3rd gen 3sge does anyone know the recommended octane? Or most likely RON since it's not a US engine? I've been bashing my brains in trying to figure out this problem and refuse to go to another race with it. After about 30 hard miles 5k RPM.... the engine starts cutting out lower and lower in the RPM range until I cant rev over 4k RPMS.

I've been running 87 octane, i'm thinking the knock sensor could be detecting something and slowly pulling timing.




If this is the case i'm an idiot, and created a lot of headache for a stupid easy fix.... just sucks cause there's no easy way to test it.

Have you tried 93? I know the 3rd gen 3sgte will only run on 93 pump do to the higher octane levels in Japan. My 3sgte will start to knock with anything less. Maybe your 3sge will benefit by running 93 pump gas.

The compression ratio is 10.3:1, so I'd run at least 91 in it, probably better with 93 since you see long periods of high RPM's.

i know 4th gen 3sge needs 93 octane(98-100ron)

so the 3rd gen which is less compression should run at 91 octane just fine. but i know you guys in every other state but california carry 93 octane.

this is EXACTLY what my car started doin...

I threw parts at it until it became right...

1. changed dirty fuel filter
result: more peppy, revs higher, but still cuts out

2. dropped gas tank, found a big square sheet of rubber in there that the fuel pump had sucked up, looked to be the floor of the intank swirl pot. removed and installed walbro pump
result: fixed the high rev cut out, but the lower revs became rubbish and stumbled everywhere, heavy fuel smell while driving (rich az fk)

3. changed rotor, cap and leads
result: alot more peppy in the midrange and up, a few low rev stumbles went away, but not completely.

4. changed spark plugs, they had become extremely fouled from the major change lean to rich
result: everything fixed, car became very powerful. 2 days later plugs fouled again.

5. powerfc ecu and tune
result: awesome, finally see whats wrong, now that i can datalog my engine sensors. looks like i had also fouled my o2 sensor as it reads a constant lean condition 0.2v, doesnt bother me since i no longer have o2 correction

as for octane. its 100RON on any *-GE JDM ecu

in a 1996 SS-II/SS-III , theres a very noticeable performance drop when running 95RON , after 4000 revs a 3SGE should come on song and start a peaky power climb all the way to 7200, but instead revs flat like a 3S-FE

Well... i was able to go out on the highway for about 40 minutes and get the engine to start sputtering so at least i can test it. Just takes a while.


After it started doing it I dumped a crapload of 108 octane booster in. Still started sputtering but the ECU might need some time to relearn itself. I'll go back out this weekend and try again.


If this doesn't fix it... i'll be switching to an aftermarket fuel pressure regulator and actually get a gauge.

CR, fuel/air ratio, octane and ignition advance are all codependant factors. Most ecu's will retard the ignition until the predetonation stops. If you put a higher octane in the car the ecu will actually advance the ignition. to utilize it until it detects detonation.
Its similar to how the engine determines the fuel trim, it constantly fluctuates between lean and rich to pinpoint the best trim. The ecu is constantly testing the limit of the ignition advance for maximum performance without detonation.
Putting a lower octane in shouldn't negatively impact an EFI motor except for a slight loss of torque. Raising the octane will marginally increase output.

Check the plugs for color, if they are pitting then you have a problem.

the octane fuel booster don't do crap!

they raise it perhaps 1 octane level and maybe 1.5 octane if that!. so if you fill with 87 octane you might get 88.5

its 93 or nothing.

Depends what you buy. Turbo 108 racing fuel concentrate, not octane booster. But yeah the regular stuff is like 1 octane or something.

Not many people run their car as long or as hard as I do. Starting out for the first 20 miles above 5k rpms you dont notice anything. Around 25 miles you start to sense something. 30+ you know something is definitely wrong.

If it's continually detonating, the ECU is going to keep pulling timing until who knows... i'm just assuming there's a point where it jsut wont accelerate anymore, but who knows. It's either that or a super lean condition, caused by some hangup in the fuel system.

The lean condition, doesn't make much sense to me though just because I would think if something was clogged, not high enough fuel pressure etc, it would just do it all the time... which would make my life a hell of a lot easier.

Filled with 93 + the race fuel... still didnt fix the problem.


Swapped computer.... seems like the computer from my spare engine suck and runs on 3 cylinders with the new computer. Next step, new fpr and gauge to monitor the fuel pressure.

Could the fuel pump be overheating? Not sure what fuel pump you are using but I could see where the pump might be wearing out if it's still the original pump. I'm a bit surprised that you don't have a wideband installed in the car, but I guess it makes sense. I certainly wouldn't take the time to look at it mid-rally.

I installed a new fuel pump before sno*drift, much quieter than my last one. I have external pumps since I have a fuel cell.

-6 lines through the interior of the car. Personally i wouldnt waste my money on wide band unless it was turbo'd.

You need to get the engine to manifest the problems and then try to diagnose it while it is still acting up. I would conjecture that it could be the ignition coil, I've seen this exact symptom before. Maybe too hot of a plug could cause this. Test the primary/secondary resistance on that coil with a multimeter, just for ****s and giggles.
Its unfortunate the japanese don't have OBD2, there is so much you can do just by reading the realtime data out of the ECU

Entire ignition system was already replaced.

firstly czwalga got your pm sorry for the late reply. 100RON or as close as you can get to it 98, or 95 if nothing else available. I forget what our rating works out to be on the US octane rating or whats the closest over there.

any chance of getting a video of when its happening?

when did you notice it started doing it? was it after some work you've done?

I may be able to get a video tonight. Problem with the video was that i wasnt trying to make it happen as i was racing, i would just shift as soon as I felt it starting to cut. So its proably not that noticeable on the audio.



I ran an event last january, it was fine. Car pretty much sat until june... in which it start happening the 2nd day. 1st day was only about 40 racing miles but it didnt happen the 1st day at all that I noticed. Happened on the 2nd day which was like 100 racing miles.