ABS/EDS Error: Troubleshooting & Fixes

Hello to the experts!

In my Corrado 2.0 model from 1994, a 16-pin OBD connector was already installed from the factory, located under the gear stick cover. I can read the engine control unit with my KEY-COM adapter. However, the ABS is still diagnosed using blink codes. In this way, I was able to identify a defective front sensor cable. Since I couldn't find the diagnostic connector, I simply connected the clamp directly to ground at the error light to initiate the blinking sequence. I have since discovered that this wire also leads to the 16-pin diagnostic connector, specifically to pin 14 (which is now assigned to CAN-Low).

To install an electronic limited-slip differential (EDS), I converted the ABS Teves 02 system to an ABS/EDS Teves 02 system. Although both are called Teves 02, they are completely different brake control units, because the standard ABS system has 35 pins, while the ABS/EDS system has 55 pins. Consequently, I had to replace all the components: the hydraulic unit, the wiring harness, and the control unit; I was only able to keep the sensors. By the way, there are also ABS and ABS/EDS Teves 04 systems – both with a 55-pin control unit – but again with a conventional vacuum-assisted brake booster, as is common to this day. According to the VW repair manual, this system may have only been used in right-hand drive Corrado vehicles.

"I've been working with ABS/EDS systems for a few years now, and I've noticed that the blinking function doesn't work. In the VW repair manual for the Corrado chassis self-diagnosis, in the ABS/EDS Teves 02 section, it states that different control units were used with varying fault memory readouts: up to 07.91 with blink code output, and from 08.91 with fast data transmission. It is noted that only the version with fast data transmission is available as a replacement part, which means that even older vehicles are automatically equipped with fast data transmission. Since you can't determine the vehicle's production date, it is recommended to always start the self-diagnosis process using the fast data transmission mode."

Since the modification, I have two additional wires connected to a 2-pin connector under the shift lever cover: gray-yellow and gray-red. Strangely, common 2x2 adapters don't fit. In an American wiring diagram for the Corrado, I also see two diagnostic lines that come from pin 42 (gray/white) and pin 23 (yellow) of the engine control unit. The yellow wire is labeled "Canada only," and the other one I believe is the K-line. In any case, my KEY-COM adapter can't do anything with either of these lines. I've also tried various start baud rates: 9600, 4800, 2400, 1200, and 240 – all without success. Could it be that this generation of engine control units requires both K and L lines to establish communication? The L-line was originally intended only as a trigger line to initiate communication, but it was quickly discontinued. My original diesel adapter should be able to handle fast data transfer from the beginning, but is an old software version required for that?
Best regards,
Andreas.

Hi.

A friend has an early Golf III Cabriolet that could only be read out using my 1551 diagnostic tool. VCDS was unable to establish a connection to the control unit, and the car also had the 2x2 connector installed.

Best regards, Rüdi.

Hi,

Yes, very old engine control units (ECUs) require K-line and L-line communication, and the KEY-COM is capable of supporting that.
K-Line is pin 7 in the diagnostic connector, and L-Line is typically pin 15.
Relatively modern engine control units do not require an L-line if the "new" ABS system only has a K-line. In that case, you can connect the K-line to the free pin of the L-line in the diagnostic connector.
Otherwise, connect both correctly.

Best regards, Rainer.

Hello Ruedi,
Were all the pins in the 2x2 connectors used for all the connections in the convertible? (If you remember).

Hello Rainer,
I tried reading the data again today using the correct wiring: Unfortunately, I still didn't get any response from the control unit.
Of course, I tried both options, as I don't have any documentation to identify which pin is the K-line and which is the L-line. The KEY-COM has a pull-up resistor of approximately 500 ohms connected to battery positive on both pins (7 and 15).

"This can be measured. During the connection attempts, I observed what was happening using a multimeter. A multimeter is too slow for this kind of thing, but at least you can see that there are complete transitions between high and low levels, and it seems to me that the same thing is happening on both lines."
But is that the intended purpose? My adapter is labeled "Dual-K." How does the adapter or software know that it's not just reading pure K-lines, but that an L-line control signal needs to be sent?


Best regards,
Andreas.

Hi.

Quote:

that there are complete switches between high and low levels, and it seems to me that the same thing is happening on both lines. But is that the intended purpose?

Yes, it is. If the control address transmitted via the K-line is incorrect, there will simply be no response or an incorrect response on the corresponding K-line.

The 5-baud initialization should ideally "trigger" the L-line, if it exists. However, this is just speculation on my part, as even in my younger days of tinkering, the L-line was already outdated...
I'm currently looking at https://community.dieselschrauber.org/en/viewtopic.php?t=3089, and the cable color of the L-line is yellow. This might match your description.

Without a wiring diagram and pinout for the connector of your ABS control unit, you're unlikely to get very far. What is the VAG part number?

Best regards, Rainer.

Hello Rainer,

I already have the pinout for the connector of my ABS/EDS control unit, specifically in the form of the American wiring diagram (Pin 42, gray/white).
and pin 23 (yellow) connects to the diagnostic connector.

Unfortunately, the cable colors don't match because I have gray-yellow and gray-red.

I unplugged the connector to confirm that it's a 55-pin control unit. However, I can't read the part number right now because I can't get a screw on the bracket loose (it's not fully threaded and is stuck...). I'll continue working on it tomorrow.

Best regards, Andreas.

With a good screwdriver, it worked. The part number is 357 907 379 A.

Interestingly, the sticker on the unit says "ABS control unit" and nothing about EDS. However, I've observed this function working on several occasions, such as when starting with one tire on a slippery shoulder, or when downshifting and engaging the clutch on a slippery road surface, where you can hear the EDS valves clicking.

How do I properly connect to the L-Line now?

Andreas.

Hi,

It might look wrong in the picture, but the K-line (yellow wire) is sufficient in the example. Are you sure you have everything connected correctly on your end?

Die K-Line ist im Bild in der letzten Kontaktreihe, der 5. Pin from the right.

Best regards, Rainer.

Hello Rainer,

"I haven't connected anything directly to the control unit, because I bought a complete, undamaged wiring harness for the conversion. Since it came from a Corrado, I was able to easily install it using a plug-and-play system. I had the hydraulic unit replaced at a local workshop (it was practically done at the same time as the brake fluid change)."

Regarding your picture: My 55-pin control unit looks completely different; it resembles the image labeled 'ABS EDS control unit number 1HO 907379 A.JPG'. And I also found a picture online of a 1H0 907 379 D, which doesn't look like the one in your picture either (ABS_Steuergerät_aus_94er_Golf_3_2E.JPG).

Admittedly, it has always seemed strange to me that these two lines have a different connector than the one used for the 2x2 diagnostic adapter. Also, they are unusually thick for communication lines. The previous ABS wiring harness didn't have this branch at all, because, as I said, it was still a 35-pin control unit where the state-of-the-art blink code output was implemented. Nevertheless, from the factory, my car didn't have any 2x2 diagnostic connectors installed, but as mentioned earlier, the blink code output line (i.e., the ABS fault lamp K47) is connected to the 16-pin diagnostic connector, specifically to pin 14 (the lamp lights up when I short this pin to ground).
Okay, I'm going to remove the control unit again to check with a multimeter whether the supposed K/L wires actually lead to the expected pins 42/23.

Here's a closer look at the connector. I misspoke in my previous post; the K-Line is the 5th pin from the right, which is pin 42.

Best regards, Rainer.

Hello Rainer,

"Yes, your new picture clearly shows that it's from a 55-pin Teves-02 control unit, and it matches the situation in my car. Using a multimeter, I've determined that the gray/yellow wire is connected to pin 42, which ends in the center console in my car. The American wiring diagram also matches, except that the wire color is gray/white there. This must be the K-line. The second yellow wire in the diagram, labeled "Canada only," goes to pin 23, just like in my car, but the wire color is gray/red in my case. This must be the L-line."

Unfortunately, I also made a mistake, and the part number is not 357 907 379 A, but rather 535 907 397 A.

Then I have some good news: Since the error is no longer occurring sporadically, but the warning light is constantly on, I was able to identify the problem at the control unit connector: According to the wiring diagram, the sensor connections are as follows:
Front left: between pins 48 and 30.
Front right: between pins 47 and 29.
Rear left: between pins 46 and 28.
Rear right: between pins 45 and 27.
Everything has a resistance of approximately 1.1 kOhms, except for the one on the front left, which is infinite, meaning it's broken. I then measured directly at the sensor to determine if the left cable was faulty (I had to replace the right one back in 2010), and: It's clear that the sensor itself is defective, as it has an infinite resistance.

"However, I don't want to have to keep driving until the fault becomes permanent next time, but rather be able to read the error codes. Surely, this should be possible with my control unit as well (The casing looks different again from all the pictures above, as it's a die-cast housing with cooling fins on the edges and closed flatly with metal covers at the front and back; green sticker). Or is it possible by blinking the light, and am I doing something wrong? Shouldn't the wire of the error lamp be connected to ground while the ignition is turned on, and then the connection should be disconnected after about 5 seconds, or what?!"

Best regards,
Andreas.

Halt! Kommando zurück! Keinen Kopf mehr zerbrechen über die L-Leitung, denn während ich es absendete ist es mir klar geworden und habe auch gerade draussen an meinem Auto die Blink-Codes dieses Teves 02 ABS/EDS, Teilenummer 535907397A erfolgreich ausgelesen! Da"How" follows.

Andreas.

Rainer, I'm sorry, but this is a completely confusing story: The thing about the 5 "s" on the mass bridge during ignition is correct, but the trick is figuring out which of the incredibly many wires to use!

The two changes, from 2x2 diagnostic connectors to a 16-pin OBD connector, and from blink codes to fast data transmission (K-Line), have created a huge compatibility issue.

To me, it seemed completely illogical that, according to the VW repair manual, simply replacing the ABS control unit and without modifying the factory-installed wiring harness could suddenly cause the K-line to be wired differently, instead of displaying a blink code. Somehow, the wiring harness must already have taken both possibilities into account, but how?
As reported, in my original ABS wiring harness, the wire leading to the ABS lamp also served as the ground connection needed to initiate the diagnostic code output. However, the K-line requires its own dedicated pin, independent of the warning lamp.

In older ABS control units with 35 pins, there are two diagnostic pins, namely pins 26 and 27, which, in all wiring diagrams I've seen, are connected together within the wiring harness and lead to both the warning light and the diagnostic connector.
By disconnecting the control unit, it cannot be separated. I should have forgotten about that, because it remains a potential issue.

If you think about it, there are actually three lines for diagnostic purposes in the 55-pin control units! In addition to the warning light, there are two more (K/L lines). It seems that when developing these units as a transitional solution for devices that still output blink codes, the developers didn't come up with anything better than "misusing" the already pointless L-line to simulate the blink code output, because it works for me: On the gray/red line (which I've confirmed comes from pin 23), I have approximately 10V when the ignition is on. The gray/yellow line from pin 42 (K-line) is dead (0.0V) in my case.

"However, pin 23 is also not internally connected to the error lamp. Nevertheless, initiating the blink code output function works on this "new" pin practically "as usual." Unfortunately, pin 14 of my OBD connector still carries the voltage of the error lamp, which is why I thought that if it's related to blink code output, it *must* go through this pin. However, that's not the case."
Several error codes are now appearing, and these are stored even without power supply (to reach the next error, you need to short the ground connection again).
1112 - ABS inlet valve, front left.
1222 - ABS main valve.
2133 - Pressure sensor for ABS.
1233 - Speed sensor, front left.
1241 - Speed sensor, front right.
1311 - Rear left speed sensor.
1243 - Rear right speed sensor.

So, it seems like almost everything should be broken down! I think the large number of error messages is the consequence of continuing to drive despite the defect. I'm sticking with my assessment that only the front left sensor is faulty, and I'll get back to you when I've replaced it.

Best regards,
Andreas.

"Bingo: After replacing the front left sensor, the warning light went off on the next drive, after about 1 km on a country road. (It took some time because the past few weeks had no trips that couldn't be easily done by bike.) Today, just to be sure, I checked the system again: the response was 4444, meaning no errors are stored – which is good."


Hello,
Andreas.