To Hopefully End all Rich Threads

[Speed Junkie]

This "SR20DET running rich" thing is all over the forums and they all end up being forgotten and no one ever posts a solution. Or they post they typical "check your plugs" comments that lead nowhere.

Context:

I have a kouki S14 SR with S15 turbo, ehxhaust and typical IC piping, S15 injectors, Walboro 255, Z32 MAF, Kouki S14 ECU with Nistune board, and wiring specialties pro harness which is about a year old as well. The car was tuned by Matt Shue (www.shuetuning.net) and has ran correctly for about a year.

I have replaced O2 sensor with an OEM bosch one that is a few months old. I've also replaced the ecu temp sensor. The engine and surrounding engine-related wiring is in good condition and not hacked up or jimmy rigged in any away. The spark plugs are less than a year old and correctly gaped. Coils are also good.


The Issue:

Five minutes into a road trip I started noticing that anywhere between 10 in/Hg and atmospheric the AFRs gauge would max out to 10.0 (so actual AFR is lower). Between 20 in/Hg and 10 in/Hg (steady cruising load) the AFRs would hover properly at around 14.3-14.7. The car WAS NOT tuned to run like this. Not to mention 10 AFR anywhere under atmospheric is way too rich.

While the engine is running like this, it would idle at 11-10 AFR and if you let it idle long enough it would max out at 10 and struggle to stay on. And once I got going to cruising speed and between 20-10 in/Hg the AFRs would go back to 14.3-14.7.

I later discovered that some times, if I shut the car off and immediately turned it back on, it would go back to normal. But just a few minutes back on the road it would slowly get worse and eventually end up behaving as it did before.

I hooked up my laptop to it using nistune trial software just to see if I could find any obvious faults but it's hard to tell when I dont really know what normal behavior is. BTW, I had and cleared a typical knock sensor trouble code but these codes are know to be common on these engines. Not to mention I didnt have the rich issue before regardless of the code. I did notice that while it was running rich, the short term fuel trim would drop to max -20% over the course of a few seconds, hold -20% for another few seconds, then snap back to 0- -5%. During this time, AFR at the gauge did not change from 10. This would repeat about once a minute. I remember looking at it last year when it ran right and it was definitely not doing this. It seems like the ECU is trying to cut back fuel to hold stoichiometric but something is not letting it. While this is happening, injectors are at 2.xx ms PW (compared to 1.xx when it's idling correctly). I also compared it running with my friends S13 SR (using free consult software) and all the main readings such as MAF v, O2 v, injector PW etc, were showing similar readings and behaviors (during normal idle).

Are there any other things I can look at in nistune that could point me in the right direction? The other free consult software is pretty limited and wont let me see things like short and long term fuel trim but I can always get a new free trial for nistune.

Maybe this weekend I will record a video of the nistune software open while the car is idling rich. and post it as reference.

Let me know what you guys think.

[dbeiler]

The MAF reading is normal, correct?

So it's getting extra fuel into the cylinders somehow. I can think of several possible causes:
Base fuel pressure to high.
Fuel injector leaking down.
Failed fuel pressure regulator dumping raw fuel into the manifold through the manifold reference vacuum line.

[Speed Junkie]

I'm going to double check the MAF reading but keep in mind that the actual injector pulse width was higher than it should have been (2.XX ms VS 1.XX) which is an "electrical" thing and not a "mechanical" thing.

[Perfect Balance]

Just to chime in, (I'm the friend with the s13 in the first post)

It's not a mechanical issue like "injectors are stuck".

Something is causing the injector pulse width to jump up almost a full millisecond from 1.8ish (I think) to 2.5 or so, without any noticable change in any other reading from the datastream that would seem to cause it. THIS is what's causing the rich condition, the change in injector pw. The question is why is it doing this?

[LoSt180]

Could it be a glitch with the Nistune board? Also, it's a 20 year old computer, I'm sure funny things are bound to start happening.

[dbeiler]

If NIStune is commanding a longer injector pulsewidth, you'd be seeing that on the short term fuel trim.

Does the OEM o2 sensor data in NIStune match your wideband data? Rich and rich? Lean and lean?

[Speed Junkie]

I took the ecu apart myself before sending it to nistune and looked at all the terminals and solder points and everything looked practically brand new. And I'm sure they did too. I also inspected it after getting it back from them. Either way, here's a screen capture video I took of the whole thing.

https://youtu.be/qJgtnShU1j8

[dbeiler]

Your ECU and Nistune is working exactly as they should be. The injector pulse is being lengthened as your MAF reading rises. The o2 sensor is reading a rich condition and leaning out fuel accordingly.

You need to completely disregard the injector pulse data and focus solely on the o2 sensor, MAF, and fuel trim data.
You should already have completed my recommended testing on the base fuel pressure and leaky fuel pressure regulator. These are simple tests done with simple tools that can save you a ton of diagnostic time.

[Speed Junkie]

I’ll check those two asap and get back here once I do but I don’t see how too high fuel rail pressure will increase MAF voltage.

[Speed Junkie]

So I checked fuel pressure (placed the gauge after the fuel filter and before the fuel rail). At idle it sits right at 40psi. Which is more than OEM spec of 36 psi due to the Walboro 255. The car was tuned at this pressure btw. I drove around with the gauge connected and it behaved properly. I actually drove it around for like an hour without any issue. Almost exclusively 60mph cruising though. The problems seems to happen at idle or low rpm driving. When it started acting up the gauge did not read any differently and was always behaving in accordance with the vacuum/boost gauge.

With that being said, I drove the car twice yesterday and the problem has in fact gotten much worse (when it happens). So now the car barely stays on and when I hold the throttle at like 2k RPMs it oscillates badly and sounds like it's misfiring. This is all during idle. As I drive off it struggles to even accelerate, still hesitating and misfiring. If I give it full throttle it hesitates and coughs some more until it "pops" out of it and accelerates (still not at the same rate as when it's working). All of this is with the AFR gauge reading 10 AFR (maxed out).

Well at least now I have a much more substantial problem that might make it easier to find a trouble shooting item in the Silvia manual.

Let me know what you guys think.


EDIT:
The Silvia manual trouble shooting steps are pretty much the same for all of the stall/hesitation/low rpm performance issues. I checked the injectors (listening and disconnecting) and they’re good. Checked the coils and they’re also good. Pinched the PCV hose and the rpm didn’t rise and vacuum didn’t change. I’m gonna try to do a vacuum leak test tomorrow but the fact that the issue is intermittent, or at least associated with low rpm operation is what’s throwing me off.

[Kingtal0n]

Just disconnect your O2 sensor. This will throw computer to open loop and ensure you aren't getting a fuel trim which is typical 2-3min into driving.

If that doesn't help you have a boost leak, or the tune is off for some reason now.

[Speed Junkie]

I removed the O2 sensor and it behaved normaly for a short bit and then progressively got worse. In fact I'd worse than with it connected in the sense that when it started to go rich it really didn't come back from it. Eventually it was just so rich that it would barely accelerate. During this time the short term fuel trim was at 0% and would spike to some negative value when I'd let go of the throttle. Which seems normal if that's how the algorithm accounts for decel enrich.

I also tried switching to a different working MAF to no avail. Where else can I go from here?

[Speed Junkie]

Alright so as of this weekend, the problem has been fixed. There were two issues. First, the knock sensor was definitely bad but that wasn't what was causing the really bad issue and bizarre behavior. When I replaced the knock sensor the car improved slightly but still had the original issue. The second issue was the harness ground for the MAF sensor. I remember Matt Shue telling me to keep it disconnected as a precaution but at the time it didn't actually make any difference in how the car ran. Well during this time I had it connected to one of the PS reservoir bracket bolts (which has a rubber bushing) so it was probably a bad ground to begin with. Either way the problem went away entirely when I did this.

I did get some useful information back from Nistune when I had them look at my tune so I'll post it here for anyone that is planning on using them:

"We like to keep K constant (and TP which is calculated from it) near original factory values. This then results in the TP load scales not needing modification

What has happened is the Z32 MAF change increased K where it was, TIM adjusted for injectors, and the TP scales were altered to make the cursor operate in those scales

Resulting tune from this shows large increase in K constant, and the TP scales. I would suggest if doing a re-tune to make the TP scales factory values, lower K constant until the load cursor operates in those scales, and get mixtures correct using the TIM parameter."


These MAF cars are really sensitive but also fairly simple. As I mentioned above, all of the factory service manual troubleshooting steps for rough idle/low rpm performance pretty much point to the same things. Check for vacuum leaks, check for MAF voltage, check for working coils, working injectors, knock sensor, O2 sensor, and proper TPS voltage/calibration. Those are pretty much in order of descending importance. And of course, if your MAF sensor harness comes with a ground, leave it unplugged.

[RalliartRsX]

I was about to come in here and say electrical (in particular grounds). But glad you got this fixed! However, why is your MAF grounds separately from the rest of the harness??

Good info for anyone else :)

[Speed Junkie]

The wiring specialties Z32 MAF harness has a separate ground cable which I think is for wiring loom/shield. Can't remember exactly what it is for but it is not necessary to connect it.

https://cdn6.bigcommerce.com/s-b03t074/products/996/images/2533/Z32_MAF_Connector_Plug_and_Play_Sub_Harness__54807 .1442697435.1280.1280.jpg?c=2

[heywier427]

Interesting. I have the Wiring Specialties loom & Z32 maf with it grounded.

No running problems what so ever. Maybe I'll un-ground it and see if anything happens.

If I dont pick up 67hp, I will just put it back.

Glad you got it sorted.

Edit: my tune is based off of enthalpy, but edited in tunerpro by me when I went bigger.

[Speed Junkie]

Yeah Matt from Nistune did confirm the following on a reply email:

"You should only have 12V, signal and signal ground (ECU signal ground) connected to the MAF from the ECU. Do not connect to chassis ground, otherwise it may mess up MAF readings."


Other fun fact from his email:

"ECU uses a calculation where TP = MAF/RPM x K constant"


TP stands for theoretical pulse width which is what the ECU uses for engine load. This is derived from the above equation.