Cliffjumper126

Alright guys, I need some input, as I am completely dumbfounded. I can't find my issue anywhere on the site, and I'm stumped, so I want you guys' opinions.

Lately, when I start my car cold, it runs perfectly fine until the RPMs drop to ~1.2, then it starts shaking and the revs go up and down. Pushing the clutch makes no difference. If I rev it, it revs fine, but when the RPMs go back down, it starts shaking again. If I try to drive it while it's doing this, it bucks and jerks and the throttle acts funny. If I'm driving during this time and I floor it, the RPMs barely go up, and after about 10 seconds of holding the throttle wide open, it jerks and accelerates like it should. Then it starts doing this all over again. Eventually, it always stops acting like this and runs normally. There are no codes apart from my usual ones from my gutted cat & disconnected air pump, so the ECU isn't noticing anything (surprise, surprise)

The coils & wires have about 10k on them, and the plugs have ~3k. I don't think it's ignition related because it runs great once it stops this crap. I'll get a video during a drive in a bit. I did notice that the STFT went from 0 (before it starts acting up) to -24.9%
The LTFT was 7.0% and the load value was 50something...wtf

-Not a coolant seal, coolant level is perfect, and it doesn't steam or smell like coolant
-Not the control rings, oil level hasn't really changed, and there's no smoke.
-Wires are all seated correctly.
-Cleaned MAF, no difference.

Strange RX-8 Behavior When Cold - YouTube

RIWWP

iTrader: (2)

Hmm.

With a STFT maxed out negatively, it's seeing you WAY rich and it's trying to pull every bit of fuel it can. A few possibilities:
- O2 sensor is shot, and it's reporting rich when it shouldn't be. If it is, then this could easily actually be a problem with the O2 heater circuit, having trouble getting it heated up as fast as the ECU thinks it should be, causing a disconnect between what it's reporting and what is actually flowing past the sensor.
- MAF heater circuit could be failing, so it's seeing a high G/S MAF value when it's just that the MAF isn't hot enough yet, so the ECU is adding tons of fuel, which it then sees as way too much.

Just theories...

9krpmrx8

iTrader: (46)

Sounds like a vacuum leak that gets sealed up when the car warms up.

But I did have an engine that was dying from a coolant seal failure and I had no smoke, noticeable coolant loss, etc. It was only leaking coolant into the housing for a few seconds at start up and then would seal up once things started to warm up (quickly in Texas). It would idle rough and then smooth out. I only confirmed this when my oil sample showed traces of coolant in the oil and then I pulled the engine and sent it to Pineapple racing where they found a slight failure of the coolant seal upon tear down.

9krpmrx8

iTrader: (46)

At idle those items are not in play.

RIWWP

iTrader: (2)

It was running in closed loop, so that would mean it's using the O2 right?

Wouldn't a vacuum leak show positive STFT, not negative?

9krpmrx8

iTrader: (46)

At idle it just runs on an "idle map", it will even idle with the MAF unplugged. As for STFT, well generally I just look for jacked up STFT, I have never paid attention to whether or not they were + or - honestly.

9krpmrx8

iTrader: (46)

Cliff, just go over everything with a fine tooth comb. I have seen some silly things cause pretty serious running issues. Try following the FSM to narrow things down a bit.

Engine, troubleshooting, rough idle

M A Z D A

RIWWP

iTrader: (2)

I would think that the ECU is using the MAF and O2 at closed loop idle, since there is a STFT adjustment. If neither was being used, there would never be an adjustment, because it wouldn't even be looking to see if something wasn't correct. Having a STFT adjustment means that it's seeing an amount of air flowing past the MAF, and then it's seeing through the O2 that it's not ideal.

A vacuum leak would mean less is flowing past the MAF than the engine is using, so it's injecting too little fuel, the O2 sees lean and adds positive STFT to compensate.

9krpmrx8

iTrader: (46)

Ya think?

j9fd3s

well looking at the OBD data the only thing that really jumps out at me is the calculated load value, 50% seems high. although the airflow number was in a unit that means nothing to me.

so AFM, o2 sensor, and of course vacuum leaks/ weird broken stuff

its possible the coolant temp sensor has a dead spot, but it seemed like it was giving rational values.

9krpmrx8

iTrader: (46)

I missed his video, yeah 50% CL is strange.

wcs

iTrader: (2)

Wow that is really weird.

You cleaned and reset the ESS before, right?

Anything else maybe ... just brain storming here

Different place you got gas?
Heavy rain or other hazards encountered recently?
New parts?

Like 9k and RIWWP said ... starting going thru everything again.

Cliffjumper126

Ok, starting from the top:

RIWWP, the 02 sensor is brand new. Literally less then 2k on it. The MAF, however, I could see having problems. Wouldn't a failing MAF cause hot idle problems? Because my hot idle is fine

9k, I feel like it isn't a coolant seal because it runs great when it's first started. The **** only hits the fan after it's been running a few minutes. I guess it could be a vacuum leak, but I feel like it isn't...

wcs, I have tried cleaning the ESS and a reset. It didn't help :/

9krpmrx8

iTrader: (46)

Okay, is your coolant temperature sensor the original one?

RIWWP

iTrader: (2)

I suggested the MAF because it works by heating an element, which is then cooled by the air passing it. How much it is cooled tells the ECU how much air is passing. If it's not being heated up enough by the time the ECU expects it to be, it could be seeing the cooler sensor as lots of air being ingested by the engine. This would cause it to dump lots of fuel in and show a high calculated load. The O2 sees how rich it is, and is trying to lean it out.

Cliffjumper126

Yeah 9k it is

RIWWP, that makes sense. Since it eventually runs correctly, then that would be when the heating element is finally warm enough to read correctly?

RIWWP

iTrader: (2)

That would be a guess, but yes, that is what I'm thinking. You can swap MAFs with an 8 that runs fine to validate?

I have trouble pointing the finger at most of the usual cold-start suspects, since it runs fine at first, and only once the air pump turns off and (presumably) it switches to closed loop from the initial open loop from startup that it runs into an issue.

Since calculated load is heavily based on the MAF data as compared to maximum MAF, it has to play a role. If it's not a direct roll, something is fooling the ECU into thinking it is.

Cliffjumper126

No, I don't have anybody I can swap with, sadly. There aren't really any other 8s around here (which is one of the reasons I bought it). I don't have the money to go buy one at the moment, and I won't until I get paid on friday. If it doesn't fix anything, I can always return it

9krpmrx8

iTrader: (46)

Then I suggest replacing it eventually. It is a very important piece of the puzzle and it is cheap. But at the same time, I highly recommend against trying to get the problem diagnosed by trying different things suggested here. You need to follow the proper procedures. Problems with the MAF sensor itself are very rare.

Cliffjumper126

Good idea, especially since they're only ~$25. I do know not to just start replacing things right and left lol, I don't have the money for that, but there isn't really a procedure for this issue though, since it's so strange. Seems to me like the only way to troubleshoot would be to just try different things.

9krpmrx8

iTrader: (46)

Well if the ECT is done then it can cause all kinds of issues. Can you read AFR's with your scanner when it is doing this?

Cliffjumper126

All I can read are fuel trims. And there has been a new development. Last night it just started doing it all the time. And unbelievably, it actually threw an accurate CEL for once. Last night, I got P2096--Post Catylyst Fuel Trim System Too Lean.

9krpmrx8

iTrader: (46)

I suggest sending an oil sample in to get tested to Blackstone Labs or doing a coolant system pressure test.

Cliffjumper126

Ok I'll do that. I edited my post btw

Cliffjumper126

Now I've got P0172--"system too rich" as well. I'm going to take a plug out and look for coolant