On a 98 Mazda 626 GF 2L I'm working on the engine seems to shake / wobble for about a second / second and a half when going to WOT from any starting rpm. It also almost sounded like maybe one of the belts was even momentarily slipping. If I accelerate gently everything is smooth and there is no shake or wobble. It's an automatic and I did this all in park. Not sure if this is related or not, but the idle seems to be low and a bit shaky / erratic. Adjusting the idle air screw doesn't seem to have much effect one way or the other.

I'm assuming this to be abnormal, as I don't experience any shake going to WOT on my other car, a 99 Almera which also has a low occasionally erratic idle and even stalls a few times a day ( haven't had time to work on it yet ).


So I just did a cigar smoke test, and I noticed that smoke looked like it was coming out the number three injector. On closer inspection, it looks like the smoke was coming out where the grommet holds the injector in to the manifold. I also noticed what looks like a dark stain around one side of the grommet ( I shined a light on it and it's definitely not a shadow ):

enter image description here

If smoke can get out at the injector grommet, I'm assuming air can get in.

I also took out the number three plug, and upon close inspection one side of the plug seemed to have a yellowish hue to it.

Could this be the source of my low rough idle and the shaking / wobbling when pressing down hard on the gas? Maybe the excess air is causing that cylinder to run lean and hot, and causing the yellowish tint on the plug?


So I got a scan tool. Yippie! I fixed the small air leak from the #3 lower injector seal, and couldn't see any other air leaks ( doesn't necessarily mean there aren't ), but still have the same problem. I reset the ECU and logged a 16 minutes drive. During idle and cruise STFT stayed at zero, but during moderate or heavy acceleration ( which is mushy at first, then solidifies after a second or two ) I see STFT go around 9% - 14% then return to zero at cruise or light acceleration. Engine load seems to be over 60% during these injection spurts. The engine seems smooth at higher rpm while cruising. LTFT stayed at zero the whole time.

At the end of the drive I ran the car's internal "Key On Engine Running" self test which didn't show any errors

The 98 US WSM lists the following possible causes for "Idle speed too low and excessive engine shake":

• Air leak from intake air system parts
• Spark leak from hightension leads
• Improper operation of IAC valve
• Erratic or no signal from camshaft position (CMP) sensor
• Erratic signal from crankshaft position (CKP) sensor
• PCV valve malfunction
• Restriction in exhaust system
• Inadequate fuel pressure
• Fuel leak from fuel injector
• Engine overheating
• A/C system improperly operation
• Purge solenoid valve malfunction
• EGR valve malfunction
• Low engine compression
• Poor fuel quality
• Air cleaner restriction
• Electrical connector disconnection
• Fuel pump mechanical malfunction
• Fuel injector clogged
• Vacuum leak
  • Are you still having this issue? Commented Dec 8, 2015 at 15:06
  • @DucatiKiller After fixing that fairly minor air leak, I smoke tested it again and couldn't find any other leaks, but it was still giving the same behavior. Then I got a scanner and saw that the O2 sensor was staying pegged to zero volts and the ECU showed the car was never getting out of open loop mode. Also noticed that my ignition timing was all foobarred also. You can read about it and see the graphs here: mechanics.stackexchange.com/questions/22686/… Commented Dec 9, 2015 at 18:30

3 Answers 3


So I finally figured out the problem.

First off I had a defective O2 / Lambda sensor, which I replaced. However, this had little or no noticeable effect on the problem. The valve cover gasket was completely shot, with oil constantly coming out and also leaking into the spark plug holes. I fixed the gasket, which also didn't solve the problem. However, I read a post ( which I can't find at the moment ) in which someone was suffering the same symptoms and also had oil leaking into the plug holes. He changed the gasket and the plugs, wires and coil which solved his problem. So I took a known good plug and used it to do a spark test. I saw I had good spark at each wire, but decided to change the plugs since the insulators looked a bit burnt, hoping that would help. Well, it did!

enter image description here

I had actually asked about the plugs previously, but everyone said they looked fine, so I went on a wild goose chase checking a million different things.

I'll try and follow with a more detailed answer later.

EDIT March 20th, 2016

This has eliminated about 80% of the problem, but after driving around for a while I realized that the problem is still there, just at a greatly reduced intensity, including both the hesitation and the rough idle. The replacement plugs I used are NGK G-Power plugs with a fine tip, which according what I've read in various places: "a much lower voltage is required to jump the spark gap".

If this is true, it makes me suspect that the great reduction in hesitation and wobble when revving the engine may be due to the lower voltage requirement, indicating that the coil pack is getting weak. I've attempted to test this theory by using a adjustable gap spark tester, however my results have been inconclusive since I don't know what voltage I should be testing, and what the resultant spark should look like.

EDIT Nov. 25th 2016

Thanks to a comment by @FredWilson on a different question about recurring arcing it became clear that when there is arcing on a plug, there is usually also a carbon trail on the inside of the plug boot:

enter image description here

This will contaminate the new plug fairly quickly and cause the misfire to return:

enter image description here

EDIT Nov. 28th 2016

So I pulled out all the plugs, and cleaned the porcelain sections and the plug boots with IPA and 320 grit sand paper to remove the carbon trails, then I slathered some 3M Silicone Paste ( dielectric grease ) on the porcelain part of the plugs and dabbed a little into the boots.

So far so good, the hesitation is gone and the increase in power is noticeable. It seems to have helped somewhat with the idle as well, although that's really a separate and intermittent problem, which I suspect is due to either clogged / leaky injectors or a wonky fuel pressure regulator.

  • The plug on the left has the black streaks that a classic sign of arcing. The spark is leaking to ground along the outside of the plug insulator. Commented Dec 29, 2015 at 15:47
  • @FredWilson So what we're seeing on plug 1 is like what's described here? 3 : Flash-Over - Spark leakage from terminal to metal shell ngk.com.au/spark-plugs/technical-information/… Commented Dec 30, 2015 at 14:14
  • Yes, your plug shows the flash over noted on that site is what that plug exhibits. It is often not as clear as that picture shows. Commented Dec 30, 2015 at 14:48

This may well be the smoking gun

That "grommet" is actually a fuel injector seal. It is designed to keep air out and prevent fuel from escaping the intake manifold.

If it is compromised then fuel can leak out, which will explain the presence of the dark stain.

A quick way to confirm this would be to turn the key in the ignition to the 'On' position to allow the fuel pump to pressurize the fuel rail. The pressure in the fuel rail should cause fuel to leak out if the seal is damaged or torn.

A damaged seal can also explain the rough idle and shaking, not only because unmetered air can get in, but metered fuel can wind up not reaching the cylinders as well, resulting in a lean condition that explains the physical appearance of the Cyl #3 spark plug.

  • It's not the O-ring which seals the injector to the rail, it's the grommet which hold the tip of the injector in the manifold / head, kind of like the PCV grommet. Commented Oct 3, 2015 at 16:52
  • @RobertS.Barnes : That's right. There are two O-rings on any fuel injector, one between the fuel rail and injector, the other between the intake manifold and injector. Both can leak fuel if compromised. It may let in unmetered air as well if the pressure differential between the manifold and atmosphere is negative.
    – Zaid
    Commented Oct 4, 2015 at 13:07

If you have an air leak that can definitely cause roughness in idle and during acceleration. So definitely fix that first and then see if the problem persists.

You must log in to answer this question.

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