This is a real life case study about finding a BAD fuel injector on a 1994 Honda 2.2L equipped vehicle with a few simple testing techniques. Now, having said that, the info in this article applies to any Honda vehicle equipped with a 2.2L or a 2.3L fuel injected engine. Let's jump right into the story:
A few days ago, my buddy Charles called me to ask for help with his daughter's Honda Accord. He had just replaced the engine (with a used one) and the Accord was now misfiring.
He had already checked/replaced a few things to no avail and suspected he had a BAD fuel injector, but just wasn't sure.
Now, just to summarize my story a bit... when I went over I performed 4 specific tests to find the BAD fuel injector (cylinder balance test, spark test, compression test, injector noid light test, swap the fuel injector). As mentioned at the beginning, these were all simple and easy tests and I'm sure you'll want to add to your bag of diagnostic tips and techniques. OK, this is what happened:
You can find this tutorial in Spanish here: Cómo Encontrar El Inyector Averiado (Estudio De Caso) (at: autotecnico-online.com).
Cylinder Balance Test First
Over the phone, my buddy had mentioned that although the Accord was not OBD II equipped (and thus unable to provide misfire diagnostics/codes) he knew it was the #3 cylinder that was ‘dead’ and causing the rough idle.
Now, working way longer than I would have care to as an automotive tech, I know the first rule of any diagnostic is to test everything yourself. So, even though Charles said it was the #3 cylinder, I didn't take his word for it. I did a cylinder balance test to make sure.
This is what I did:
- Removed the metal locking clips from all four injectors.
- Started the engine.
- Unplugged one fuel injector at a time to see if it had any effect on the engine idle.
Now, in case you've never done a cylinder balance test before... the purpose of the cylinder balance test is to see which injector, when it's unplugged, doesn't cause the engine's idle to roughen even more. Why? Well, because if the cylinder is dead to begin with (either due to a lack of spark, fuel or compression), unplugging the fuel injector from its connector will not make a difference!
To clarify this a bit more: Since Charles' daughter's Accord was already running on only 3 cylinders... unplugging a fuel injector that belongs to a cylinder that is working will affect engine idle quality drastically!... thus letting me know that cylinder is NOT ‘dead’.
OK, getting back to my story: As I was unplugging the fuel injectors... it turned out that it wasn't the #3 cylinder that was ‘dead’... it was the #2 cylinder. Whenever I would unplug the #2 fuel injector from its connector, the engine idle was unaffected.
Spark Testing All Of The Spark Plug Wires
Now, that I knew which cylinder was the one that was misfiring (or ‘dead’), it was time to find out what was missing from the mix. This mix is the air/fuel mixture and spark that the #2 cylinder needs to work.
So, at this point in my troubleshooting, there were three specific possibilities for the #2 cylinder misfire:
- Lack of Spark.
- Lack of Fuel.
- Lack of Air (low engine compression).
I asked Charles for his spark tester and some jumper cables and started checking for spark (from each spark plug wire).
This is what I did:
- I connected the spark tester to the spark plug wire for cylinder #2.
- I grounded the spark tester with the jumper cables directly to the battery negative terminal.
- Had my bud crank the engine while I observed the spark tester.
- The spark tester sparked! Cylinder #2 was getting spark.
- I repeated this spark test on the other spark plug wires just to make sure they were OK.
It was very important for me to eliminate the spark plug wires (and indirectly the distributor cap) from the get-go because they tend to fail. I've seen it too many times where either one distributor tower or one spark plug wire stops transmitting spark, whose end result is a misfire condition. Well, the above test results, I now knew that the spark plug wires and the distributor cap were OK.
The next step was to remove and inspect the spark plugs for cracks or any other obvious problem (my bud mentioned they were new).
Since everything looked and tested OK with the ignition system... my next move was a compression test.
Compression Testing Cylinder 1 And Cylinder 2
In most cases, when I test the engine compression, I test all the cylinders. The Accord has 4 cylinders... so testing all 4 is a breeze, but I didn't.
Having solved so many misfire conditions, I've developed a feel for them and have learned from experience that it isn't always necessary to test all of the engine cylinders for compression (especially if I have identified the misfiring or ‘dead’ cylinder). In this case, I only needed to test 2 cylinders... the one with the misfire and any one of the other 3.
I'll explain why in the process of what I did:
- I already had all of the spark plugs removed from the previous test.
- So, I tested #1 cylinder for compression and got 184 PSI.
- Then tested #2 cylinder and got 180 PSI.
- I stopped here...
I stopped at #2 cylinder because #2's compression reading told me loud and clear that it's compression ability was good and that low engine compression was not the root cause of the misfire.
Now, to further explain this to you: Since cylinder #1 had 184 PSI... the lowest compression I needed to see from #2 (to know that the used engine Charles had just installed was bad) was 158 PSI. Why? because if a cylinder compression value varies more than 15% from the highest reading... that cylinder with the low compression will misfire!
This is the math I did... 184 X .15 = 27.60 (we'll rounded it off to 28). Now, 28 needs to be subtracted from 184... 184 - 28 = 156 PSI. So 156 PSI would be the lowest compression reading cylinder #2 would need to have for me to say “Aha!, this is the problem!”
OK, so far I had eliminated several things: 1) the ignition system (distributor cap, spark plug wires, and spark plugs) and 2), low engine compression in cylinder #2... what next? Let's turn the page and find out.