Oh, Nemo

Today's post covers my challenges with Nemo's electrical and computer system simply trying to get him to start, or maybe a signal on the OBD-2 port. The car-work hours that I did not want to spend sanding on Zed, I spent doing this... until this took over. Yeah, I know. I could use another hobby.

No Start

Like so many issues with these old cars, our challenges with Nemo started with him not starting. His battery had been sitting on a trickle charger all winter, but I had left his hood cracked to make room for the alligator clips, so this ultimately was self-inflicted. Love those. But, the path we took to get there was interesting. Nemo is sitting under the fabric carport blocking a garage door. Behind that garage door sits Oliver, the 1978 MGB convertible, that I would very much like to be road testing and otherwise driving around. So, enter my motivation for moving Nemo. Insert key, turn ignition to "RUN". I see some familiar idiot lights, including the battery, oil and check-engine. Turn to "START" and he rev'd but would not start. I went around to the tailpipe and could not smell gas, so I start with the assumption that it is fuel delivery.

Before I did anything else, I considered what I experienced when I turned the key to "RUN". I did not hear the fuel pump cycle. I did not hear the radiator fans kick on. So, something is not getting the "RUN" signal. So, I pull out the UltraGauge and plug it into the OBD-2 port. I get a never-ending "scanning..." message. Well, that's not good. I double-checked the UltraGauge in Hapy, and the computer linked within a few seconds. Okay, we have trouble with Nemo's diagnostic system which could prevent a start or we have multiple things going on. With this car, the latter is probably true. First, I tried to isolate the OBD-2 issue, and kill many many hours doing so.

OBD-2 Issue Diagnosis: General

The early B5 versions of the A4 had some interesting 1 or 2 year only issues. One of the oddities around 1997, when Nemo was sold, was the engineers choice to integrate the stock stereo with the "K-line": the signal processing wiring between the computer, many components and the OBD-2 plug. In these late 90's cars, it is the K-line you are subscribing to, not the more modern "CAN BUS". Okay, cool. There are many modules plugged into the K-line, but they are all in parallel. the diagnosis theory goes that when you get no signal on the K-line it is because one of the modules is faulting out, grounding the K-line. Once you can identify and remove the faulty module, you will get a signal on the K-line. You will probably get a failure code for that module, but you will at least get a signal. So, away we go, trying to figure out what modules are even on the K-line, and then isolating them.

OBD-2 Issue Diagnosis: Radio

The stock radio wiring was unceremoniously cut to pieces by the idiots at Car Toys when they installed an aftermarket stereo. Based on the wiring diagrams I have found and the commentary on the interwebs, I am not sure this car actually had that K-line-in-the-stereo thing. I did check the wires for the right color combinations, and for wires that were the correct colors for the K-line. I was unable to find them. There are, however, many wires sticking out of the hole where the radio used to be... before it was ripped off in Eugene. I intend to re-integrate an original plug before I put all this back together, but that's another day. I moved on to the next possible culprit: the climate control system.

OBD-2 Issue Diagnosis: Climate Control

found in climate
computer

The climate control system in these cars is another little computer. In some models, it is held in place with a couple brass clips. In Nemo, I had to remove the center console cover plate / face and found 2 Phillips screws holding it in. Once threaded out, the unit removes through the front. I unplugged the unit and checked the OBD-2 for a signal. Nope. I did, however, find a soda/beer can pull tab inside it (picture on the right, here, laying on the carpet after I removed it from inside the climate control computer thing). So, that was interesting. Next.

OBD-2 Issue Diagnosis: ABS (Brakes)

The fine folks on the interwebs have found that the ABS sometimes causes K-Line problems. The way to identify if it is the problem is to disconnect the plug. The ABS unit is on the driver (left) side, next to the washer bottle. You simply pull up on the silver metal tab and the plug lifts out of the socket. I found a piece of electrical tape floating around in there. Another question-mark from the prior owner. Neat. I checked the OBD-2 for a signal. Nope. I shot the socket and the plug with DeOxit and left it disconnected for the rest of my tests. I figured I might as well leave it better than it was (DeOxit, not leaving it unplugged). Next!

OBD-2 Issue Diagnosis: Dash Cluster

Moving on, we consider the dash cluster. The entire thing is held in by 2 Torx-head screws from the front, hidden behind a cover plate just above the steering wheel. Pull the cover plate, remove the screws and it pops right out. I unplugged the various cable plugs and checked the OBD-2 again. Nope. Grr... Next!!

OBD-2 Issue Diagnosis: ECU

example A4 ECU 

It seems like every system on the engine that has more than one wire into it delivers some signal to the computer. So, this seemed like the next logical step. I figured if I could get a signal after unplugging the small plug on the ECU, I would have shown that the issue is in the engine compartment and I could put everything in the cabin back together again. If the issue remains, then it could be the ECU. So, I disconnected the small plug from the ECU and we still didn't have a K-line signal. Grr...

So, I plugged the ECU, ABS and the dash cluster back in, turned the key to RUN and heard the click-snap of some relays firing, but no fuel pump nor fan spin. I could not remember if I heard the relays before, but this changed my thinking. If at least some of the relays are firing, then the ECU is telling them to. If there is a Check Engine Light, then the ECU must be at least partially working. Perhaps the K-line / OBD-2 issue is completely unrelated and this is just a simple fuel supply issue. So.. new plan: get the fuel pump to fire.

Fuel Pump Test

As is so often the case in laboratory settings, the pre-work for the experiment is where all the time is taken. Such is the case here. I started by removing the plate covering the fuel pump. This sits behind the rear seat in the trunk on the passenger (right) side, and is held in place with 3 very short Phillips screws. Once the plate is removed, you can see the top of the fuel pump. There are 2 fuel lines, one supply and one return, and a single electric plug. I found that slightly wiggling the plug allowed it to pull off the pump. YMMV. Once removed, you will see 4 pins. 2 of them are for the fuel level and 2 are for the pump. I started my tests with basic resistance and continuity-to-ground. Things seemed fine, so I shifted focus to test-firing the pump by applying 12V to the pump.

This is where the experiment set up took some time relative to the actual test execution. I ran a pair of wires from near the battery (not hooked up) back to the fuel pump. At the battery, I taped 10A fuse to the "red" wire and alligator-clipped the other side of the blade to the positive battery post. I alligator-clipped the "black" wire to the negative post. Back at the fuel pump, I applied the red and black wires to the fuel pump and it fired right up. I held the wires there a few second to demonstrate that it would not just run for a second and die. So, we have a good pump, but the pump won't fire from a relay signal. So, next we check the voltage from the relay, after we put the experiment wiring away.

Fuel Pump Relay Test

A4 Fuel Pump Relay

Back up front, I took a red wire, stuck it into the voltage-supply pin in the wiring plug at the fuel pump and tossed the other end through the cabin, over the outer edge of the driver seat. My plan was to compare the voltage sent to the pump at the pump. Since the relay only pops for a second when you turn on the ignition, I had to have my test equipment near the ignition key. I stuck my black lead against the metal fuse box mount, held the red lead against the wire I tossed through and turned the ignition to "RUN". I got 5V. I checked other spots around the fuse box and got 13V, so I started thinking that the relay was bad, dropping the voltage from 12/13V to 5V which, of course, is not enough to run the pump. So, I swapped out the relay... and the fuel pump still would not run. Great. I concluded that the 5V I read was actually the base charge for the fuel level sender and not the supply for running the pump. 

So, I checked the voltage at the supply-side of the relay: 13V. I jumpered across the relay pins, and the pump fired up. So, we're back to the ECU is not sending a signal. Fearing it is the ECU, I went looking for any other reason. I recalled that when Dot (the white 2000 VW Jetta 2.0 which broke the timing belt) failed, one of the theories for not starting was a crank position sensor. The thinking is that if the ECU doesn't know where in the revolution the crank is, it doesn't know when to fire, so it doesn't try. It is not as clear whether a failing crank sensor would prevent the fuel pump relay from getting a signal from the ECU, however. As much as I have tried to avoid it, it felt like I was sliding into just-swap-parts problem solving. At $35US for the least expensive part-to-swap (crank position sensors start closer to $60US), this can get needlessly expensive very quickly.

Crank Position Sensor (CKP)

A4 Crank Position Sensor

I have mentioned before my belief that the Audi engineers are sadists. Only someone who takes pleasure in other's pain could arrange for replacement items to be so hard to get to. Case-in-point: the crank position sensor. The crank runs the length of the block (obviously), so given a blank canvas, an engineer could potentially place a sensor anywhere along that axis. Chosen location? Directly under/behind the oil filter so you either have to remove the oil filter (forcing an oil change, which arguably is due) OR you need to perform contortions to get an Allen-head wrench in there. If you go this route, consider that the further the bolt comes out, the closer you get to the oil filter. At least there is only one fastener. These sensors have a long lead on them that plugs into the engine harness... behind the coolant bottle. In Nemo's case, this plug was fairly easy to get to. There are methods for testing the sensor, of course.

The Audi/VW sensor has 3 wires, from pin 1-3: power | signal | ground where the power and signal connect to the ECU and the ground finds its way to the chassis. To test, 12V and ground need to be supplied to the sensor plug (once disconnected from the harness) and a multimeter checking the signal wire for an AC wave signal while the engine is turning. No signal = bad sensor. Since no local shops had a CKP, I returned to sanding Bondo in Zed while I waited for the part to arrive.

Once I had the part in hand, I checked the resistance between the pins. I got 500ohms +/- between the middle pin and one of the outer pins. Using that as a baseline, I swapped out the CKP and then tested the one I had just removed. Same resistance. So, I concluded there was nothing wrong with the old one and put it in the stash of A4 spare parts (with the fuel pump relay).

ECU Swap

At this point, I had run completely out of ideas fo4r why the ECU was not sending a "go" signal to the fuel pump relay. I hit the local junkyard looking for a just-dropped-off A4 to scavenge the ECU. None of the A4's there had the right ECU, but I did get an ECU box cover: Nemo's had a good-sized hole in it where I thought water might have entered. Without another option, I hit eBarf and found an ECU with the same part number as the one as was currently in Nemo at a salvage yard on the East Coast (with a 30 day return policy!). Hopeful, I swapped out the ECU's which consisted of pulling out the big plugs in the back of the old and plugging them into the new one, since the box was still lid-less. I had some concern that energizing the ECU could fry it if the underlying condition still existed.

Still, I turned the key to "RUN" anyway... and heard the radiator fans and fuel pump cycle for a second before shutting off. This was how things used to work. So, I turned to start and he fired right up. I have concluded that my first guess was the correct one: I had left the hood open a little bit to make room for the alligator clips on the trickle charger, and with the hole in the ECU box lid, water got into the ECU and partly fried it. If I had not had the experience with the chipped ECU for Hapy when the computer semi-worked, I would not have been so quick to accept this. The Check Engine Light lit, so the A4 chat-boards all assume that means the ECU is fine. NOT TRUE. It could be partly damaged, like Nemo's.

The climate control is still in pieces, the lower dash panel is off, and the radio wiring, well, it's the same as it was: a mess. While I can move Nemo out of the way so Oliver can go for a spin, Nemo is not ready for driving around just yet. The re-assembly will wait, and I'll probably post on at least the re-wiring of the stereo plugs that the idiots at CarToys cut apart. I did, however, install the ECU box lid, complete with threading in the screws to hold it tightly shut. No more water getting in there.

Thanks, as always, for following along-