is switched to a fixed value for the initialization process, trigger more than just an error memory entry? Does the MIL (Malfunction Indicator Lamp) turn on, and could it potentially also go into limp mode?

In my opinion, no, as long as the signal voltage does not leave the target range for too long.
To avoid ever getting an "open" sensor reading, I would connect a dummy resistor in series with the sensor, which is normally short-circuited via the normally closed contact of an additional relay.
Upon startup, the relay must engage and simulate a cooler engine temperature by drawing current through a high-value resistor.

Quote:

Wenn man den Schlüssel etwas länger festhält (gefühlte fünf bis sechs Kompression) läuft er als wenn nichts gewesen wäre.

Then just do it. What's the problem with starting for just 1 or 2 seconds?

Anyone who dealt with steam engines in the eighties knows what it means to have a long start-up time.