I had some problems which I suspected were related ( diagnostic index at pg. 150 in 98 US WSM ):

  • idle speed too slow and excessive engine shake pg. 167
  • Hesitation / Stumble: Momentary pause at beginning of acceleration, or during acceleration pg. 170
  • Performance poor under load (i.e., power down when climbing hills) pg. 174

In the end, the cause turned out to be the spark plugs, which seem to have gone bad due to large amounts of oil collecting in the spark plug holes due to a completely shot valve cover gasket. One thing I'd like to point out is that defective spark plugs are not listed as a potential cause for any of these problems in the WSM. The only ignition related issue listed for any of these is Spark leak from hightension leads.

Even though I solved the problem without changing out any other parts ( other than a defective O2 sensor ) or doing any unnecessary maintenance ( think cleaning IAC, EGR, etc ) I still feel like it took me a long time to track down through a very long process of elimination, and a lucky guess.

I'm wondering if maybe there is some way I could have diagnosed the problem more quickly and in a more direct and certain manner.

What I mean by that is something along the lines of, "Oh, conditions X, Y and Z exist, so it must be the spark plugs."

So I guess I'll describe the steps I went through and maybe all y'all can make some suggestions how I could have improved my diagnostic process.

This is a 98 Mazda 626 GF 2L ATX. While this international version supports OBDII via J1850 PWM, it has no OBDII connector ( only the proprietary Mazda DLC ) and I am not sure how much of OBDII it actually implements as far as the types of errors it can detect and report.

Miscellaneous things, new air filter, PCV grommet, air intake hose and valve cover gasket, which I did just because they were needed but without relation to trying to solve this problem. I also did a homebrew seafoam of the intake and the piston heads. I also had to remove the throttle body to replace a burst coolant hose which runs under the intake manifold, so I cleaned the TB while I had it off and replaced the TB gasket. There was also a bad thermostat which I replaced. Also had to replace the battery which was dying. Did I mention I got this car real cheap :-)

Initially, I didn't have a scan tool available, and there was no check engine light, so I just looked up the problems in the WSM and started testing the things which seemed easiest to test, and which were common to all three listed problems, like voltages of the EGR solenoid, the camshaft position sensor and the crankshaft position sensor ( plus the gap between the sensor and the crank teeth). I also did a power balance test which was inconclusive, and checked that all the timing marks lined up correctly. I also checked cam to lifter clearances which were all in speck, although a few were clearly wearing at a much quicker rate than the others.

I did a cigar test for air leaks, and found a very small one from an injector grommet, which I fixed, but didn't have any effect on the problem.

I eliminated transmission problems since it was shifting into gear without any delay, and I didn't see anything unusual in the ATX fluid.

I unplugged the IAC while idling, the car stalled, which is what it's supposed to do according to the WSM.

I did a vacuum test to see if anything stood out, but the vacuum was steady at 17 @ 650 rpm and 18 at 2k rpm, although the WSM says minimum is 18 at sea level. I'm at about 950 meters above sea level, so that's probably a somewhat OK reading.

After a few weeks a scan tool arrived from the states. There were no diagnostic trouble codes, KOEO and KOER self diagnostics ran without error, and long term fuel trim was close to zero, as was short term fuel trim except during moderate to hard acceleration, when it got as high as 14.5%. I tried graphing various parameters during test drives, like TP_v, MAF, IAC, Engine Load, Fuel trims and Spark Advance and couldn't find anything that really stood out, other than that engine load seemed excessively high to me ( although I'm not sure if it actually was ), STFT jumped up to 14.5% on acceleration and the spark advance was all over the place, with a big dip then jump when accelerating.

At this point I asked a question regarding what seemed like wacky spark advance, and someone suggested looking that the O2 sensor which turned out to be defective.

I went for some more drives, did some more graphing and saw that at idle STFT was now oscillating between around 2.5 to -2.5 and LTFT was sticking around 1 to 1.5%. Revving the engine to about 3k RPM would cause both STFT and LTFT to jump to about 5.5% with LTFT dipping to -5% after the rev before quickly returning to about 1%. Nothing else really stood out to me, although maybe I just don't know what to look for.

I then did a spark test using a known good plug to check the wires and coil, which seemed OK, and a compression test which was also OK. I don't quite understand the vacuum numbers being at minimum acceptable levels in light of the fact that my compression numbers are no where near the minimum ( actually the compression was almost like a new engine ).

On a hunch I replaced the plugs, based on reading about someone who had also had a bad valve cover gasket, with oil leaking into the plug holes and similar symptoms to mine who fixed the problem by replacing the coil, wires and plugs. As I said, that completely solved all the above problems.

Is there anything I could have done to narrow this problem down to the spark plugs more quickly?

  • 1
    I'm glad you finally got it figured out. I don't have time right now to post an answer ... I'm betting one of our better people will be doing so soon. This is a great question, by the way. Commented Dec 27, 2015 at 12:13
  • Yes Indeed. What @Paulster2 said. Great question! Commented Dec 27, 2015 at 18:17

1 Answer 1


The WSM is useful to a point but note a couple of key facts about them; they are written for the dealer level technician and they are not meant to be complete. So when it says spark leak a trained tech knows to look at the whole ignition system. The diagnostic material is only written to the cover the most common problems. The OEM’s train there techs to call the technical help departments if the answer is not found in the published materials. Unfortunately, I know of no substitute for experience, following formulas and checklists can get one only so far down the road.

Here is my standard approach to this symptom. A quick test drive to verify the symptom. A visual inspection of the secondary (high voltage) ignition system; plugs wires, cap rotor, ignition coil. This includes looking for oil leaking on to the ignition wires and spark plugs. Then replace any parts that have obviously failed; followed by another test drive. 90% of the time this solves the misfire. If it is still present an ignition scope is used to identify which cylinder is at fault and if the problem is in the ignition system or elsewhere. If that looks OK then we check compression and fuel mixture.

The reason weak and leaking spark causes misfire on acceleration is that the mixture goes lean. Fuel is a better conductor of electricity than compressed air so it takes a higher current to initiate the spark across the plug gap. Electricity will always take the path of least resistance. If the insulation is poor it will jump to ground at that spot. One can usually see the arc site because it has carbon tracking.

  • So oil leaking onto the wire or spark plugs is a potential cause of failure for those parts? Looking at the plugs I took out, do you see anything that would indicate "obvious" failure? Did you mean fuel is a worse conductor so it takes a higher current? Commented Dec 29, 2015 at 13:53
  • Forgot the link to the pic: mechanics.stackexchange.com/a/23671/7132 Commented Dec 29, 2015 at 14:08
  • @RobertS.Barnes Oil softens and breaks down the rubber in the wires which compromises its insulation capability. You can't see this, experience teaches us that wires that have been soaked in hydrocarbons should be replaced. Fuel is a better conductor than air, so a lean mixture takes a higher current than a rich mixture. Commented Dec 29, 2015 at 15:41
  • @RobertS.Barnes On the spark plugs in the post you linked. The black steaks on the plug on the left are a classic sign of arcing due to spark leak. Commented Dec 29, 2015 at 15:51
  • Would an inline spark tester have told me anything useful in a case like this? Commented Dec 30, 2015 at 9:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .