Engine noise, only at very certain circumstance

[BBuzecky]

First off, 2004 Mazda 3, 2.0L Automatic 4-speed, 73,000 miles

Noise has developed over the last few thousand miles. It's always hard to explain a sound, but I would say it almost sounds like noisy injectors or almost like a diesel engine, just a constant small rattle. A few interesting notes...

-- Only happens in 4th gear, with slight pressure on the gas pedal. I can adjust how loud the noise gets to a point by gas pedal pressure. It will be very faint if I just barely have pressure on the pedal, if I put a little more pressure the sound gets louder. However, if I put a moderate amount of pressure on the pedal, the car will rev from about 2000rpm to 2700rpm (normal). The instant the rpms move up due to moderate pressure, the sound dissappears.

-- The sound happens with an ice cold engine, and also with a fully warmed engine. Thus, can happen at any time.

-- Sound never happens at idle, startup, or 1st through 3rd gear. Only time is in final gear, with very little gas pedal pressure.

Since I can somewhat control the sound through light pressure on the pedal, I would guess it has something to do with fuel injection, or maybe a valve adjustment?

Finally, car uses Shell Gasoline, and is normally treated with Redline Fuel Treatment, so adding fuel injector is not the solution certainly. At any rate, what's your opinion?

 

[The_Jeff]

It's a ping, you notice it because of your torque converter lockup. Also likely the PCM advances the spark a little in top gear for MPG, not to mention high gear is the hardest for it to pull, making pinging easiest.

Try high test gas, or ignore it, or try all sorts of things to fix what's harmless under light load. I'm personally a fan of cooler spark plugs and a coolant flush.

 

[ARCOgraphite]

Not sure about the ping on dead cold engine - but knock/ping was my first though too - and still the most likely issue. BTW I was noticing noisy injectors when i started using Shell regularly - it went away with another brand of gas - then came back and wont go away now! Maybe its an ethanol issue.

 

[ARCOgraphite]

If it is ping, you would only need maybe 2 gal at the most shell v-power in a tankfull to quell it. I would NOT run a complete tank full of it OR any 89 octane. Your ECU would retune for max timing and a diff trim, then when you went back to 87 w/ E10, all heck would break loose (afa ping and overadvanced timing). Hey! Been there done that

 

[Carbon]

Originally Posted By: ARCOgraphite
If it is ping, you would only need maybe 2 gal at the most shell v-power in a tankfull to quell it. I would NOT run a complete tank full of it OR any 89 octane. Your ECU would retune for max timing and a diff trim, then when you went back to 87 w/ E10, all heck would break loose (afa ping and overadvanced timing). Hey! Been there done that

I am sceptical of what you said. Are implying that the engine takes hours rather than seconds to adjust to a change in octane? Or are you talking about the transmission taking longer to adjust?

 

[BBuzecky]

Thanks for the responses. I am going to fill up with 6-7 gallons of Shell Premium and see if it affects the issue. I'll update the thread.

 

[holla]

Wow, that sounds so much like my 96 civic. I've been fighting something like that for about 5 years, only mine is linked to cold weather. Anyhow, I've tried this:

exhaust manifold (it was cracked but that wasn't the noise)
valve clearance
tune-up (plugs wires cap rotor (twice actually))
all kinds of filters/fluids
ignition timing (you can set the "baseline" but of course the ecu takes over to a degree and does what it wants)
injectors

I once used 89 octane for a full tank and saw no difference, so I ruled out ping for about 5 yrs, esp since it's linked to cold. Just last week I finally bought a full tank of premium with a bone-dry tank and haven't heard it in several days. It seems (I've been wrong before, see above) it might be ping, and because of spark advance I believe ping CAN be linked to a cold engine. Take a high compression ratio 4cl w/ 200k miles, maybe with a little carbon increasing CR further, over-advance spark when cold , maybe have a bad knock sensor or ckf or HO2S without throwing a code, and boom.

So, I'd say give a full tank of premium a try. Its really the only way to rule ping out. Don't waste 5yrs like me because you bought mid-grade and it still detonated. And I've NEVER heard not to switch grades of gas, your ECU will adjust in minutes and you won't even know it.

 

[AdRock]

There could be carbon buildup on the pistons causing the compression to raise just enough that you need to go to a higher octane fuel.

 

[PT1]

I would:

1. fill up with a brand other than what you are using and get at least 89 octane
2. give it an Italian tune up on the freeway

re-assess.

 

[ARCOgraphite]

Originally Posted By: Carbon
... Are implying that the engine takes hours rather than seconds to adjust to a change in octane? Or are you talking about the transmission taking longer to adjust?
The octain adjustment for catastrophic operation would be instant (if you got a load of Kero) through the K-sensor. But the look up tables can take time to build and time to remap. All depends on your software and power of the ECU on your particular engine - I had noticed "premium" vehicles to have more elastic and adaptive ECU than lets say Hyundai in the past. I Feel it ( as does the car) in New England in Winter with many a warm drive home from work and a sub zero morning start.

 

[ARCOgraphite]

Originally Posted By: BBuzecky
Thanks for the responses. I am going to fill up with 6-7 gallons of Shell Premium and see if it affects the issue. I'll update the thread.
Hope it works out, We've had some apparantly "below rated octane" last month around here - everything was pinging. 2 gals of V-power, if "fresh" will do it. If not, you need a decarboning.

 

[Gary Allan]

Quote:
It's a ping, you notice it because of your torque converter lockup.



That was my first thought, given the conditions cited. However there would be no reason for it to just be noticed in the last few thousand miles. That is, unless tune is due for a refresh.

...a simple test would be to tap the brake while it's occurring. This usually releases the converter lock. If he could cycle it in and out at will doing this, that would kinda nail it.

 

[BBuzecky]

Gary, I have attempted tapping the brakes and it always stops the sound. When I get back on the accelerator slightly, the sound comes back.

The sound does not happen every time I drive, but has happened more and more often in my travels. As said, I never heard this sound in the first 70,000 miles on the car.

Lastly, I am going to clean the throttle body out this spring. Mazda has released a TSB stating that they can get clogged up with carbon deposits, and affect the idle. Sure enough, my idle has gone down-hill in the past year. It will idle properly at around 750-800rpm, then rumble a bit and goto 500rpm, almost dieing, then going back up to normal idle rpm. On the Mazda 3 boards, virtually every user with 60,000-80,000 miles on an '04 or '05 have complained about the situation. Maybe this built up carbon in the throttle body is affecting the pistons?


I am going to fill up with premium this afternoon, and open it up a bit on the highway. It's reassuring that most on here think it's just ping, and likely not a serious issue with the engine. I will let you guys know how it develops, gratefully appreciate the responses.

 

[mechtech2]

Computer programs don't add ignition advance, they subtract it from the knock sensor's input.
If he needs 89 or 91 octane, if it is recommended, he should use it.
A computer will normally not relearn more advance.

 

[ARCOgraphite]

IIRC most start from full timing map at key on cold start and retard selectively at knock onset. KS Usually dont care about ping. Will not most 32bit ecu obdii keep retard in adaptive remap table for warm restart or for more than 1 cold restart cycles?

 

[ARCOgraphite]

Originally Posted By: BBuzecky
...


I am going to fill up with premium this afternoon, and open it up a bit on the highway. ...
You really only need 2 gallons on 10 or 3 on 15. More is not better - the delayed burn can cause MORE carboning.

 

[Carbon]

BBuzecky, have you checked the sticker inside of your gas filler cover on your Mazda? I expect it says something about using regular/ 87 AKI, but some cars say "premium recommended".

 

[Gary Allan]

Quote:
Gary, I have attempted tapping the brakes and it always stops the sound. When I get back on the accelerator slightly, the sound comes back.


Do it while maintaining your foot on the gas pedal at a steady state of the condition. If the brake tapping (just the light circuit needs to be activated) causes the noise to go away, then it would point to the converter clutch. This should be occurring somewhere at lower speeds where you're at max shear load in high gear. Progressive throttle (slowly) should maintain the "groaning" until the clutch releases ..where you'll experience a jump in rpms.

 

[ARCOgraphite]

Just another thought, the converter should not be locking on a cold engine, in fact it should not go into OD until warm. - unless MAzda changed this programming recently for increased fuel mileage.

 

[BBuzecky]

UPDATE:

So I have ran through a good two tanks of premium Shell fuel. The problem is still evident. I am going to be cleaning out my throttle body this summer, as carbon build-up is a problem on my car. I am thinking that as a result of the throttle body, I have carbon build-up in my engine which may be causing the noise? At any rate, it's not octane related, does anyone have insight?