Intermittent Fuel Cut

s300 and SManager software questions & answers
Post Reply
User avatar
jbretthowell
Posts: 2
Joined: Mon Apr 02, 2018 6:01 pm

Intermittent Fuel Cut

Post by jbretthowell »

I’ve been having an intermittent issue for awhile now that I can’t seem to track down. It’s a 91 Civic with a normally aspirated D16 running on an S300 v3 that I autocross and occasionally drive on the street (though the problem seems to only manifest while autocrossing). It seems to happen after at least a few runs (car is hot), and after having been run hard at full throttle, then coming off the throttle to slow for a course feature, then getting back on full throttle again. The motor just seems to lie down - like it’s not being given any throttle. Usually after pumping the pedal a few times, it seems to eventually clear after a couple/few seconds. Sometimes it’s good for the rest of the run, but occasionally it will happen again later in the run.

I finally managed to capture a data log (attached) of it occurring last Sunday (around 0:35 to 0:40), and despite the throttle being at 100%, my injector pulse width and duty cycle both drop to zero (A/F ratio jumps to 19.09 which makes sense if there’s no fuel). RPM and other sensors all seem to read normal otherwise (indicates fuel pump is on, etc.).

WBO2 sensor is brand new (connected through an Innovative LC-2), as is the TPS (new) and main relay (new). Injectors are RDX 410cc (I believe dead times are properly configured), and fuel pump is a Walbro 255LPH (new as of last year). I ran it with the tank full this time to try to rule out a pickup problem.

It was tuned when it all went together a little over a year ago, and I’m going to take it back to my tuner in the next week or so, but because it’s occurring intermittently, I doubt I’ll be able to replicate it on the dyno (or even the street since it seems to only happen when the car is being driven hard).

Some have suggested the fuel pump may be cutting out, but I question why that would cause pulse width and duty cycle to drop to zero - I mean, I recognize that the fuel pump cutting out would mean there wouldn't be the expected fuel pressure reaching the injectors and therefore not enough fuel being delivered, but I don't understand why that would cause pulse width and duty cycle to drop to zero. Something seems to be cutting the injectors, and since the ECU does not seem to monitor fuel pressure as an input that would control the injectors, I would expect pulse width and duty cycle to report normal values in the data log (the injectors would still open even though less/no fuel would be delivered since there is no pressure). Maybe I'm thinking about that wrong - please feel free to set me straight.

Could this be caused by the main relay? While the one in the car is fairly new, it's not from Honda, and I know non-Honda relays can be suspect (I've got a couple Honda relays on order and am planning to swap the relay to test this theory).

Anyone have any suggestions of what to check? I’m happy to provide more info as needed. I've attached my calibration as well as the datalog.

Thanks in advance.
Attachments
stumble.s3d
(634.92 KiB) Downloaded 68 times
20210522_ready.skl
(43.83 KiB) Downloaded 72 times
TTR
Posts: 106
Joined: Wed Mar 31, 2004 8:48 pm

Re: Intermittent Fuel Cut

Post by TTR »

Disable the lean protection and give that a try. It definitely eclipses the threshold for it. Looks like the wideband signal is wonky as well, it shouldn't be stepping the way it is in the log. You should be using D10 ELD for a wideband input after removing R136 and R138 resistors inside the ecu. Looks like the motor is lean at wide open.
User avatar
jbretthowell
Posts: 2
Joined: Mon Apr 02, 2018 6:01 pm

Re: Intermittent Fuel Cut

Post by jbretthowell »

I think I figured it out - TTR got me looking in the right place (thanks).

I had lean protection triggering at a relatively low AFR and engine load that I believe was getting triggered by a overrun fuel cutoff that was kicking in at high throttle setting relative to the configured minimum TPS reading (that value was "learned" by reading when the TPS was just replaced) with a delay that was longer than the protection trigger time.

The overrun fuel cutoff was getting triggered at low but not closed throttle openings triggering a lean condition that lasted longer enough to trigger the protection and cutting the engine until the throttle was released long enough to trigger the restore...I was pumping the throttle, but not lifting completely off it long enough to restore it quickly...then when did finally lift long enough...

This makes sense and fits perfectly with what I am seeing in the datalog and what I've experienced recently. Fingers crossed my tweaks will solve this issue.
Post Reply