1997 Audi B5 A4 2.8L V6 12V (AFC)

[davisev5225]

If the 2.8L is anything like the 2.7T, then the heads are "flipped" when installed. If the cam sensor you can find is at the back of the engine bay, then the other cam sensor is probably at the front.

 

[o2bad455]

If the 2.8L is anything like the 2.7T, then the heads are "flipped" when installed. If the cam sensor you can find is at the back of the engine bay, then the other cam sensor is probably at the front.

Good thought, but please read my post #14, last sentence. To recap, it's simply not possible to mount one of these particular Hall cam sensors at the front of a single-cam head. The twin-cam heads can only because they drive the second cam internally rather than with the timing belt.

 

[nugentp]

According to Bentley, the "8D0 906 266B" ECU is correct for MY96 AFC motor, but somewhat confusing references are made to other ECU part numbers. I am struggling to match which of these Hitachi MMS systems maps to each P/N - its the MMS which explains use of 12V on the CMP sensor(s). I might try to do some more digging.

The MY96 version should have ECU with 5 x multi-pin connectors (labelled A-B-C-D-E on the schematic for J192) - thats the one with single CMP sensor - G40 on schematic but VCDS calls out G163 for the fault code as the CMP is on Bank-2.

While the MY97 version seems to have a single multi-pin ECU connector and that version of schematic uses J220 for the ECU and we have G40 and G163 on this engine. Notable absence of the G4 crankshaft position on this version tells you the later AFC engine must have switched to a toothed flywheel sensor with a gap to convey the reference position along with rpm.

Talking of flywheel teeth - on the strongest likelihood that you have the early version with two crank sensors (G4 & G28) you can unbolt those from rear of the engine and carefully rotate the engine by hand to verify the presence of the crankshaft reference pin. The Bentley manual I have does not explain where that should be in terms of degrees BTDC but I expect it will be same as the AAH motor.

Point being - you could have perfectly good G4, G28 and G40 sensors but the absence of a mechanical pin to drive the G4 signal means it can never synchronise, but that doesn't explain why you have the 17800 error code. I have seen the missing G4 pin once before on an old 5cyl engine.

Honestly though - 5 minutes with an oscilloscope will root cause what is wrong. The other possibility is that all the signals exist but G4 and G40 are not aligned correctly - perhaps as mechanical timing is wrong so you should double check that.

 

[nugentp]

The more I look at the Bentley book for the B5-A4, the more i am convinced the 'MY97' schematic for the AFC engine is utter garbage - it is showing an early flavour of partial DBW for idle control (with J338 but no sign of throttle pedal position sensors) for one thing and the ECU pinout shown is suspiciously identical to the later AHA engine with the 5V heads. Only notable difference is lack of CANbus wires on what the Bentley shows as MY97 AFC engine. I call nonsense on that and that Bentley has a mistake here.

All this confusion has come from VCDS association of the CMP fault code for Bank-2 to explicitly call out G163 when it should in fact be G40 on this engine. This is a bit of a minefield for VCDS to be fair - I am not for a second suggesting anything else - such anomalies are very hard to cater for - especially on engine variants that only existed for such a short amount of time.

 

[o2bad455]

The more I look at the Bentley book for the B5-A4, the more i am convinced the 'MY97' schematic for the AFC engine is utter garbage - it is showing an early flavour of partial DBW for idle control (with J338 but no sign of throttle pedal position sensors) for one thing and the ECU pinout shown is suspiciously identical to the later AHA engine with the 5V heads. Only notable difference is lack of CANbus wires on what the Bentley shows as MY97 AFC engine. I call nonsense on that and that Bentley has a mistake here.

All this confusion has come from VCDS association of the CMP fault code for Bank-2 to explicitly call out G163 when it should in fact be G40 on this engine. This is a bit of a minefield for VCDS to be fair - I am not for a second suggesting anything else - such anomalies are very hard to cater for - especially on engine variants that only existed for such a short amount of time.

I fully concur, thanks! I'll just add one more data point that I noticed from viewing actual photos of used 8D0906266B ECUs on eBay. Most were "MMS-411B" like mine, but one was "MMS-411C" despite having same ECU part number with "B" suffix.

I have ordered a 2-channel scope. A 3-chan was out of budget, but I think I can meas 2 of the 3 at a time to finally sort it out. I'll follow-up when I do. Thank you for all of your help!

 

[NEtech]

OP, please post an complete Autoscan with latest VCDS version.

 

[nugentp]

I fully concur, thanks! I'll just add one more data point that I noticed from viewing actual photos of used 8D0906266B ECUs on eBay. Most were "MMS-411B" like mine, but one was "MMS-411C" despite having same ECU part number with "B" suffix.

I have ordered a 2-channel scope. A 3-chan was out of budget, but I think I can meas 2 of the 3 at a time to finally sort it out. I'll follow-up when I do. Thank you for all of your help!

i *think* the MMS-411C may have additional OBD capabilities (or better standards compliance for whatever US regulations were in effect) - I noticed it somewhere but cannot recall where.

Good job on buying a little 2-channel scope - you don't need anything fancy but its a game changer for engine diagnostics and removes a huge amount of guess work (and part swapping) when looking at no start situations like yours.

 

[nugentp]

One more daft thing to check is that the G4 and G28 plugs from the two crankshaft sensors are correctly assigned - there is a connector mounting plate in/around the bulkhead and you should see BLACK coloured male & female plugs coupled for one sensor and GREY coloured male & female pair for the other sensor. If those are accidentally crossed up then it will never start and I don't know if the OBD goodness is of 1996/7 is clever enough to tell you they are crossed up. The 17800 error still points to some specific badness on the CMP sensor.

 

[o2bad455]

Wrap-up:

There's only one cam sensor at the rear of the left bank cylinder head on this 12V AFC engine. The Bentley calls it G40, so the VCDS report of G163 was apparently a red herring. There is no G163 on the engine.

The actual issue in this particular car was a broken left camshaft. The shaft had a clean break hidden under the second to last bearing cap just before the cylinder #6 lobes, apparently due to oil starvation at the rearmost camshaft bearing since that's badly galled. So the lobes for cylinder 6 and the sensor wheel itself were simply not turning. The oil starvation issue likely originates in the bottom end, so I probably won't bother trying a simple cam and/or head replacement (even if I could find one).