VAGCOM: Engine computer not responding

Hello

"Out of curiosity, I also built an adapter without an optocoupler, but without ICs, but only with transistors, diodes, LEDs, and resistors."
According to the Oszi measurement, this device has significantly "better" transmission properties than my opto-adapter.

All STGs could also be easily addressed – except for the engine. From there came further error messages, in a surprising variety:
No response from the control unit.
Too many Communication Errors to Continue
Can't synchronize Baud Rate
Controller uses Unknown Protocol (each time with a different number) . . .

In a subsequent test with my "disabled" Opto-adapter, everything worked perfectly.
So, I connected the oscilloscope to the K-Line and observed the signals in the communication with the individual STG's. The timing interval between the signal peaks was consistent everywhere – approximately 100 usec = 10 KB.

However, the zero-level signal coming from the STGs was different on the K-line: ABS and WFS practically showed 0.0 volts, while the airbag showed approximately 0.5 volts and the motor showed 1.0 volts.
So, it seems that the 1 volt from the motor was not enough to make my adapter recognize a 0, even though the calculated switching point varied between 4 - 6 volts depending on the voltage at the COM port.

During a review of the adapter, I found a manufacturing defect that was actually capable of lowering the 0-detection threshold to around 1 volt.

Once the error was fixed on my adapter, the device is working perfectly.

However, despite the superior "speed" of the adapter, I am not able to achieve higher sampling rates with the opto-part.
Therefore, the experimental adapter will be recycled

By the way, the Noxon's minimal switch actually has a similar Achilles' heel as my adapter before the correction... anyone who has purchased this "popular" switch as an adapter might experience similar problems.
Conveniently, it's often not possible to non-destructively inspect many purchase adapters to verify the issue , but one would have to perform a tedious and relatively complex measurement with the adapter on the laptop to determine the threshold. . .

@ulf

mit der Abtastrate kommst du auch nicht wesentlich höher. Bei den TDI Steuergeräten dieser Generation ist deine Abtastrate die Grenze! Mehr geht nicht. Da kannst du den Adapter optimieren, wie du willst!

'For some gasoline engines, you can get up to 4-6, but that's only for the GTI and VR6 engines! Your 2.4 adapter is already optimal! It can't get any better!!!!!!!!! This also proves that the one I installed on my carburetor ultimately worked, although with tricks, but still!'

and I am well aware of the error messages with your latest version. completely different. sometimes like this, sometimes like that. it was really frustrating!

I think that if you want to create a general design, then use the 2.4 as a base and add the diode array to it, and that's it. The part with a reasonable PCB should actually be reproducible by almost anyone who has ever used a soldering iron.

joergs wrote:

With the sampling rate, you also won't get significantly higher values. For the TDI control units of this generation, your sampling rate is the limit! There's no more. You can optimize the adapter as you like!

Hi Jörg

Now I also know it
However, Rudi had emailed me that he could get higher rebates with his own, non-Optos-based design than with the Noxon circuit. That naturally piqued my interest, and I wanted to try something similar. . .

Quote:

I think, if you want to design a general one, then take the 2.4 as a basis and build a diode gate on it and that's it. The part with a reasonable PCB should actually be reproducible by almost anyone who has ever held a soldering iron!

Yes, but the 2.4 reacts with noticeable bit length changes, especially on RXD, even with brutal voltage fluctuations on the board.
According to ISO standards, diagnostic devices should be fully functional up to 8 volts.
While I haven't tested the 2.4 GHz band under these specific conditions, it's likely to fail quickly if the settings are not properly adjusted.

With a 78L05 for the opto-power supply on the driver's side, you can definitely resolve this issue.
And RXD, signals with at least 5V / usec are generally better than those with just 1 V / usec. ... You don't know which "quirky" COM ports there are that are also suitable for things like VAGCOM.

I'll send you my current theory of the "Endless Adapter" for your comments.

Hi Ulf,

However, Rudi had emailed me that he could get higher rebates with his own, non-Optos-based design than with the Noxon circuit. That naturally piqued my interest, and I wanted to try something similar. . .


Probably, the reversed wording would have been better. The Noxon is slower than my homemade

Best regards, Rudi

Hello Ulf!

I had the exact same problem yesterday!

All the control units could be accessed, but the MSG could not.
Error Message: Connection unstable; too many communication errors!

The fascinating thing is: It was working perfectly fine on Saturday!

I am using the following configuration: PC Desktop 486/66, Windows 95,
Adapter 'free' for Noxxon, VAGCOM 208.2

Previously tested on the following vehicles: A4 AJM MJ00, Polo AEE MJ98; Galaxy 2.3l MJ99
Works perfectly with VAGCOM factory settings, as mentioned, until yesterday.

I also tried version 208.1 (Eng.) just to be sure; the same effect, plus an additional error: When reading the data blocks, the following Windows message appears every time: 'Application is closing due to an invalid operation' (It seems to be addressing invalid addresses).

Your theory about the zero-level of the MSG is interesting, but I honestly can't imagine that this level is 'subject to daily fluctuations.' I will check the adapter anyway.

IMO, VAGCOM is far from reliable, therefore it reacts extremely sensitively to even small deviations in the configuration.

Greetings, Jürgen

@juergen

The resulting errors are a combination of adapter problems, VAG-COM synchronization problems, and STG bit length fluctuations.

Regarding your crash issue: I'm familiar with this problem with version 2.082. It's been appearing in various forums quite often lately. Therefore, I'm not concerned that it's due to my installation or my hardware. A new DRV version is announced for the next few days! Maybe it will run stably.

Regarding your communication problem: Which VW part number is the STC that is not reading properly?

@ Ulf,

I have, as promised, continued to follow the story with K-lead = 1V.

My MSG also provides a low voltage of 1 V!

So, I connected a 100 Ohm potentiometer between the emitters of the transistors and GND in my Noxon adapter, and I measured the voltage between the base and emitter (UBE) of the relevant transistor.

As expected, the measured values for Poti = 100 Ohm and Poti = 0 Ohm are almost identical (0.78 V and 0.02 V respectively) and are within the acceptable range!

Furthermore, I have measured UBE in various STG tests; the results were always identical!

So, it definitely isn't due to the 1V low-level!!!!!!!!!

Incidentally, I measured between all STG terminals from K and GND up to 40 V !!!!, but I have absolutely no idea where they come from.

@ jörgs

The MSG is a Marelli 1AV, serial number 032906030 (Motor AEE).

Your comment confirms my suspicion that the problem is related to the VAGCOM!
It's almost outrageous to demand 120 EUR for a single license for something like this (Bill Gates should take note).
However, I have already looked for an alternative and will post it!

Greetings, Jürgen

@juergen

Your control unit is also a problem, just like the dino-1z-stg. You can also read about it in various forums.

However, I wouldn't say that VAG-COM is 'bad.' It has its weaknesses. That is correct. But immediately talking about such things seems a bit exaggerated!

What are your settings for the options in VAG-COM? I've been able to achieve good results by adjusting these settings!

Juergen wrote:

So, I connected a 100 Ohm potentiometer between the emitters of the transistors and GND in my Noxon adapter, and I measured the voltage between the base and emitter (UBE) of the relevant transistor. As expected, the measured values for Poti = 100 Ohm and Poti = 0 Ohm are almost identical (0.78 V and 0.02 V respectively) and are within the acceptable range! Furthermore, I have measured UBE in various STG tests; the results were always identical! So, it definitely isn't due to the 1 V low-level!!!!!!!!!

Hi Jürgen

???
With 100 Ohms in parallel to B-E of the K-sensor transistor, no logic 1 will be detected on the K-line at all:
If a potentiometer with a resistance of 150 kΩ / 100 Ω is used, with a value of 150 kΩ between the base and the collector of the sensor transistor, then a voltage of approximately 900 volts would need to be applied to the input in order to generate a B-E voltage of 0.6 volts.
???

@ jörgs

I've tried all sorts of settings - the same (bad) result every time!
Just to reiterate: on Saturday, it worked perfectly, and on Tuesday (with the same configuration), it didn't work anymore.

To VAGCOM:
If the program cost 20 EUR or at least had free updates, I wouldn't complain.
But I still stand by my point: demanding 120 EUR for a banana program (which ripens at the customer's location) is outrageous.

@ Ulf

The 100 Ohm resistors are not located between the base and the emitter, but between the emitter and ground. When a transistor is switched on, approximately 1V drops across the 100 Ohm resistor, therefore, according to my thinking, this should solve the problem of the K-line being low = 1V.
But as already mentioned: the transistor also switches on perfectly without a 100-ohm resistor.
Therefore, I conclude that, at least with the Noxon adapter, the 1V level is not a critical issue.

Greetings, Jürgen

Juergen wrote:

The 100-ohm resistors are not located between the base and the emitter, but between the emitter and ground.

Hi Jürgen

Aha, jetzt hab ichs verstanden.

Quote:

When a transistor is switched on, approximately 1V drops across the 100 Ohm resistor, therefore, according to my thinking, this problem of K-line Low = 1 V should be solved.

IMO not: that means you have a partial emitter follower.
If you reduce the base voltage, the emitter voltage will remain approximately 0.6 volts. This means that the voltage drop from Ub = 1.6 volts to 0.6 volts. Due to the 100 Ohm resistor, the current flowing through it decreases continuously (as also through the optocoupler at the collector), until the load current returns to its normal value !! only really drops to zero at around 0.6 volts.

More effective would be, for example, 22kOhm in parallel to E-B. In conjunction with a base resistor of 150 kOhm, a (trigger) threshold is achieved in the range of 5 volts.

Quote:

But as already mentioned: the transistor also switches on perfectly without a 100 resistor.

I find it hard to believe - until I had seen the voltage difference, for example, between the collector and the power supply on an oscilloscope. IMO, this is definitely not a rectangle, but rather a long ramp on the falling side of the K-line, which makes the slew rate at RXD even slower than with Optos, which is unavoidable.
With your 100 Ohm, the signal slopes are likely even flatter, and therefore "less digital" than in the original circuit.

Quote:

Therefore, I conclude that, at least with the Noxon adapter, the 1V level is not a critical issue.

I see it differently - see above.

Hi Ulf!

With the emitter follower, of course, you are right; however, I thought that it was only about small deviations in the circuit, and that the 100 Ohm potentiometer could be used as 'fine-tuning'. (It is human to err!)

Since I was too lazy to take my oscilloscope to the garage, I measured with a multimeter (shame).

The story of the signal flanks is particularly interesting with regard to the apparently fluctuating signal times.

If I wasn't so annoyed, I would install a Schmitt trigger.

Greetings, Jürgen

Hello,

I have a suggestion for the 'sensor'.




What do you think?

By the way, I also found another adapter.
Likely using a DC/DC converter with MAX232 and MC33199 for K-Line, and aiming for 200kbps.
However, I haven't found much information about the MC33199, which seems to be a Motorola product.
Greetings

data sheet can be found here:

http://e-www.motorola.com/webapp/sps/site/prod_summary.jsp?code=MC33199

Greetings

Nicklas wrote:

Hello, I have a suggestion for the "sensor". What do you think about it?

Hi Nick

If the PC-COM port needs to be on the left, there should not be an even number of inverters on the way to the K or L line (74 unit and transistor stage).

And only with a +5 Volt supply can the 7404 not represent any negative levels for RXD. . . if this is the complete adapter, then IMO it will not work.

Hi,

The entire schematic is available this weekend, I'll build and test it then.
I just wanted to discuss the part of the K-Line read transistors that uses the Z-Diode. Should it work out okay, or?
The resistors are just estimates, I should recalculate them, or if I have plenty of time, I can simulate them in PSpice, but I prefer to build them manually *g*
by the way, what do you think about the suppressor diodes on the K/L line and VBat? I think they should be attached to every pin that goes into or comes out of the car.

Greetings