Old Hyundai tuning issues

At my wits end on this one. Car in question: Tubey, my 2003 Hyundai Tiburon GT 2.7V6 6-spd. Car has an intake, headers, no cats and no rear O2 sensors. Has run great for about 7000 miles. 1800 miles ago I replaced the fuel pump with a new OE and a new coilpack. Car has new wires and plugs, fuel filter is part of the pump assembly.

About 2500 miles ago the car started to have a hard starting issue. If you held the throttle down to turn off the injectors it starts hard. Had an issue with the IAT plug that had come loose. Thought that was the problem but it wasn't. Car had 220k miles so I decided to replace the fuel pump and the coilpack due to originality and age. I also have a hot track issue where it drops a cylinder after 15 minutes of WOT track use above 70* and hoped this would help that as well. That is always cured by a half lap at 70% use and then it's good to go for another few laps. 

So, car made the 900 mile trip to NOLA fine, ran the event with the hot track issue of dropping a cylinder. Fine, that issue isn't fixed but due to events that happened (splashing through a large puddle on track) I feel this one is a heat soak issue. Since it's been here since last year I'm not inclined to believe the two are related. Once the car came home (900 mile return trip) I ran the codes and had new ones. Now we have a P1166 and P1167 to go with the rear O2 sensor codes that have been there from the beginning and a Lean condition on both banks. Car is hard to start but once running runs great. 

I've checked for vacuum leaks, replaced the catch can for the original (also new) PCV hose.Put new clamps on the intake as a few wouldn't stay tight. I changed out the MAF at the same time as that CAN be an issue with these codes. Reset the ECU and for about 2 days it started normal. Starting Yesterday it went back to the hard starting as before.

I'm at the end of my rope here. What the hell? I love this crappy crappy car. I hate when it does this but I can't be mad. It's 18 years old with 220+k miles of abuse. 

If it had low fuel pressure could that cause both hard starting and the lean condition at WOT? Maybe? 

New data point... Long Term Fuel trim on B2 is stuck at zero while B1 is showing -0.79%. 

mazdeuce - Seth said:

If it had low fuel pressure could that cause both hard starting and the lean condition at WOT? Maybe? 

Problem is it's not lean. IT's really really rich, especially on start up. 

Is it possible that the O2 sensors are faulty, giving normal readings while actually being lean? Just throwing things out there. 

In reply to barefootskater (Shaun) :

I've got one new OE one to try. The rear sensors aren't there to give readings so they should be consistent.

But the proble with that idea is the start up. the O2 shouldn't have any input for startups.

In reply to bobzilla :

I'm going to go a bit different direction - do you know which cylinder is dropping when you're on track? I'd be curious if that always happens to the same cylinder, that you may be able to track it down to an injector problem, stuck valve, lifting head, whatever. 
 

If it turns out that it's always the same cyl having issues on track, I wonder if the problem could be worsening & now presenting its self as a hard start condition too?

bobzilla said:

But the proble with that idea is the start up. the O2 shouldn't have any input for startups.

Not directly, but is it possible the ECU applies the learned fuel trims to everything, including startup?  I know at least some do, so if the trims get skewed way rich, it ends up running rich even when it's in open loop at start and WOT. 

Pete Gossett (Forum Supporter) said:

In reply to bobzilla :

I'm going to go a bit different direction - do you know which cylinder is dropping when you're on track? I'd be curious if that always happens to the same cylinder, that you may be able to track it down to an injector problem, stuck valve, lifting head, whatever. 
 

If it turns out that it's always the same cyl having issues on track, I wonder if the problem could be worsening & now presenting its self as a hard start condition too?

Nope. It doesn't record that there's even a miss when it happens. 

Not a Hyundai, but somewhat similar in behavior on my sons SX4. 

I'll start with the short version, followed by why I'm asking

TL:DR

Does Hyundai use downstream sensors as a data source for fuel trims?

Could you have an upstream sensor that is biased and appearing to read correctly but with an offset causing the mixture to be off?

If reset to factory baselines does change the behavior? Does it continue to start hard ?

Back to basics: grounds, clean throttle body, fuel pressure, temp sensors (intake temp and coolant temp match on cold start), etc. 

Long version of what let to the questions:

Suzuki SX4 a Rich condition at idle, rear O2 showing <1 volt, upstream O2 showing normal readings (close slightly negative but close to 0mA).  Replaced rear O2 sensor and due to cable length/routing added extension that moved it further from the cat and out of the flow. STFT was +9%. Cleared codes and no change  

We disconnected the battery, and upon restart it wouldn't idle, and developed a whole new set of codes. 

Cleaned MAF and removed and cleaned throttle body (no separate idle circuit), and after disconnecting battery to get back to factory defaults it would idle again. The rear O2 sensor code returned. 
 

Checked all circuit voltages, etc and decided to replace the upstream O2 even though it showed proper readings and behaved correctly when throttle was opened and closed. Apparently the sensor was bias and thought it was lean even though it was causing the ecu to run the engine rich. 

In reply to No Time :

The car was fine for a year and thousands of miles with no rear sensors installed at all. Still not installed but having the issue. I've read conflicting data on whether they use the rear sensors for LTFT. The Fuel pressure isn't easily checked. they don't use a normal type of fitting to check. I may pull the battery cables and reset again and see if it changes. Intake and coolant temps are correct (or within a couple degrees) of ambient when starting cold. 

I'd look at live data for the upstream sensor to see if it's behaving properly. Is it a wide band or simple O2?

Does the Tiburon have a separate idle control, like a GM, or is it drive by wire with idle  handled by the motor on the throttle plate?

The Suzuki uses the throttle plate, and I didn't think the throttle body looked bad (especially compared to carbureted vehicle from my youth), but it made a big difference in how it started and idled after disconnecting the battery. Over time it would learn to compensate for the crud, but disconnecting the battery would put it back to where it was hard starting and not idling. 

In reply to No Time :

cable and old school IAC, Idle's fine once started. Starting requires holding the throttle open to start

I have had 2.7's of that vintage injectors fail sticking wide open, also had a couple ecu's fail holding injectors open, 

mine were 2001, to 03 Sonata's

In reply to warpedredneck :

It would be pretty cheap to have them sent out & cleaned, and definitely not wasted $ regardless. 

bobzilla said:

New data point... Long Term Fuel trim on B2 is stuck at zero while B1 is showing -0.79%. 

Exactly zero, or pretty darned close?  0.00000 would indicated that something isn't doing anything- which is really, really odd.  Assuming that the -79 is taking fuel out, that matches your starting procedure of going WOT- which should be turning the fuel off during crank.  Old carb trick that all OEM's picked up- a way to turn the fuel off when starting.

In terms of when the fuel trim is really, really relied on- when the O2 sensors are not on at all- cold start and idle.  So all of the learning will be applied there.

The rear sensor does have an impact on fuel trim- but in that era, I doubt that it will force the system to run rich when it fails.

Based on all of the info, I think it's running rich.  really rich.

alfadriver (Forum Supporter) said:

The rear sensor does have an impact on fuel trim- but in that era, I doubt that it will force the system to run rich when it fails.

In this era, Hyundai's electronics were too dumb to recognize that <.5v or >.4.5v on a 5v-referenced sensor meant there was a wiring issue so that sensor should be ignored, so random opens or grounds would do some amazingly bad things.  I had one that would kick down to 1st gear at 35mph when decelerating, whacking your head into the steering wheel, because there wasn't any fault recognition code.

I would not assume either way that they were using the rear O2s for trim, or that if they were, they bothered to stop trimming if there was a sensor fault.   The electronic controls functioned a lot like the script for "Brazil" in their bureaucratic chunderheadedness.

Pete. (l33t FS) said:

 

I would not assume either way that they were using the rear O2s for trim, or that if they were, they bothered to stop trimming if there was a sensor fault.   The electronic controls functioned a lot like the script for "Brazil" in their bureaucratic chunderheadedness.

The reason I think they do is that everyone was doing it by then for quite a few years.  We were doing it in 1996, and have been ever since.  The amount of trim is pretty small, and the failure  mode is to just stop doing it.  At least then- now more is done to compensate for a dead rear sensor.

Kia was buying their ECU's, and I know every system on the market used the rear sensor as trim- it's a pretty key part of making emissions as we all learned how to do it.  2003 was at the beginning of LEVII, Tier 2 emissions- so everyone was doing everything we could to meet them.  

alfadriver (Forum Supporter) said:

bobzilla said:

New data point... Long Term Fuel trim on B2 is stuck at zero while B1 is showing -0.79%. 

Exactly zero, or pretty darned close?  0.00000 would indicated that something isn't doing anything- which is really, really odd.  Assuming that the -79 is taking fuel out, that matches your starting procedure of going WOT- which should be turning the fuel off during crank.  Old carb trick that all OEM's picked up- a way to turn the fuel off when starting.

In terms of when the fuel trim is really, really relied on- when the O2 sensors are not on at all- cold start and idle.  So all of the learning will be applied there.

The rear sensor does have an impact on fuel trim- but in that era, I doubt that it will force the system to run rich when it fails.

Based on all of the info, I think it's running rich.  really rich.

Does this car actually have an upstream sensor for each bank of the V6? I know my GM V6's of that era would only have a single upstream, but the ECU would still have a bank 2 that would report a zero value.

In reply to alfadriver (Forum Supporter) :

Yes. Showing absolutely 0.00. Watching the sensor 1 read outs on the way home and they are fairly consistent .12-.93v. The stft is wild though. Bank one is typically -3% to +5% while bank two is all Positive 2-20%. 
 

i have a spare set of injectors somewhere from the second engine. I may try replacing them on back 1 since they're the easiest to get to. 

In reply to FooBag (Forum Supporter) :

There are supposed to be 4 sensors. One upstream and down stream of the cats in each bank. This car has no cats and the header had one of the o2 pointed at the starter. 

bobzilla said:

In reply to FooBag (Forum Supporter) :

There are supposed to be 4 sensors. One upstream and down stream of the cats in each bank. This car has no cats and the header had one of the o2 pointed at the starter. 

Just to be clear, you have two upstream installed and  no downstream O2 sensors? 
 

Edit: it looks like the upstream sensor is the same for both banks. Before changing injectors you could swap sensors between banks and clear the learning to see if the fuel trim behavior follows the sensor or is consistent based on the banks. 

bobzilla said:

In reply to alfadriver (Forum Supporter) :

Yes. Showing absolutely 0.00. Watching the sensor 1 read outs on the way home and they are fairly consistent .12-.93v. The stft is wild though. Bank one is typically -3% to +5% while bank two is all Positive 2-20%. 
 

i have a spare set of injectors somewhere from the second engine. I may try replacing them on back 1 since they're the easiest to get to. 

Something it making it not correct, even with working sensors- and that's really odd.  If this has an issue with stuck injectors, that would be a good thing to check into.

In reply to No Time :

yes, two up stream (one on each bank) and no rear sensors.

More data points. 3k rpms short term fuel trim is almost identical on both banks. At idle bank 2 (rear bank) is getting a lot more fuel (+5-10%) and the front bank is getting a lot less (-3 to-7%)

pulled front injector rail and pressurized th system. No leaks there. I'm leaning towards the rear O2 at this point even though I'm not getting any codes on misfires, lean nor rich and none of the 1166/1167 codes. Just hard starting.