What most people think when it comes to "diagnostics" in relation to an automobile is that it simply comes down to "plugging it into a computer" and then that magic computer tells a mechanic what part is faulty. I wish - I really do - it would be as easy as that... In reality, the computer (the right name for this device is a "diagnostic scan tool", aka "scanner") can only read some data from the memory of the onboard control units. This data might include diagnostic fault codes, live data, and freeze frame data associated with some of the stored fault codes. In some cases, the component mentioned in a fault code description can indeed be faulty. If that's the case, it would in fact appear like the computer found the problem. Only it is not always such a case...
On this day, the 2008 Dodge Nitro comes in with a whole light show of warning lights and messages on the dash, wipers running non-stop, and a long, long list of fault codes in numerous control modules. By the way, none of those fault codes said, "there is a broken wire under the passenger seat, it is a white cable with an orange stripe". Nope, the majority of codes were like "lost communication with this or that module", "low voltage" and "CAN bus fault". There was also a message "no bus" showing up on the instrument cluster instead of the odometer mileage.
The customer has already tried to replace the TIPM module - the unit with relays and fuses under the hood, behind the battery, that occasionally gives troubles on Chrysler/Dodge vehicles. That only made it worse - nothing worked with another TIPM at all. The customer re-installed the original TIPM and brought the car to my shop.
By the way, if you think that computer data bus on a car looks something like shown on these pictures...
...then you will be disappointed. This one below is an actual piece of the data bus from this very Dodge Nitro. This is only a pair of wires twisted together. Because control modules on a car only exchange short messages, not a big files, these two wires are enough to get the job done
Because of the multiple indications regarding the data bus and communication problem I started by setting up a digital scope and checking the CAN INT signal waveform. Normally, the CAN bus signal waveform is expected to look like this when reading voltage from both of those two twisted wires (red and white graphs)
This is a waveform of a single message being transmitted on the CAN bus. It takes only 230 microseconds for this message to be present on the wires. Or, in other words, about four thousand messages like this one could be transmitted in only one second on a network made of two simple twisted wires. This is why I need to use a scope (and a high-quality one, by the way) to diagnose CAN bus problems: it is not possible to see a difference in good and bad signals with a multimeter or a test bulb (connecting a test bulb to the CAN network will actually just shut it down; it is the same as shorting the network wires to the ground). Or even with a cheap scope with a low sample rate. Here is another waveform, taken from a different CAN bus operating at a slower speed. Each of those short signal peaks actually the same messages as above, only at a different zoom level (zoomed out 1:500 here as opposed to 1:1 on the previous screenshot). This waveform fragment is about 200 milliseconds long - one-fifth of a second - and yet you see 21 messages got transmitted over this short time. The messages here actually represent real-time data going from the TIPM module to a scan tool when it reads live data from the module. These are not the messages that onboard vehicle modules exchange with each other; those would have much higher density and therefore would be less "visual" in this example
Now here is the waveform on the CAN INT bus on this Dodge Nitro
The modules connected to the network can't recognize this mess as valid messages and therefore they can't communicate with each other. Therefore nothing works properly on that network anymore.
The next classic step, in this case, would be disconnecting all the modules connected to this network, one by one until the signal restores to normal. If after disconnecting some module signal gets restored to normal, then this module will be pronounced faulty. Sounds easy.
This ain't a small job, though! The units on this network are all around the vehicle interior: instrument cluster, HVAC control module, radio, both front door modules, audio amplifier, DVD screen - if there is one...
It will take a lot of labor just to get them all located, removed, and disconnected. Some of the modules also might be shown on a diagram but not actually present on a vehicle. Therefore extra care should be taken to ensure that all the modules got verified for their presence or absence on the vehicle. For this reason, I decided to start first by checking the most likely locations for the wiring problems - front door wiring harnesses and their connectors. This is where wires go from the body to the doors, and because of the constant bending back and forth over the years the wires here can get broken or shorted to each other.
After some fighting with the incredibly tight placement of those wires and connectors, I found some corroded pins on the passenger side. Nevertheless, disconnecting both door harnesses didn't restore the normal network signal pattern. That sucks... The next step was a bit radical: I cut the network wires at the TIPM connector and checked the network signal on both sides of this cut. The signal indeed restored a normal pattern on the TIPM side of the cut. Therefore, the problem must be on the other side, not in the TIPM. Meaning that proceeding with all the work of disconnecting all the network modules is inevitable now...
Removing all the seats, center console, center dash panel, instrument cluster...
...trim panels along the door openings, and the big right trim panel in the trunk - because this is where the audio amplifier is and I need to get to it to try it disconnected (some other modules if present would also have been installed close to this location).
Discovering along the way lots of rust, moisture, and dirt under the floor mats where massive wiring harnesses are located. A couple of very bad-looking ground connections were found there too...
Holy mackerel! Any of this shit can cause a problem; where should I even start here?
Disappointingly, after all this work to get to the control units, disconnecting any of them did not make any difference! Only if ALL of them got disconnected at once was making the signal go back to normal. But once I had any of those units in any combination re-connected to the network, the signal was going haywire again...
Wiring issue then, maybe? All this moisture and rust around the wires must have done some damage. Measuring the resistance between the network wires hasn't shown anything wrong or unusual. There was no short circuit between the wires nor a short circuit to ground or power. Therefore it is time to start stripping the wiring harnesses in those wet and dirty locations, hoping to find some corroded splice or some rotten broken wire.
No luck here either! The wires were dirty but not damaged anywhere. I also fixed the suspected bad grounds under the driver's seat by connecting them with a jumper cable to a good ground: nope, this made no difference either.
I had that other TIPM here too, the one the customer tried to install. And while that unit didn't even allow for turning the ignition, the network signal with it remained OK regardless of whether the other modules were connected to the network or not. What if the original TIPM itself is at fault? What if something failed in the TIPM module in such a way that it can't support network operation once there is some other "load" connected to the network?
I disassembled an original TIPM as much as I could and checked its internals for any sign of moisture, corrosion, burning: any damage. Nope, nothing there...
The new TIPM module costs $2500, by the way. Tough call. What if my suggestion about TIPM's internal fault is wrong? The best advice I could come up with at this moment was to try another used TIPM, trying to get it from the same year, same model, same trim level car. Maybe it will work...
The worst thing here was that while I was doing all this work, the car stopped operating the starter, and the code came up regarding the loss of communication with the Vehicle Security Module. Damn, now it looks like I made this even worse than it came in with! At least the engine was running before, now this damn thing is completely dead! Oh, one of those days when I hate my job...
By the time another TIPM have arrived, I came up with another idea: let's try to check the network signal at the TIPM and at the same time in a few other locations of the CAN INT bus. Maybe there will be a difference in signals after a certain location, and maybe - just maybe - it will help me to find a wiring fault somehow?
Indeed, there has been a visible difference when simultaneously comparing signal patterns taken from different spots. And finally, I got lucky: I found the broken wire! It was under the front passenger floor mat. The wire got rotten and broke inside the insulation without any external sign of the damage to it. Probably some microscopic crack or a pore in wire insulation allowed moisture inside the wire and caused copper to corrode creating extra resistance in the network. I made a temporary connection at the point of breakage and verified that CAN INT network signal now remained normal with all the modules connected. Then I cut out the corroded section of the wires entirely, made a repair insert with twisted wire pair, and soldered it in.
That is some success, but the car still doesn't start, and I have to look into that now...
I see there is no communication now between TIPM and Vehicle Security Module, Airbags Module, and Occupant Classification Module. They are all connected to another network, the CAN C network, along with a few other modules.
Is there any common circuit for these three modules? Common fuse, common ground?.. Nope, they are all separate, nothing common except for the network connection. Let's use the same method of comparing network signals I just used on the CAN INT network. Whoa, there is a clear difference between signals: they do look healthy but they are different as if they are taken from two different networks altogether. So somehow, the part of the network got separated from the rest of it.
It took some time to trace the network wires again. The fact that network wires changed colors along their course and it wasn't reflected by the wiring diagram that I used added to the time and frustration, for sure! But thanks to the wiring harnesses being already all stripped and laying loose, I traced the network wires to the point of the break relatively quickly. Guess what: apparently, you can't start Dodge Nitro if the passenger seat connector is disconnected! Because some genius designed the network cables to be joined inside the seat wiring harness! For real, dude?! What was the thinking here? And why it is not reflected on the wiring diagram? The thinking here actually was likely a limitation to the "branch length" - the length of the piece of wire branching from the network to the module. Still, the things like this should be shown on a diagram!
Put the passenger seat back on, connected, and - voila - it is alive! Engine starts, codes cleared, lights are gone. Finally!
Do you think it is all done now? Not yet. Now all the stripped wiring harnesses need to be wrapped back in tape, secured with proper cable holders and routed in the right way, so they won't get damaged again. I also restored bad grounds by re-routing them to a different ground connection point in a nice dry location. Cleaned all the rusty debris, re-assembled all the removed interior parts, seats, floor mats... Lots of work again. And finally, it is ready to be billed out.
And here is the worst part: believe it or not but I've spent 16 hours looking for that one broken wire! This is the reason why I can't provide any labor time and cost estimates for the diagnostic jobs. The time that might be required to find the root cause of some mysterious problem is totally unpredictable. Previous experience doesn't really help much here. I had similar issues with network communication faults, and the typical pattern is "check for the moisture under the floor mats, if there is water, remove and dry the mats, fix all the rotten wiring you can see and the problem is fixed". This time though it was not so typical and took much longer to find a problem. Plus it involved the unnecessary cost of a module that got purchased with no actual need for it (good thing it wasn't a new and expensive module!) I guess I could be more efficient if I would deal with the electrical diagnostic problems like this all the time, without distracting myself with routine jobs like removing the engines or replacing wheel bearings. Unfortunately, the market in a small city like Grande Prarie, AB wouldn't support my business enough should I decide to specialize exclusively in electrical diagnostics. Or at least this is what I think at this point.
Comments