The financial demographics of the customers we serve have often impacted my diagnostic strategies. A lot of the vehicles we service are overwhelmingly neglected because budgets are tight. Where I work, it is almost an everyday scenario that a customer's Check Engine light concern can often reveal 10 to 15 different trouble codes stored in various ECUs. These are not "ghost" codes as we sometimes call them, these cars are that broken.
Network diagnostics have often troubled my diagnostic mindset because I'm too quick to assume that if there are multiple codes, there must be multiple problems. When I finally identify a problem and another exists, I have gone off on another diagnostic plan only to find the same problem I found earlier. This detour has often resulted in a lot of time wasted.
Let me give you an example.
Enter The Mercedes
A 2001 Mercedes CLK430 with 130,000 miles and a 4.3L V8 engine showed up at our shop with the concern of an intermittent parasitic drain. Yay!!! Since it came in on a Monday, I thought to myself, whoever worked on this one would still be pulling their hair out the following Monday.
130,000 miles
4.3l v8
Automatic transmission
Original concern: parasitic drain
I dodged a bullet on the first concern…a battery drain. Josh, one of my teammates, got dispatched the initial check, so I was off the hook for now. We had discussed some of the latest methods being used for testing. We knew not to disturb the fuses because this can cause erroneous test results.
When this vehicle went to sleep it ended up pulling about 0.03 amps. This car is old, very broken in a lot of different ways and there are a lot of fault codes. Should we consider the DTCs and, if so, which ones do you chase for a parasitic drain that is intermittent? In my experience every situation is different. We may have a series of CAN faults across multiple modules. We may have only one fault in a single module. One, both or neither may turn out to be the cause of the drain.
Luckily, this intermittent parasitic drain revealed itself in other problems that we could duplicate. The windows and the power seats stopped working while my coworker pulled the vehicle out of the shop. Since we spent our initial diagnostic time on a draw we could not duplicate, we had to sell more tests to the customer to track down the window and seat issues. They approved the additional testing and the diagnosis got re-dispatched to me.
Focused on the search
My coworkers know that any day of the week I will trade ball joints and brakes for some diagnostic work. Sometimes I'm a glutton for punishment but mostly I love the challenge. So I get started on this vehicle with an auto scan using an Autel 908. What I can see right away is that I am missing a lot of computers on the network (Figure 1).
I’m not fully sure I know every module that should be there but right away I know the seat modules, which are common causes for parasitic drains, are not talking with the scan tool. My first step in any diagnosis is to check fuses as long as the concern can be caused by blown/missing fuses or improperly placed fuses. I found none of the above. My next step is to start building a network topology. What I end up finding with a little research is that there are multiple networks on this vehicle. I also find out that the interior CAN bus cannot be accessed from the DLC. The seat modules are on the interior CAN along with the front Signal and Acquisition Module (SAM) and several others (Figure 2).
Knowing this is imperative to the next step. We need to ask ourselves what modules on which networks can talk to the scan tool? What I have found at this time is the front SAM is the only module on the interior CAN of the gateway that can talk with the scan tool. This is important because we can, therefore, conclude that our CAN bus wiring is good to the front SAM and whatever the problem is must be after the SAM. With this knowledge in hand, I can funnel my test methods to get the most accurate results.
One article simply cannot do justice on tackling multiplex diagnostics. There are far too many faults that can be present in a bad network and they all deserve a thorough explanation. The technique that I chose to use in my diagnostic test plan is not complex, but it was tactical. After analyzing the network topology I thought I may have open wires after the front SAM. I thought this because a short to ground or voltage will pull all the modules down. A bad ground or power feed to a CAN bus module can affect the whole bus. This is the reason I thought I may have an open circuit.
If you have ever performed a voltage drop test with a multimeter you can do this test with a scope on a wired vehicle network. The idea is simple enough. Place one channel probe at one location where the scan tool can communicate with a module and the second channel probe at a different location that scan tool cannot talk with a module. When the vehicle is working normally, the bus data packet transmission (the pattern you see on your scope) should be identical. When a wire is broken or there is a high resistance issue you will have either mismatching data packet transmissions or experience voltage differences between the two channels. In the case for this vehicle, I placed my scope leads at the front SAM for CAN Hi and LOW and the next set of leads I placed at the seat module for CAN Hi and LOW. Here's what I saw (Figure 3).
What we can see here are two traces that look similar and the other two that are not. If you are familiar with the way a CAN bus looks like, you may also point out that it looks nothing like any bus transmission you've ever seen. CAN bus LOW has different bias voltages and the patterns do not match at all between the front SAM and the seat module.
What's wrong with this picture?
This is where I should have stopped and refocused my efforts in finding the open/high resistance circuit. What I did instead was to sit and analyze the other CAN bus HI. I called my good friend, Pedro De La Torre, and said, "Hey, look at this pattern for me, I know there is a high resistance issue on CAN bus LOW but why is my other bus looking so messed up?" I started to suspect two different problems instead of finding the answer to my first problem.
Luckily for me, I didn't go too far off into left field. I experimented. Well, since I knew my open is after the SAM, what can I say about the problem with my CAN HI? What can I do to see if the bus can look normal? I cut both bus wires after the SAM and recorded the network out of the front SAM. To my surprise, (Figure 4) that looks like a CAN bus to me. This now proves to me another way that both problems are after the front SAM. Now on to finding the problem in the vehicle.
My plan of attack once I've gotten an idea about what the problem may be is to divide the network at the easiest point. I want to break up the network right down the middle so I can isolate half of the modules to see which side the problem may be on. On many vehicles, multiple splice packs house all the different busses. Sometimes these splices are easy to get to and sometimes they take hours to get to. This splice pack on this Mercedes was in the front kick panel below the carpet. This provides an excellent area for moisture intrusion to take place and corrode the pack. I located the splice pack and (Figures 5, 6) I found the source of both of my problems...the broken bus wire and CAN bus that is shorted together. I said I get lucky sometimes and this was my day!
Since I got lucky I want to explain what steps I would have done if the problem was located elsewhere in the vehicle. After gaining access to the splice pack I would remove the wire from the splice pack that comes from front SAM and load test the wire. Since we are removing the circuit from its voltage potential and it is not being stressed in any way using an ohmmeter, I would use an applicable bulb to make sure the wire can carry current. Whenever you load test a circuit, you need to know how much current the gauge of wire you’re testing can handle so you don't smoke a wire or get inaccurate results. Not putting enough electrical strain on the circuit may cause you to not detect a voltage drop issue. For example, using a headlight bulb to load test a starting/charging circuit is not going to cut it. If the problem wasn't located between the splice and the front SAM I then move farther past the splice to whatever module I can easily access till I find the source of high resistance.
Now that I've found the problem, I have to fix it, of course.
