In the end, it's all just voltage

I often get calls from shops struggling with ETC faults and I was called to look at this truck after it had been setting throttle body codes and running rough. When I arrived, I found the shop had already installed a brand new throttle body but the same symptom was present so, they reinstalled the original unit. A code P2100 set almost immediately (when the ignition was turned on) indicating a fault with the ETC motor (Figure 1) The engine would start but run rough.

Before I dive too deep into this diagnosis I am thinking about what the most likely causes of this concern may be, using my knowledge of how most electronic throttle body systems operate. Most of these “fly-by-wire” or TAC (Throttle Actuator Control) assemblies operate similarly in that the ETC motor is wired directly to the PCM. And, the PCM controls the position of the throttle plate by changing the polarity of applied PWM (Pulse Width Modulated) voltage to a DC motor, connected to the throttle plate. If a voltage is applied to one terminal of the DC motor and, ground to the other, the motor will turn in one direction. When polarity is reversed, the motor will turn in another direction. As a safety feature to prevent a runaway-engine, the throttle plate is held to a closed (or almost closed) position by a throttle return spring. This means that the PCM is typically only opening the throttle plate. But the PCM doesn’t need the throttle plate to be slammed-open or closed, it needs to control the position of the throttle plate. This is accomplished by pulsing the voltage of the ETC motor (PWM). By changing the PWM duty cycle, the PCM can have precise control over the position of the throttle plate.

In addition to the motor, most throttle bodies have two TPS (Throttle Position Sensors) that read the throttle position (if the PCM doesn’t know where the plate is, how can it control it?). There are two sensors because of the need for redundancy. If there was only one sensor, and it became skewed, it may show the throttle closed when in reality it was open. The PCM (thinking the throttle was closed) would operate the ETC motors, opening the throttle even more! This type of situation could cause an engine to rev to WOT (Wide-open throttle) while a customer is attempting light throttle. Of course, this could potentially cause an accident.

By adding a second TPS sensor, the PCM can compare the values from both sensors and determine if they are both operating correctly, and accurately. It is important to note that both TPS sensors will read different values at different throttle body positions. So, if somehow both TPS signal circuits were shorted together, the PCM would be able to determine that the signals were not correct (they should never share the same voltage at any given throttle angle).

This same principle that applies to redundant TPS sensors also applies to APP (Accelerator Pedal Position) sensors. Depending on the manufacturer, the PCM will have some type of strategy that will put the vehicle in “limp home mode” so, the engine can’t rev past a predetermined level. This is in case the throttle plate were to stick open or if a sensor failed. “Limp-home” is often achieved by disabling cylinders. I have seen plenty of technicians chase a misfiring engine but, the misfire was intentionally caused by the PCM, to regulate engine speed, due to a malfunction in the ETC system.

The ETC system and its components
We can break almost all electronic throttle control systems into four main components:

-The throttle body assembly (which includes the ETC motor and TPS sensors)

-The APP sensor

-The wiring

-The PCM

Because the PCM must maintain control of the throttle plate at all times, the circuits are typically not shared with any other components (i.e. the 5v reference for the TPS sensor is not likely shared with the MAP sensor). In the case of this Jeep however, the 5v reference circuit for the TPS is shared with other sensors (Figure 2). What this means for us is that although ETC diagnostics is usually isolated to the four above-mentioned (potentially failing) components, we need to consider an issue with the other sensors that share the 5v reference, with the TPS sensors.

With that knowledge in mind, I began to think about what tests I need to perform first. I decide not to focus on the APP sensor (because there are no APP codes). Since the TPS sensor/ETC motor are all part of the throttle body (and we had one nearby), I tossed it on and verified that the code did reset, right away, with the new unit. Although possible, it wasn’t likely that both the new and original throttle bodies were causing identical issues. I also decided not to focus on the TPS signal or 5v reference circuits (because I would expect the PCM to set TPS related fault codes if there was a TPS related concern).

Narrowing the area of focus
At that point, I was focused on the PCM and the wiring between the PCM/ETC motor, as potential causes of this symptom. I researched the P2100 code and found the diagnostic procedure directing me to check the resistance of the circuits with an ohmmeter. As I expected, two circuits connect from both poles of the ETC motor directly to the PCM. I typically don’t like to use an ohmmeter to test circuits. In the case of this ETC motor, an ohmmeter can only check continuity of the circuit. It is not able to test the circuits' ability to carry the amount of current required to operate the motor. But, because the code would set immediately with the key turned on, I was expecting to see an open circuit (or very excessive resistance if the wiring were faulty). I quickly checked both circuits with my ohmmeter, while wiggling the harness but found no excessive resistance, whatsoever.

At that point, I was ready to condemn the PCM as faulty but,  I first wanted to look at the voltage on the ETC circuits and see what (if anything) the PCM was doing. Using two channels of my scope I connected to circuit K124 and K126 (the ETC motor circuits) at the throttle body and referenced my ground lead directly to battery-negative (Figure 3). I then cleared the code and cycled the key a few times while recording. I want to point out that I cleared the codes (which may or may not have been a critical part of this test). Service information does not tell us exactly what happens when this code sets. It might be safe to assume that the PCM may remove power from the circuit. If it detects a fault (meaning if the code is set), we will not see anything on those circuits with our scope. On the other hand, P2100 is a continuously-monitored code and, because the key must be on to clear the code, we may find ourselves in a catch 22 of trying to record the exact-instant when the failure is present. In our case, we were able to catch at least some sort of activity, and recorded this screenshot just as we keyed the ignition on.

Looking at this screenshot we were confident that it wasn’t correct (Figure 4). Since there wasn’t anything else on the ETC circuit, it was easy to condemn the PCM. As a rule of thumb, we never condemn a PCM without testing the power and grounds. And by testing, I mean load-testing — and we’re about to find out why load-testing is so important.

Flushing the fault to the surface

Many years ago, I was taught to use a sealed-beam headlight (to test power and ground circuits) rather than a voltmeter or a test light. The reason for this is that just because a circuit can carry enough current to light up the tiny filament in a test-light bulb. It does not mean it can carry enough current to operate the component on a vehicle. Consider the multiple strands of copper wire in the heavy gauge B+ wire, like the wire supplying the starter motor. Let’s say that the wires at the battery terminal (connected to that starter wire) are severely corroded, so that only one strand of wire is making contact and connecting the circuit. If we were to use our voltmeter (with one lead connected to the negative battery post and the other at the at battery positive wire of the starter) at KOEO we should see battery voltage present because, there is one strand connecting the circuit. But that one-strand is not capable of flowing enough amperage to crank the starter. The same principle applies to a test light. Just because that circuit shows 12v available (or illuminates a test light), does not mean it can do the same, under load.

And there are many different ways to load test a circuit. One common and easy way to load test a circuit is by using a voltmeter (while the circuit is loaded). In the case of our corroded starter wire, that would mean checking the voltage while cranking the engine. We may see 12 volts with the starter at rest and 4 volts while cranking the engine. This is because the single strand of wire (resistance in the circuit) is using 8-volts, of the available voltage and all we have left at the starter is 4-volts. You may also notice that the portion of the wire, connected by the single strand, gets hot or maybe even “glows!” That’s where the voltage is being used, in the same way a light bulb or heater element converts energy into light or heat.

I, on the other hand am lazy and for me, using a sealed-beam headlamp (to load test most circuits) is typically quicker. This is because a sealed-beam headlamp will draw enough current to load most circuits on a vehicle. It certainly won’t work for a starter circuit but for most modules and low current devices, if the circuit can light up a sealed beam headlamp, it can certainly carry enough current to power a module. If a circuit is not healthy, it is easy to tell. The bulb may visibly illuminate but, is dim. This comes in handy especially, when you have to cycle a key inside the vehicle, while performing a load test outside.

So, before I condemn the PCM on this jeep, I want to print out my wiring diagrams,connector views and test all of the powers/grounds with my sealed-beam headlamp. I typically like to spend a few minutes studying the wiring diagram and using multiple colored highlighters to highlight the pins I want to check, for either power and ground. In this instance, I needed to check circuits F1, F42, A209, and F942 (for power) and circuits Z130 and Z816 (for ground). It is important to point out that I am doing this testing with the PCM disconnected. Circuit F42 gets its power from the ASD relay and the ASD relay is energized by the PCM (when PCM supplies the ground path). It grounds circuit K34s which means we must ground circuit K34s at pin 3 of PCM connector C3 to test for a healthy power on circuits F24 to the PCM (Figure 5).

As we are going through the testing I found all of the circuits to have healthy power and ground. They light up our sealed beam brightly but, when I got to circuit F42, we had no illumination (Figure 6). I double-checked that I had the correct wire grounded at the PCM, to close the relay and could even hear it clicking when I cycled the ground. I checked and found 12v at that fuse (with the relay closed and the sealed-beam headlamp disconnected) but, as soon as I connected the sealed beam the voltage dropped to 3.6 volts, on both sides of the fuse. This told me that we did have a voltage drop and, that it was upstream (of the fuse) in the circuit, towards the relay. Had we only checked for voltage at the PCM connecter (either with the key off or the PCM disconnected), we would have missed this voltage-drop because we needed the circuit to be loaded, to see it.

With that information, I decided to take the testing to the relay and installed a relay-testing adapter. I found (with the headlight disconnected) I had B+ voltage on terminal 87A (the output going to fuse #16) (Figure 7). When I reconnected the headlight (loaded the circuit) I measured the same 3.6 volts at terminal 87A (Figure 8). Next, I checked and found healthy battery voltage at terminal 30 (the supply power for the switched-side of the relay) (Figure 9). This indicated that the relay was using up almost 9-volts of the available 12.5-volts. What I didn’t mention is how I already knew that since it burned my fingertips when I removed it!

Retracing my steps
So, going back to the original scope screenshot (that we captured during the failure), we were able to identify exactly what was happening and how the failing relay was able to cause the code to set. We can see that the PCM is supplying 12-volts to each side of the DC-motor but, as soon as it grounds the circuit, (our yellow trace) to operate the throttle blade in one direction, the voltage on the power circuit drops to around 5-volts. From there, you can see the PCM trying to flip the polarity. But, the waveform does not look clean and the voltage stays around 5v. After replacing the relay, we can see in our known good waveform that the voltage will switch (on and off) with crisp transitions and, the voltage stays steady at B+ voltage (Figure 10).

The voltage needed to run the ETC motor for this Jeep, was supplied by the ASD relay. And because the ASD relay was not able to allow enough voltage through, the ETC motor failed to control the throttle plate. I was unable to find specific code-setting criteria but, we can assume one, of two things caused that specific code to set:

- Either the PCM was expecting to see the TPS readings change (when controlling the ETC motor), and because there were no correlation issues between the two sensors, the PCM flagged the fault in the ETC.

-Or the loss of voltage in the ETC motor circuit under load was detected by the PCM as a circuit fault, which is specifically what the code describes.

Either way, any technician could have easily made the same diagnosis. With a procedural diagnostic approach, a little understanding of how these systems operate, and an understanding of voltage drop. I didn't use much aside from a scanner, scope, a voltmeter, and a headlight bulb to figure this one out. When you consider the fact that you could probably find all of the necessary tools second-hand (to complete just this one diagnostic) , for less than the cost of a new PCM /throttle body for this Jeep, it makes you wonder why a large percentage of shops and technicians aren’t willing to learn and apply these techniques.