RNN14 Pulsar strange issues.

[Ian]

Ok, get ready for a long read.

First some backround. I'm a professional mechanic with more than ten years of experience, and lots of time spent diagnosing issues. I've also been building and tuning Nissans and Subarus for many years. My experience with Nistune specifically is limited, but I'm well versed in Romraider, ECUflash, PCLink, and a few other tuning programs.

The car is a 1991 RNN14 Pulsar GTI-R. Started with Stock engine, with Just FMIC downpipe, and Greedy Emanage.

Current owner purchased car. It ran and drove as far as I'm told, normally.

A build list was compiled.

The ECU was shipped to Florida. A type 3 realtime board with feature pack was installed along with a flex fuel board.

The ECU was reinstalled, and the Emanage removed. Either immediately after, or shortly after, (owner isn't sure), a hesitation developed between 2200 and 3000 rpm. If you're heavy on the throttle, it will pass right through the range with very little issue. If you're light with the throttle, there is very noticeable hesitation. More so in lower gears. The lower the load, the worse it is.

The owner asked me to install the following parts.

Fuel pump
ID1000 injectors
Fuelab adjustable FPR
AEM flex fuel sensor (continental)
Z32 MAFS
AEM UEGO
New ignition components
Nismo thermostat

Starting with a raw RNN14 FP file, I used Nistune to auto resize the injectors. Then I resized the MAFS. I 14.7d the majority of the low load fuel map. The car starts and runs. Manually adjusted latency to 990 and it idles/cruises at 14.7. Heavy load is very rich.

The hesitation still exists. As RPM approaches 2200, the wideband shifts leaner and leaner. Light enough on the throttle, and it will push right past 17 to 1 off the gauges scale. It will only happen in a 300-500 rpm zone. Only under light throttle. Above 3 it will rerun to 14.7ish cruising or 10 or 11 to 1 WOT. I have No ECU codes. STFT and LTFT read 0 all the time.


My first thought was o2 sensor. On Nistune it read full voltage all the time. Bingo. So I unplugged the sensor. Now it reads 0 volts. No change to drive ability. Strange. Still no codes. Very strange.

Tried disabling closed loop control. No change. Something is fishy

Then I noticed that the ethanol content read 0. When it should be 10% with regular fuel. Shouldn't effect drive ability, but Stranger still.

Pulled the ECU, and opened. Discovered the green wire between board and diag slot was melted in the middle and stuck to the case. Wire itself did not appear compromised, just shielding, like someone bumped it with a soldering iron. Cut out damaged section and soldered/heat shrinked. No change.

Checked voltages for flex sensor. 12v, 5v, ground are good. 5v unplugged. Sensor pulls line down to 4.85 volts plugged in. Hmmm. Sensor Unplugged the LED on the flex fuel sensor flashed. No LED with sensor plugged in.

Also checked Distributer timing. It's spot on.

Is it possible to have bad o2 wiring, and a bad flex sensor, or is it an issue with the ECU? I don't think my base file is the issue because it was a problem before on the factory ROM. Whoever installed the boards/desoldered the resistors on the back didn't do the prettiest job, but everything appears intact.

If nothing else, I may dump some ethanol through the flex sensor tomorrow to see if maybe it's reading on the low side.

Hopefully someone has some insight for me.

TLDR: It's broken and I hope it's not my fault.

[Ian]

started over with a fresh rom tonight. made injector change and maf change. Tweaked latency, and TIM a bit, then pulled these logs.

1111_2017-06-26_2054_33.csv

1

(52.48 KiB) Downloaded 255 times

NIStune_2017-06-26_2051_32.csv

2

(26.99 KiB) Downloaded 241 times

wideband isnt currently linked to the laptop, but I may go through that trouble if I have to. Dont have anyone to drive for me at the moment, so they got a bit long, sorry. Both start in second gear, and pass through the no no zone. The wideband shifts, from mid 14s to full lean at 2300 rpm or so, and back to normal near 3K.

Is it possible a problem with the load indexes is causing this problem? I feel like it shouldn't because it did it before the MAFS/Injector swap.

[Ian]

Here is a copy of the retry rom based off the 54COF *** FP rom thats in the rom pack. matched the one that was preloaded to the board when I initially connected.

RNN14retry.bin

(32 KiB) Downloaded 245 times

[Matt]

Just some questions. I will look through the logs and BIN file later today

1. Why was the Emange installed if the rest of the car was stock. Was the car ever run without the Emanage in the way?

The ECU was shipped to Florida. A type 3 realtime board with feature pack was installed along with a flex fuel board.

The ECU was reinstalled, and the Emanage removed. Either immediately after, or shortly after, (owner isn't sure), a hesitation developed between 2200 and 3000 rpm.

2. Did the car run the same after the board (but before the Emanage was removed?). Just to make certain that the hesitation was not caused by the installation of the board. At this point it would have been good to stop here and work out the problem before progressing with the build

If you're heavy on the throttle, it will pass right through the range with very little issue. If you're light with the throttle, there is very noticeable hesitation. More so in lower gears. The lower the load, the worse it is.

Sounds like it might be going lean or perhaps timing pulled out during that time. I'll need to check the logs and see what is going on

[Ian]

It's an JDM import. No history really.

I concluded that the emanage was there to compensate for the FMIC manual boost controler and downpipe. As far as I could tell, everything else was stock.

I had No interaction with the car before Nistune install.

The Emanage was disconnected, and Nistune board installed over the winter, and the car was brought to me just a few weeks ago. There was really no turning back.

Before changing injectors/MAFS I did reconnect the emanage and test drove. The WOT fueling was a tad leaner (11.5/1 vs 10.5/1), but otherwise the car seemed no different. The MBC was cranked all the way out when I got it, so 7psi, slow.

I spent some time last night rescaling the load min/maxes, and smoothing some stuff out. I also deflagged the fuel table in case the ECU is still trying to compensate for the unplugged o2.

Got the flex fuel sensor working after realizing I never reflagged flex fuel after starting a fresh rom again, tried like 4 fresh roms scaling things different ways just in case.

Thanks for your help.

[Ian]

Here is the newest revision of the file in the car now. Rescaled load limits and whatnot. Also started set up of the flex fuel/timing maps.

Idles and cruises pretty much dead nuts on stoich when warm. Need to work over the fuel map now that Im happy with my K,TIM,Latency tables. WOT is a pretty solid 12 fading to 10 at redline. Boost is set to 1 bar, (Trust MBC).

RNN14retry4.bin

(32 KiB) Downloaded 212 times

Did a few pulls, but the fuel tank was getting low, and I needed to tune for E85 which will be primary fuel anyway, so I went and dumped some in.

Here is a WOT pull. 3rd Gear, 15 PSI, 50ish% ethanol. I pulls smoothly and cleanly through what Ive dubbed "The no no zone" at WOT. Same lean symptoms if I tried to do the same at low throttle.

8888_1.csv

(9.17 KiB) Downloaded 236 times

It drives me nuts how well the car runs outside of the low load 2500 rpm range.

Here is a log of a super slow pull in second gear. I really wish the owner would have bought a innovate or zeitronix wideband so I could log A/Fr. But the wideband reading is just a symptom. Looking at it, the timing does taper down in the exact same region that the system goes lean, but It stays low.

[Ian]

I was able to get the o2 sensor functioning this evening. Turns out someone had damaged the base of the sensor, and the signal wire was beginning to short to the heater circuit inside the base of the sensor. I was able to separate them for now, but it needs a new sensor in the long run. Fuel trims remain near zero while idling, or cruising outside of the "no no zone"

As I pass through the "no no zone" under light load accel, I still get a lean out, but the STFT increases significantly, and dampens it. The o2 hangs lean for more than 2 seconds before switching back to rich, which is followed by a lean misfire. Looking at the log, I might just think the sensor is lazy, but it reacts just fine in other situations.

8888_8.csv

(9.56 KiB) Downloaded 240 times

Also, I cant seen to read these log files with anything other than Nistune. Im used to reading CSVs in a spreadsheet format, but for some reason, neither of the CSV readers I have will open the files, even though they are tagged CSV.

[Matt]

I can open the log files in microsoft excel (2010) just open as CSV file inside the software

I've looked at both of the logs but many things have been unticked (injection time is the main one I was looking for). I know you cannot get the AFR into the log but can you get the injection time so we can watch that and can you also post a picture of the dyno graph

Timing looks good anyway from what I can see provided. Maps look good also. What if you increase Max Load (TP) table to 255 for now to see if that helps at all. Why was fuel cut table changed?

playback_2222.PNG

(131.82 KiB) Downloaded 3081 times

[Ian]

I'll go pull some more logs in a bit. I've a hunch to test first.

I tried pushing max load all the way to 255, it actually seemed to make it a bit worse, maybe because of the drop in resolution. I'm not sure. That system functions a bit different from the Subaru/Group N/Standalones I consider my forte.

I bumped the fuel cut table up because it would boost spike slightly in 4th and 5th and hit fuel cut if I rolled into WOT at the right RPM. Did I go too high?

[Matt]

The max load table just adjusts the upper limits of how much load is read by the ECU (think of it as a clamp on load). Having it too low will limit your injection at the top end of fueling

Load TP cut will perform a fuel cut when you measure too much load at a particular RPM.

The fuel cut/recover tables are for deceleration (they cut the fuel on decel and then recover when RPMs are lower) to prevent fueling on decel

[Ian]

I mixed up the tables. I adjusted the TP load cut.

I dont think Ive touched Fuel Cut table. Ill have to do some comparing.

If the Max load table is set high enough not to overrun, it shouldn't matter right? It looked really funny to me being at 255 for half the table. I only recently changed it after reading about others changing it. Doesn't it determine which load section of the fuel/timing tables get used vs actual load? At present the stock turbo will never flow enough air to push beyond 150ish on the load scale, why scale all the way to 255? Or am I thinking of the wrong table?

Nistune doesn't give me the option to log Injector Duty Cycle. So I went and pulled a bunch of logs with ECUTalk. Or at least I thought I did, it doesn't seem to have saved any of them. Stupid computers.

Once the o2 is warmed up, the car drives nearly flawlessly. Cruising at 65is kph in 5th hangs right in the no no zone, and It maxes the STFT occasionally as it tries to correct, then starts ramping up long term. Tough to hold it there long as Im forced to increase throttle to maintain speed which increases load and the problem vanishes.

I deleted the Blow Off Valve in case it was leaking. No change. Also rechecked base timing, and fuel pressure. Both were solid.

[Ian]

Is it possible that its dropping to overrun fuel cut mode at such small throttle input levels. As far as I can see the TPS is registering fine, and it reads closed throttle, when its closed, and the closed throttle warning goes out with the slightest throttle input while stationary.

I know that ITBs can make for very sensitive tip in enrichment and whatnot due to the large throttle blade area and lack of plenum/short runner length.

Does this setting have anything to do with it? Hmm, guess Ill just have to try flattening it out. Either that or I can just disable overrun fuel cut I guess. More tests for tomorrow.

Screenshot (7).png

(78.48 KiB) Downloaded 3058 times

[thommo]

I had a very similar issue with my Nissan Stagea
Story goes, the 02 sensor was dead so I bought a generic sensor that was advertised to work with the RB25 Neo.
Car ran OK for a week or so then started to do some strange things.

Cold start and initial driving was fine.
As soon as the engine came up to temperature at light throttle and cruising it would start to hesitate.
During this period the AFR would get leaner and leaner causing the car to stumble and miss more and more.
If I kept the same throttle opening it would almost die.
However as soon as I opened the throttle and brought it onto boost the engine ran perfectly.
I did log after log and and it kept showing the same thing, AFR leaning out, 02 voltage reaching as high 1.3volts (not oscilating) and short term fuel trim going crazy.
I ended up buying a new genuine 02 sensor and the car has never had that issue.
Turns out either the generic sensor was either faulty or was not specifically meant for the car.
My suggestion is your issue probably isnt to do with Nistune itself and more to do with your 02 sensor system.
For roughly $100 its worthwhile changing.

[Ian]

I had a very similar issue with my Nissan Stagea
Story goes, the 02 sensor was dead so I bought a generic sensor that was advertised to work with the RB25 Neo.
Car ran OK for a week or so then started to do some strange things.

Cold start and initial driving was fine.
As soon as the engine came up to temperature at light throttle and cruising it would start to hesitate.
During this period the AFR would get leaner and leaner causing the car to stumble and miss more and more.
If I kept the same throttle opening it would almost die.
However as soon as I opened the throttle and brought it onto boost the engine ran perfectly.
I did log after log and and it kept showing the same thing, AFR leaning out, 02 voltage reaching as high 1.3volts (not oscilating) and short term fuel trim going crazy.
I ended up buying a new genuine 02 sensor and the car has never had that issue.
Turns out either the generic sensor was either faulty or was not specifically meant for the car.
My suggestion is your issue probably isnt to do with Nistune itself and more to do with your 02 sensor system.
For roughly $100 its worthwhile changing.

The only issue with that logic is, I have disabled o2 learning. You're describing the symptoms of a failing o2 sensor. This car behaves the same even with the sensor unplugged, and fuel trimming disabled.

Matt, I had someone drive while I visually logged, since I can't get my computer to save ECUTalk logs.

From 1500-3000 duty cycle % is steady in the 5 to 7% range depending on gear for pull. Second gear will hold 5% all the way through, with injection time right around 3.5 ms. 4th gear is 7% and 4ms, give or take.

The A/Fr's shift from stoich to full lean and back to stoich.

MAFS voltage is steady, timing is steady, manifold pressure is steady. The only thing that really changes is RPM and A/F.

So either a large amount of air is bypassing the MAFS at only 1 RPM range, the MAFS is not reporting properly at only 1 RPM range, or the injected amount is shifting at a certain RPM. Whether the ECU is not commanding what it's reporting or that signal isn't reaching the injectors, I'm not sure. Or maybe there's a big shift in VE at that range, but why only show after Nistune install?

Flattening the decel fuel cut min throttle had no effect. I also played with a few other tables with no effect. Is there a way to fully disable overrun/decel fuel cut? I'm running out of ideas.

[Ian]

Any new theories Matt? Pretty much out of ideas. New o2 sensor arrived today, but that isnt a solution. Wont have time to work on it the rest of this week, as I'm taking my S13 endurance racing all weekend. But Id like to have some things to try when I get back in town.