During the initial diagnosis of this customer's concern, the original technician could hear compression leaking from the vehicle when the engine was running. At first, he suspected an injector seal to be leaking, creating a density misfire. A smoke machine was used in the cylinder to identify a potential compression leak external to the cylinder. After the smoke machine was hooked up, smoke was present in the intake manifold on TDC compression. A mechanical problem was now suspected but being a GDI engine, this test was not conclusive enough to rule out a carbon build-up issue. The customer authorized the shop to tear the intake manifold off to identify the potential problem. Before the tear down was done, I suggested we prove another way what the other technician was seeing with the smoke machine. These tests and analysis were performed in order to prove what was wrong with this vehicle before expensive tear down was done.
Get a routine, follow your routine and back your diagnosis
As a technician that works on a lot of vehicles that have been to multiple shops, I have to stay centered on my diagnostic routine. Being centered on a routine first means we need to have one. As I have grown stronger in my diagnostic strategies, it has helped me funnel my testing in a way that is logical and produces results. Earlier in my career I simply started with eliminating possibilities until the answer was found.
While deductive logic is the foundation of how we all work it should not be the sole method. Strategy based diagnostics start with the scan tool more often than not. This article isn’t based on scan data strategy but that is where this diagnosis started. Currently I am funneling my testing methods so every test is justified by the previous test’s results. My very first tests are going to start with the easiest and least time consuming to rule out the most possibilities. When it comes to a misfire, the scan tool can offer a big funnel when it comes to weeding out fuel, ignition or mechanical issues.
Once I have made a conclusion that this is likely a mechanical problem, I will stop testing for ignition and fuel and focus on what the data has led me to believe too far.
Starting my routine
I’m constantly learning new techniques to better understand/diagnose the systems that I’m working with. Luckily for misfire concerns I think there is a fundamental first test that we all should think about using to gain a direction as to what our next test should be. This first test for me is secondary or primary ignition analysis. Secondary and primary ignition analysis has not changed much since scopes were first used on the internal combustion gasoline engine. Certainly obtaining these waveforms has gotten a little more difficult over the years with the invention of the COP coil and transistorized coil packs.
However, the fundamentals of how ignition systems work is the same. Power and ground is provided to a primary coil with few windings. Once the ground side is released, the collapsing magnetic field surrounding the primary coil is induced into the secondary windings. Since there are more windings in the secondary coil, voltage is increased significantly. Now a high voltage in the secondary winding follows the easiest path to ground which we all hope is in the combustion chamber near TDC compression.
Utilizing ignition waveforms allows a technician to look at compression spark and fuel all in the same test. This gives us the quickest and best diagnostic direction for additional tests. i say direction because like I stated earlier in the article I’m not willing to condemn any one part/parts off of one test. I’m going to use my results from the secondary pattern to make a hypothesis on what is going on and what my next test should be. So I grab my newly purchased COP wand and start looking at this misfiring vehicle. As I go down the line of coils the pattern on cylinder 2 has a repeatable event in the waveform at park idle no load (Figure 1).
Making a hypothesis to funnel future tests
Performing secondary analysis is tricky sometimes. You really have to trust your equipment and hope what you’re seeing isn’t noise from a multitude of other contributors. One of the best ways to trust your pattern is to get primary ignition. With primary ignition you are essentially hard wired into the primary circuit. You can trust that what you’re seeing is true. However, we don’t always have the time or ability to get there with transistorized cop designs. If we are going to use secondary for analysis, get to know the good cylinders first before going after the one you suspect to be bad.
As we look at the pattern (Figure 1), we see the point of primary turn on. We see some coil oscillations and then a rise in voltage in a triangular shape to the point where primary turns off. When the primary driver is released that is where the secondary current flow begins. We have a firing line and then a burn line. As voltage continues across the plug we see little blips /rise in voltage. This can happen because of the presence of turbulence.
When at idle the combustion chamber is rather stabilized compared to high load/high cylinder volumes. When we see this sort of behavior at idle we can draw some conclusions as to where our next testing should go. In order to make these oscillations in the burn line there needs to be movement or airflow across the spark plug. When a cylinder is sealed, the volume is being condensed but it is not blasting past the spark plug because it has nowhere to flow. When you have an existing path for the movement of air - such as a leak - then it can interact with the electron field and create higher resistance for the path of ground momentarily. If the voltage goes up and down like this it means that the energy to cross the gap is going up and down. The spark is quite literally getting blown apart. This type of analysis needs to be done at park idle no load. It also should be a repeatable pattern with a dead misfire like I have. As of right now I’m starting to suspect a sealing issue with the valve train. This test does not identify what is leaking in the cylinder but it does provide me with a direction for additional testing.
Backing up my hypothesis another way
I think at this point my number two cylinder has a problem with the inability to seal compression due to an audible noise and a secondary ignition pattern. But I still do not know what or where the leak is with definitive results. One of the ways I can back up my leak hypothesis a second way is to perform an in-cylinder cranking and running compression test with an in-cylinder pressure transducer. So I remove the number two spark plug and crank the engine over with fuel disabled (Figure 2).
In this waveform, you can see very low compression of the vertical column to the left. Arrow #1 indicates about 18psi of cranking compression. We see a leaning tower and non-symmetrical expansion stroke. Arrow #2 provides us with a blip on the expansion stroke. I have seen abnormalities in the expansion stroke multiple times when there is a problem in the valve train such as a broken spring or loose rocker. Arrow #3 shows a very deep expansion pocket. The deep expansion pocket tells me there is likely a leak in the cylinder.
So as we see in the waveform we have low compression because it leaked out. We also draw a deep vacuum because there is less volume in the cylinder than when we started the stroke. At arrow #4, we see on the intake stroke two brief pulses down and a consistent curve downwards on the intake stroke. The curve shape indicates a cylinder's in ability to fill. The short pulses down indicate potentially a late intake valve opening or a restriction to fill the combustion chamber on the intake side.
Proving where the leak is
I know beyond a reasonable doubt there is leak in the cylinder. I can show this to the customer visually two different ways and back up my previous hypothesis. I believe that there is something going on with the intake valve leaking and potentially not opening. Another way to prove where the leak is located is to perform a running compression test. This test does not always yield the results I need but in this case it did. During a running compression test and depending on where the leak is, I may have to manipulate the rpms, doing a snap WOT test or a decel from higher rpms.
This leak was present in the waveform at park idle no load (Figure 3).
Arrows #1 and #2 in Figure 3 show the expansion pocket and intake pocket. There is substantial difference in the level of vacuum on both sides. On most engines, these pockets should be similar in the level of vacuum if not identical. These pockets can differ from engine design and variable valve lift. However, on this engine they should match. But what does this difference in level of vacuum indicate?
One way I was taught to analyze this difference is by thinking about what is attached to each port of the engine. On the exhaust side, atmospheric pressure is present but on the intake side, manifold vacuum is present. When the piston is stroking downwards in the cylinder and a leak is present on the exhaust valve, this can siphon atmospheric pressure into the cylinder and create a less deep expansion pocket. What we see here is almost 23inhg on the expansion pocket. This value is above normal vacuum on the expansion pocket. What we can infer from this level is that the intake valve is likely leaking. If a piston is on its way down and an intake valve leaks we can add expansion pocket vacuum and intake manifold vacuum together which creates a more deep pocket. Adding this running compression waveform into my strategy helps me solidify that the intake valve is compromised.
My final hypothesis of an intake valve leak
My final test is a harder test on this engine unless you have made some special tools. Additionally, intake waveform analysis is quite tricky without the use of overlays and a lot of training. I performed this test in a way that makes it the easiest for me to understand.
When performing a cranking vacuum test you want to position the transducer in a central vacuum port. The design of the intake and exhaust for that matter can alter your data. The next best test to verify a grossly leaking intake valve is to perform a cranking vacuum test with an ignition sync. In the next test, I gain access to a central port on the intake manifold, disable fuel and sync up to an ignition event and analyze the vacuum pulls in the manifold (Figure 4).
I have identified each intake pull with arrows corresponding to each intake stroke. We find cylinder 2's intake stroke 360 degrees and to the right of the 360-degree ruler. After that point, we can input the firing order and follow each pull. The most significant analysis that can quickly be made it looking what happens near TDC compression on cylinder 2. Since the ignition event is close to TDC and we have chosen to sync on our affected cylinder, a rise in intake manifold pressure near TDC confirms that there is compression leaking into the intake manifold. Typically, when doing this analysis it is helpful to also have a relative compression test in the same capture. This allows the user to focus on which cylinder is low on compression and compare it to the intake pulls. Typically, a sync is chosen off cylinder 1 but it isn’t necessary. Picking the sync off of the cylinder that has the problem quickly identified where the leak was. Since we already know what cylinder is low on compression, a relative compression capture wasn’t necessary to identify the fault.
Summary of analysis
The first test with the secondary ignition analysis matched what I found with the last test, that there is a leak. The last test and the second to last test accurately identified which valve was leaking and backed each other up accordingly. The tear down on the intake ultimately identifies why we see what we see.
The tear down
The technician working on the vehicle removed the intake manifold and the engine was cranked over. As compression blew past my face from the stuck open intake valve I knew we found the problem (Figure 5). Only one of the intake valves is opening and the other one is seized in the head and has a broken valve spring and the rocker is off the valve and lifter. I imagine the intake valve stuck due to carbon build up, over working the valve spring till it broke. The piston probably hammered the intake valve into the head and therefore probably causing substantial damage to the valve seat. Weirdly enough after the valve cover was also removed the intake rocker is missing entirely from the engine. We managed to find most of the valve spring and one keeper still laying in the head. Does this need an engine? Maybe we can get away with a cylinder head but who knows what happened to the piston and cylinder walls. The customer declined to fix this vehicle. We were able to give him visual test results that confirm that the high dollar repair is needed. If you structure your tests to give you the best value out of your time you can find the problem faster, easier and more accurately more times than not.
