Bitten by a misfire

First things first

The vehicle that was about to eat my lunch was a 2003 Ford Mustang equipped with a 3.8 liter powerplant. The Check Engine light (also referred to as the MIL light) was illuminated and when I started the car to bring it into the shop, it was quite evident from the tone of the engine’s idle cadence that one hole wasn’t firing.

I connected my scan tool in global OBD-II mode nd looked to see what DTCs the ECM had lying in wait for me. One was, as I suspected, a P0301 defined as “Cylinder #1 Misfire Detected.” There was also a second DTC recorded by the computer. This one was a P0316 which translates to “Misfire Detected On Start-Up/First 1000 Revolutions.” This one was important to this diagnosis because it meant that the vehicle’s computer has detected a misfire in one or more cylinders right after the engine started spinning.

The next stop was Mode $02 (the “$” sign means it’s a hexadecimal number used in programming) to take a look at the freeze frame data recorded by the ECM at the time it decided to turn the light on. The OBD-II misfire monitor is a continuous monitor, constantly testing and updating the results during any given drive cycle. For the ECM to turn on the MIL, it must see the misfire occur at least twice in a row, and under the same load and engine speed. The freeze frame information can help you determine what the load and engine speed conditions were at the time the fault was noted providing some clues as to the cause of the misfire.

In this case, all it did was confirm what I already knew. The engine misfired cold, at idle. But at least I had done everything right so far.

What is a misfire?

Often a repair is made based solely on the initial code description shown on most scan tool screens. This typically results in a repair that doesn’t fix the problem and a waste of time and money. This is especially true when it comes to misfires. Misfire codes are one of the top 10 codes in the United States. Do you want to guess what parts are sold most often to attempt a repair?

Yep! Spark plugs and ignition coils.

Throwing parts at a problem in the hope that it will fix the car is not a professional approach to a customer concern. When diagnosing any kind of DTC, you have to understand how the computer is testing the system. The computer, not the owner, becomes your customer. If your repair does not meet the computer’s test requirements, the code will return, the Check Engine light will come back and your customer is going to be less than happy.

Most OEMs use a strategy that uses crankshaft speed to detect a misfire. The ECM monitors the CKP signal. When one of the cylinders produces less than its fair share of power, the crank will momentarily slow. Any condition that results in a cylinder’s loss of power could be the cause of a misfire code.

With the atypical cadence the engine was displaying while cranking, the most likely cause of this misfire was a loss of compression. Checking engine mechanical condition should be one of the first steps taken when dealing with any driveability problem. I had my old UEI scope already on the work bench, so I connected a high amp clamp to it and checked the relative compression using the starter current draw as my indicator. The pattern on the screen was painfully obvious, with five nice little peaks followed by one missing one.

Hmmm, I wondered what cylinder it could be?

Diving into number one

I removed the plug from the suspect cylinder and threaded in a traditional compression gauge. And before anyone writes me to say I should have used a pressure transducer with my scope, this story took place long before the in-cylinder pressure testing technique was invented!

When you perform this test, you need to disable the engine so it won’t start. On some vehicles, all you have to do is hold the accelerator pedal to the floor so the ECM enters “Clear Flood” mode to prevent the start.

I cranked it over and the result was a measly 54 psi. Nowhere near the 120-160 psi specification.

Next, I grabbed my cylinder leakdown tester to isolate where the loss was occurring. A cylinder leakdown tester is just a tool that allows you to pressurize the cylinder in question with shop air and has two gauges fitted. One is the air pressure entering the cylinder, and the other is the pressure being maintained in the cylinder. If the cylinder is tight, the two pressures should agree. If there’s a leak past the rings or the valves, the second pressure will be lower than the first. That difference can be considered as a percentage loss, with approximately 10% being the maximum acceptable loss. The primary advantage of this tool is that you’ll hear the air escaping. Air streaming from the tail pipe, for example, means the exhaust valve is leaking.

In the case of the Mustang, I had 90 psi going in and roughly 70 psi staying in, for a 20 psi (or 18%) difference. The air was clearly heard coming from the throttle body, so I knew I had a leak coming from the intake valves. After a conversation with both my service advisor and the customer, I removed the head and sent it out for reconditioning by our local machine shop.

Fixed! Or so I thought.

A few days later, the head was back and it was time to put the Ford back together. Time to see if it’s fixed!

Damn! The engine started well enough, but the misfire was still there and all signs still pointed to that loss of compression!

A quick check showed the same lousy 54 psi I had before. I hooked up the leakdown tester to check the valves. Maybe the machine shop had messed up, though they never had in the past. With the air flowing in, the cylinder was holding tight as a drum. The intake valve may have been leaking when I tested it and needed repair but it wasn’t the cause of the massive loss of compression.

But how can that be? If the rings and valves are sealing perfectly how can the compression be low?

Can the engine breathe?

If the engine was turning over and there was no loss of seal, then it must have been related to the amount, or lack thereof, of air getting into the cylinder in the first place. What if the valves hadn’t been opening fully, or not opening when they should have? Could I have a worn cam lobe on that cylinder? A valve train issue that prevented the valve from moving to its full lift? All I knew then was that I was going to be doing this job all over again.

As I began to remove the head, I paid close attention to everything I could think of as I worked my way down. I watched the valves open and close while comparing them to the companion cylinders on the left bank of this V6 design. I measured the installed height to see if there were any variations. I checked the push rod length to make sure there wasn’t a problem there. I didn’t find anything.

Do you see what I should have seen the first time I had this head off? Photo: Pete Meier.

Until I got the head off, that is. Looking closely at the front two cylinders, I could see what I should have seen the first time around. The ridge wear on the failed cylinder extended at least a quarter of an inch into the cylinder while its companion cylinder’s ridge was barely visible. Rotating the engine over with the head off made the problem even more apparent. The #1 piston was not reaching TDC! This wasn’t a matter of compression pressure. It was a matter of compression ratio.

I removed the piston and rod from the engine to find the rod bent almost perfectly along its axis. This allowed it to continue to move without any noise or vibration, and effectively shortened the stroke on that cylinder. Since the piston had a “new” TDC, the compression ratio for that cylinder was almost non-existent. And I should have known that a small 18% leakage from the valve wouldn’t have that dramatic an impact on the measured compression reading.

I’m not sure how this rod got bent, but hydrolock is a likely culprit. Photo: Pete Meier.

I’m not sure what caused the bent rod, but I suspect a prior coolant leak into that cylinder resulted in a hydraulic lock. While the head gasket/head was repaired, the bent rod was not. Installing a new rod corrected the issue, and I returned the vehicle to the customer a little wiser for the experience. What was the lesson I learned that you can apply to your own troubleshooting? Follow the data you collect, understand what the ECM is looking at to make the call to turn on the MIL, and never assume anything!