As technicians, we rely heavily on our scan tool. DTCs, output controls, resets, special functions, and datastreams guide our decisions on how to diagnose a problem. When we connect the scan tool and get the dreaded “No Communication” screen (figure one), the resulting frustration can often be heard across the shop. But what if there existed a simple process to diagnose these communication issues? Before you decide to punt the job to a shop down the street, realize you can do some basic checks to narrow down this simple electrical circuit and maybe make a little profit on the job. But first, to properly diagnose a circuit or component, we need to know how it works.
There are two types of Controller Area Networks (CAN): high-speed (sometimes called “chassis” or “fast”) and low-speed (sometimes called “body”). The high-speed CAN is the network that connects modules needed for powertrain control, such as the engine control module (ECM), transmission control module (TCM), and electronic brake control module (EBCM), to name a few (figure two). The high-speed CAN is a two-wire network with each wire identified as CAN+ and CAN–. The CAN+ and CAN– circuits are connected to each module in the network. They are also connected to two “terminating” resistors that are 120-ohms each. They’re connected in parallel and at each end of the CAN circuit. If you remember how parallel electrical circuits operate, you’ll know that those two 120-ohm resistors will act like one 60-ohm resistor connecting CAN+ to CAN-. These terminating resistors are found in a module, such as the ECM, or they can be connected into the harness as a standalone resistor. If you check between the high-speed CAN+ and CAN- circuit with an ohmmeter (assuming there’s no voltage on the circuit), you should see pretty close to 60 ohms, like in figure three. See there! That’s the first simple test you can do on a high-speed CAN.
To send a message on the high-speed CAN, the modules shift the voltages on the CAN+ and CAN- circuits. When no messages are being sent, the voltage on both wires is about 2.5 volts. When a message is sent, the CAN+ drives up a volt to about 3.5 volts, and the CAN- pulls down a volt to about 1.5 volts. These messages are sent relatively fast (around 215 microseconds or 1/5000 of a second), so you won’t see the full voltage changes on a digital voltmeter unless you have a meter that can record min/max at 1ms intervals (like the Fluke 87). But that doesn’t mean that the voltmeter is useless. With a voltmeter, you can still see if the voltage on each CAN circuit is around 2.5v. And when messages are being sent, you’ll see the voltage change a bit. The level of change is based on the number of messages being sent. This is because the voltmeter averages the voltage, so when a message is being sent, it only changes the voltage a little bit on average. This causes a slight increase in voltage displayed on the voltmeter. On our example vehicle in figure four, you can see between the CAN+ and ground, we measured 2.7 volts, and between CAN- and ground, we measured 2.3v (figure) five. How about that? If all you have is a DVOM, you can still use it to do some simple checks, resistance, and voltage!
There are various ways that the CAN bus fails. The most common failures are poor connections, opens, short to power, and short to ground. A module could also fail and bring down the whole CAN bus by shorting it to ground or power. If you’ve lost communication to only one module, it could be the module, a poor connection, wiring issues, or a power/ground issue at that specific module. In this case, you can typically communicate to other modules on the network, and they’ll likely have stored DTCs indicating communication failures with the module in question.
Don’t forget to check service bulletins related to your symptom. Regarding the problem that we’re going to describe later in this article, GM has an excellent service bulletin, #08-07- 30-021H, which identifies common issues resulting in CAN bus failure and no communication DTCs. The bulletin is 38 pages long and includes pictures and examples of wiring and connection failures. To summarize the bulletin, GM breaks down the failures into the following:
Backed out terminals or loose connections:
- TEHCM connections
- ECM connectors x115 and x109
- Fuse block connector X1
Inspect the ESC module connection for missing weather plugs and water intrusion
Opens and shorts in the:
- engine harness between the right front brake pipe on the left front frame rail
- wiring harness by the park brake pedal
- instrument panel harness at the left-side junction block mounting bracket
- wiring harness at adjustable pedals motor
- wiring around the transmission
- wiring between the engine and the intake manifold
- wiring harness where the chassis body mounts on the left side frame rail
- wires at the rear terminating resistor
So now that we know how the high-speed CAN functions, let’s put some of this knowledge to work. We received a 2009 Silverado 3500 2WD with about 72,000 miles on it for a transmission issue with some interesting electrical issues. This truck had a dump bed and a snowplow attachment, and it had a bit of rust/corrosion across the chassis. The following was our diagnostic process:
Review the work history:
The shop that decided not to work on it and punted it to us. They provided the work order and repair history. They originally diagnosed it as a no-start condition. The starter relay was replaced, and the engine started and ran fine. Intermittently, the transmission would “shift strangely,” and the instrument cluster would work erratically. When attempting to retrieve DTCs, the tech was initially only able to pull low-speed (body) related DTCs, and he couldn’t communicate to the high-speed (powertrain) modules. Once the vehicle sat for a bit, the tech was able to retrieve powertrain DTCs, all of which were mostly lost communication codes between modules. The truck also had a “service trailer brake system” message on the dash.
As luck would have it, once we received the vehicle, the no start/no crank issue returned, and there was a “no communication” message when trying to communicate to the network with a scan tool. This was actually a good thing! The problem was present, and if we were careful, we could narrow it down to a component or a specific point in the circuit. Careful is the keyword here. Manipulating connectors and wires could disrupt the problem and cause the intermittent to go away.
Check for bulletins:
As mentioned before, there is a great service bulletin related to this exact problem. The only issue with the bulletin is that it doesn’t define any diagnostics, just some common areas that could cause network failure. We didn’t want to start searching for these areas in fear of disturbing the components and possibly causing the failure to disappear. We decided it would make sense to narrow the problem down as much as possible and then refer to the bulletin for common areas of circuit failure.
While searching for bulletins, we also found a campaign on this vehicle to replace the trailer brake module. The campaign described many of the same symptoms we were experiencing on this vehicle as well as the network failure and DTCs. Once again, before jumping to parts replacement, we decided to do some simple tests before touching any connectors or harnesses.
First Diagnostic Test:
Our first non-intrusive test was to check the resistance between the DLC pins 6 and 14, which are the CAN+ and CAN- circuits. We found 120 ohms (figure six), which indicated that one of our terminating resistors wasn’t making the connection between CAN+ and CAN-. Now was a good time to review a wiring schematic of the CAN system. A wiring schematic is necessary to:
- Determine what modules are on the network and how the network is configured.
- Determine where in the circuit the terminating resistors reside.
- Determine where the diagnostic link connector (DLC) splices into the circuit. This will help in dividing the network for the process of elimination.
While reviewing the schematic in figure seven, we can see that each module (except for the ECM) has two CAN+ and two CAN- circuits. GM does it differently from some manufacturers by running CAN+ and CAN into a module and then delivering it back out to extend to another module. The ECM only needs one CAN+ and one CAN- , because it’s located at the end of the circuit. The schematic also shows that the ECM has one of the terminating resistors and that the other terminating resistor is found in the harness after the trailer controller.
Because this truck has a campaign for the trailer brake module and thinking that this was the source of our problem, we decided to bypass the trailer module by disconnecting the connector and jumping the CAN+ and the CAN- circuits (figure eight). By doing this, we effectively eliminated the trailer brake controller from the network. After bypassing the trailer brake module, the resistance from pins 6 and 14 dropped to 61 ohms! So problem fixed, right? To our surprise, there was still a no-start and no communication problem amongst the high-speed network modules. The trailer brake controller (or connections) obviously had problems, but it wasn’t the only problem.
Second Diagnostic Test:
Next, we checked resistance between CAN+ and ground and CAN- and ground and found an issue. CAN– to ground was 60 ohms and CAN+ to ground was 1.5 ohms. Both of these measurements should have returned “out of limits” (OL) or very high resistance values. Evidently, there was a short circuit on the CAN+ wire. Remember that nice GM bulletin? One might think this would be a good time to refer to that bulletin to check for cut or chafed wires in those common locations, but we wanted to divide the circuit up to narrow our focus.
While referring to the wiring schematic, and since we were checking the circuit at the DLC, if we disconnected the BCM and the short went away, the problem would be on the ECM and TEHCM side of the schematic. If the problem was still there, then it’d be on the other side of the schematic. We picked the BCM because it was relatively easy to get to. When disconnecting the BCM, the ohmmeter measuring between pin 6 at the DLC and ground indicated that the short did, in fact, go away, and reconnecting the BCM brought the short back. That simple test just verified that it was the ECM, TEHCM, BCM internal short, or wiring. Unplugging the ECM didn’t remove the short, and while attempting to unplug the TEHCM, the ohmmeter indicated that the short went away. The short circuit involved the TEHCM harness near the conduit clip on the transmission, which was one of the areas identified in the bulletin and can be seen in figures nine and ten.
Repair or replace?
According to the service bulletin, GM warns the tech, “DO NOT replace a module or harness” but repair the circuit instead. We all know that time is money and that diagnostic time for finding and repairing a circuit is unfortunately poor and typically not tech-friendly. Plus, if a communication wire that requires virtually no current flow has failed due to a poor connection, open or short, how confident would you be with the rest of the wiring in that harness. For the price listed on repairlinkshop.com (figure 11), the engine harness is not overly expensive to the customer, and the billable labor and list price mark-up make the job a profitable one.
Understanding how a system works is key to diagnostics, and that applies to just about everything. With network diagnostics, you need good wiring schematics and good connector pinouts. While a scope is nice, it’s not necessary for most CAN failures, and a DVOM will identify most CAN electrical failures. Realize that some vehicles have multiple high-speed networks, and the DLC pins 6 and 14 might not actually tap into the network directly (once again, read the wiring diagram). When your tests indicate a problem, consider isolating the circuits by process of elimination and consider bypassing suspected modules if possible. Just keep telling yourself, “It’s only two wires,” and you CAN do it!