03 Ford 4.6 issue

Ford 4.6 in a Aviator,  same as the Mustang ending. Getting a random misfire on #6 at start up. Swapped coils. Installed new plugs. What should I try next?

Compression or leak-down test?

Swap injectors?

Problem is, it only happens at start up. Goes away pretty quickly. Doesn't always happen at startup. I do think the #6 cylinder may be a little weak, but it runs fine and when checked while running,  the cylinder isn't showing a misfire 

Are you getting a CEL?  

In reply to Docwemple :

More time for oil to run down the cylinder walls causes weaker sealing/compression masked after start?

Or a dirty injector whose poor atomization affects cold start more than running?

I don't know, but I figure the checks remain about the same?

In reply to spitfirebill :

Yes. 2 codes. Misfire #6 and Misfire at startup (1st 1000 revolutions)

In reply to Jesse Ransom :

Ok. I hope injector, not cylinder. Wierd that it's literally startup, 1st 1000 revolutions

Is there a phantom coolant loss?  Does the exhaust smell sweet when you fire it up?

In reply to Streetwiseguy :

Oo, good question. I don't suppose the #6 spark plug looked different to the others when you replaced them?

Nope, but it's obvious that the PO tried to fix it by replacing the coil and plug. Used a cheap plug, too. I installed a new set of NGK platinums. They all looked dry and old. Just tired old. No signs of anything unusual happening. 

In reply to Streetwiseguy :

Nope. Coolant is remaining at a constant level and no milk in the oil.

You could play musical injectors to see if you might have a fuel injector leaking down.  Or, if you don't feel like risking bending a fuel rail for a what-if, you could shut the engine off by pulling the fuel pump relay, or if Ford still had the inertia switch, giving the quarter panel a solid thump  (old shop prank), so the fuel rail is depressurized.  If it doesn't misfire after doing that, suspect a leaky injector.

Disconnecting the fuel pump to shut the engine off is an old FC RX-7 method of limping along with leaky injectors, which would otherwise cause the engine to flood on startup.

I can't stress this enough with Fords.  Always go to the Ford parts counter to get ignition parts for anything after OBD1.  They aren't any more expensive, and Ford electronics don't play nice with most aftermarket parts.  Even Motorcraft spark plugs from the FLAPS are sometimes different from the plugs you get at Ford.

Ford COPs are also finicky.  If one coil starts to die and you replace it, the next weakest coil will trip a code.  The answer is to replace all the coils at the same time, or take it to a dealer and pay for them to use their fancy software to diagnose the coils that should be replaced (which usually costs as much as just replacing all the coils)

A cold compression test is in order for sure, but anytime I see random misfire on a COP Ford and see one coil replaced, it's a big red flag.

I've heard mod motors don't like certain spark plugs. No idea if it's true, or internet bs.

Fords are also not the best at cylinder detection.  It might not be #6.

A misfire on startup only, without a misfire under load, is not really likely to be ignition related.  Leaky injector or small head gasket leak making the plug wet is more likely.

In reply to Streetwiseguy :

I'm leaning in that direction.  Not looking forward to swapping injectors

In reply to Streetwiseguy :

Oh yeah... never clear codes on a misfiring Ford, it will learn the misfire as normal operation!

In reply to Pete. (l33t FS) :

Took me a minute to figure that out a number of years ago.  "How can this 3.0 Taurus with one dead segment of coil, obviously running on 4, not read misfires?"

In reply to Pete. (l33t FS) :

Dood! That explains so much on my new to me 04 F150. It was shaking like a dog E36 M3ting razorblades (obvious misfire)...and no codes. 

In reply to Appleseed :

If you don't have an IDS or similar scantool that can do a cylinder contribution graph... good luck 

Code readers that can clear codes are like antibiotics.  Good in theory but they cause sooooooooo many problems when misused

In reply to Pete. (l33t FS) :

Wow. I just learned a heck of a thing, and it's unsettling not to have known that.

It underscores how playground-education my understanding of OBD2 stuff is; I'd sort of accepted the convention of trying a fix and clearing to see if codes are rethrown. Of course *most* of my cars are either so old they're not OBD2, or have been the new-ish DD that hasn't needed troubleshooting.

That seems so screwy that you could put a car into an extra-broken state by clearing codes... So at that point you effectively have to fix it "blind" (without meaningful diagnostics) and then clear the codes so it can start from "this is what it looks like when not broken?" Is Ford more or less alone on this one?

Jesse Ransom said:

In reply to Pete. (l33t FS) :

Wow. I just learned a heck of a thing, and it's unsettling not to have known that.

It underscores how playground-education my understanding of OBD2 stuff is; I'd sort of accepted the convention of trying a fix and clearing to see if codes are rethrown. Of course *most* of my cars are either so old they're not OBD2, or have been the new-ish DD that hasn't needed troubleshooting.

That seems so screwy that you could put a car into an extra-broken state by clearing codes... So at that point you effectively have to fix it "blind" (without meaningful diagnostics) and then clear the codes so it can start from "this is what it looks like when not broken?" Is Ford more or less alone on this one?

It's even worse than that.  If i'm reading Pete's post correctly if you clear the codes before you fix the issue, then the computer thinks that the car is supposed to misfire and won't throw new codes.  You effectively have to depower the computer so it goes back to system defaults, have the computer throw misfire codes (if they still exist) then fix the misfire and allow the computer to clear the codes on it's own (i.e. log operation without the misfire condition).

In reply to 06HHR (Forum Supporter) :

I phrased it badly, but yeah, that's the crux. No diagnostic info from a system that thinks that whatever the state is after clearing codes is "normal." But that's actually better than I was thinking if you can get the computer to unlearn the misfire-as-normal and restore defaults by depowering it. I jumped to the conclusion that the learned status wasn't immediately clearable (which was perhaps silly on my part). So thanks for pointing out that it's terrible, but in a way slightly less terrible than I thought...

06HHR (Forum Supporter) said:

Jesse Ransom said:

In reply to Pete. (l33t FS) :

Wow. I just learned a heck of a thing, and it's unsettling not to have known that.

It underscores how playground-education my understanding of OBD2 stuff is; I'd sort of accepted the convention of trying a fix and clearing to see if codes are rethrown. Of course *most* of my cars are either so old they're not OBD2, or have been the new-ish DD that hasn't needed troubleshooting.

That seems so screwy that you could put a car into an extra-broken state by clearing codes... So at that point you effectively have to fix it "blind" (without meaningful diagnostics) and then clear the codes so it can start from "this is what it looks like when not broken?" Is Ford more or less alone on this one?

It's even worse than that.  If i'm reading Pete's post correctly if you clear the codes before you fix the issue, then the computer thinks that the car is supposed to misfire and won't throw new codes.  You effectively have to depower the computer so it goes back to system defaults, have the computer throw misfire codes (if they still exist) then fix the misfire and allow the computer to clear the codes on it's own (i.e. log operation without the misfire condition).

Sort of but no.  Fords of the era would...

Ok backing up.  When you "clear codes" you are not JUST clearing codes.  Primarily, you are erasing all emissions monitor test results, bad and good.  This is where you will fail emissions if you punch the "clear" button and immediately go to the test center.  All of the emissions monitors will be "Not Ready" (have not run yet) because you just told the computer to start all over again.

Some manufacturers reset all adaptives.  Fuel trims, ignition trims, etc, idle speed/throttle angle (or IAC counts), also known as "I took my car to Autozone and now it won't start", etc.  Some manufacturers do not.  In Ford's case, they would also reset the crankshaft speed variance limits.  So a computer that has had its memory wiped, be it from the magic button or from a battery disconnect, will see the engine slowing down a lot after TDC 5 and 2 or whatever, and just go okay, the crank wheel is shaped thisaway, this is just normal behavior...

There really is no "baseline" as far as I could tell... if the adaptive is cleared, then the next time the engine starts, it learns the crank profile.    Other manufacturers have very specific procedures for relearning the crank profile, others have none at all and assume that mass production is perfect and errorproof 

I do not think Ford does it this way anymore, but certainly in the late 90s and early 00s, you really did not want to clear codes.

Really, "clearing codes" is a bad idea no matter what, but, you know, sometimes ya gotta.

Well crud.. I'll be careful. It has been clearing on its own at times. It even did the flsding light, then went away