help! - Rpm jumps on deceleration

[eddyvr4]

I have a problem (2001 Legnum) where RPM is jumping markedly on deceleration, it is intermittent.

Edit: pic of car added for inspiration on otherwise dry topic:

i have a capture here with evoscan. you can see the flat line for throttle where i am off the throttle for a number of seconds, the oxygen sensor (yellow) shows an extended lean condition, and then you can see fuel going in and a subsequent jump in rpm, which is a royal pain in the arse when you are on the brake and the engine is fighting you!

i can understand the ECU adding fuel in to compensate for lean condition, but why doesnt it see i am in decel, and ignore the lean condition? is this a problem with the oxygen sensor? should the ecu be in open or closed loop mode here? should there be some decel enrichment? i need someone who understandes the ECU decel behavior too look at this and make sense of it.

my next step (folowing advise from here) is to disconnect the O2 sensor to generate a open loop rich tune and se if that makes the condition go away.


i have just replaced the ISC, and also there are no boost leaks.

I dont think the rise in load is what is causing the fuel to be added where the revs jump, because i frequently see loads up to around 50 at idle..

it looks like the ECU tolerates a lean condition on decel for about 5 seconds, then adds the fuel, which then brings the O2 back up to a rich signal. So it doesnt look like the O2 is broken.. but i dont know...

should the O2 signal go off the chart on decel?

Any assistance much appreciated..

 

[Jonson]

Is your TPS set properly?

 

[eddyvr4]

Thanks Jon, my TPS logs at 12% for a closed throttle, (ie where it is a flat line above) and 96% for an open throttle, which i believe is normal/correct.

 

[TME_Steve]

Looks like the fuel cut is coming back on a bit strong when revs drop. and the 50 load is getting up there. Try adjusting the tps so its about 13.3% at zero throttle rather than 12. Also log your iscv steps.

 

[eddyvr4]

Thanks for the input steve, thats not something i would have thought of. I agree the ECU looks to be shutting off fuel very promptly on lift off in the above.. ill do some more runs with that in mind and see if i can capture an event when it decels properly, and compare the timings.. (and the TPS value)

I have a pile of logs but unfortunately in those my idle controller had an intermittent short on the wiring loom (long story), so the idle controller pintle was in an unknown position. i thought the idle controller was causing the decel 'hang', but now the short is fixed and idle/ ISCV seems good, but im still getting the decel rpm 'jumps'

i assume i can just rotate the TPS sensor while logging at idle till the value reads 13.3%. Im suprised the ECU would be so sensitive to such a small alteration, but i guess its an simple enough test to reverse..

 

[eddyvr4]

hmm.. silly me.. seems id left the ISCV unplugged... the ISCV pintle must have been sitting dead in a position that was ok for idle, but not good for decel.

Benefit from my shame however is that we now know the above CAN be a symptom of a totally dead ISCV (or an unplugged one..)

For completeness i logged the same lift off event plugged back in and you can see the o2 takes a lot longer to read a lean condition, and there seems to be a much narrower open loop period (where the O2 reads 0) 0.5 seconds, not 5 seconds.

So steve was spot on in his reading of the chart, although the root cause was too much air getting through the throttle body on closed throttle.

so this is how things should look on decel:

Anyway this may be of use to someone in the future who has the idle hang or jump problem.

Note plugging in the ISCV with the car running did not fix the problem, but on stop/ restart then it came good. there were no CEL light with the ISC unplugged.

 

[BCX]

Note plugging in the ISCV with the car running did not fix the problem, but on stop/ restart then it came good. there were no CEL light with the ISC unplugged.

yes, this is correct, there is no feedback to the ECU to check that the ISC/IAC is working correctly. all 6 wires in the ISC/IAC pinout control coils which the ecu pulses to make it move in and out (stepper motor)

Appears the ECU just assumes the IAC will always work and will always be connected.


Also, at idle the ecu is in open loop and will be very rich to keep the motor ticking over. Only once the conditions are right (correct temp, vehicle speed, tps, engine speed) it'll start to use the fuel trims + oxy sensor input (close loop)

Other thing that comes to mind when adjusting the TPS, incorporated in the TPS is the throttle switch so the ecu knows when you've got your foot off the accelerator. If you're to one extreme, the ecu might think you've come off your accelerator sooner than you think and start affecting to decel stuff (this largely controlled by that switch).

When i get a chance, i'll post up a MUT location you can log the throttle switch.

 

[eddyvr4]

Thanks Bill.. good info.

are you sure it runs open loop always at idle? i thought once the water temp (and O2 sensor) are up to temp the ECU does run closed loop at idle, based on what im observing on evoscan). tuners sometimes play with the open - closed loop transition settings to force open loop at idle so you can force a rich idle etc, but this is not the stock behaviour.

From what i read on evolutionm.net at idle, (ie no throttle) when the O2 sensor trace reads 0 in evoscan, the ECU is running open loop, which seemed to make sense.(and can be seen in the charts above) . but im happy to be corrected, id like to understand that properly.

 

[TME_Steve]

Closed loop when up to temp at idle, bill says that too

 

[eddyvr4]

Cool..

Bill mentioned required conditions included vehicle & engine speed & TPS which gave me the impression he was referring to 'driving' conditions before ECU went closed loop, but you have set us straight!

 

[TME_Steve]

Load values up to 65% up to 1000rpm will be closed loop, 75% to 2000 etc etc

 

[eddyvr4]

well, my problem has come back again! Damn it!

however, i just went out and sprayed some detergent around the front 3 injector seals, and pressure tested the intake tract, to find the middle injector seal is leaking/bubbling a lot of air. could only check one of the rear 3 injectors, it was ok, but the other 2 are unknown. a leaky injector seal(s) drawing in air intermittently could cause this issue So, a new set of 6 injector seals.... stay tuned..

 

[eddyvr4]

I replaced the injector seals, (1 was definitely leaking) but then i would stilll occasionally still see the problem.

Its too early to be sure, but i found another issue that so far seems to have fixed it. my BOV was jamming in the open position sometimes, as a result of my water injection system, and the fact i have an turbosmart kompact BOV, this is a sliding piston type BOV, and any residual water from the water injection would pool in the BOV (because its at the bottom of the intake, right next to water injector.) the water would evaporate, and any residual crud would bind the BOV till i sprayed it with some wd40 etc. When the BOV was stuck in an open position, i would get almost no boost (naturally) , but also i got that rpm jump at 1000- 1500rpm when on decel off the throttle.

I dont understand why, because the throttle butterfly was fully closed, so the BOV shouldnt influence the ingestion of air to the engine itself, but indirectly it seems to. also i had the problem when the car was otherwise running heaps of boost, im not sure how it could have run heaps of boost if the BOV was stuck open...

anyway, ive put the stock BOV back in, and no recurrence yet. Note there is nothing wrong with the Kompact BOV, but in the orientation it sits, it didnt tolerate my water injection. the stock BOV is a diaphragm type that isnt likely to bind by a bit of residual water.

so long story short, check for a slightly jammed open BOV if you see this symptom.. at least add it to the list of things to eliminate

fingers crossed this time <mutter mutter, damn idle blip... mutter ...>

 

[mfbamba]

Is your TPS set properly?

I didn't get the memo

 

[VOLK]

I'm still getting occasional idle woes, you've inspired me to log them. What cable you using? Generic obd2 or openport 2?

 

[eddyvr4]

generic obd2 as per hotwire (lee) on this forum had a 'how to' on modding for logging on our cars.

ps my TPS logs always close to 12.15% at closed throttle regardless of whether its 'blipping' or not so ive considered that a valid suggestion but not the issue..

one other thing to note is i worked out the generic mitsubishi replacement ISCV does pass slightly more air when fully closed than the original, the pintle seems to be slightly different on the 2 new ones ive bought. As a result i have a slightly higher cold start idle than i should, and my biss screw is screwed right in. i dont think this is my problem because of the intermittent nature, but if you find the cold start idle is a bit high after replacing your ISCV, its the pintle design is slightly different now. (At least that is my experience) I could tell by physically blowing through the throttle body bypass orifice, with the throttle body out of the car, and comparing the 2 ISCVs. A marked difference....twice the flow easily on the new one in the fully extended position.... (hence screw in of BISS to compensate. yes, very scientific....

Once i can flash my ecu i will reduce the cold start isc steps a bit... its not a biggie but just noting it.

warm idle seems ok, a bit low if anything although usually within say 50rpm of target idle, seems to depend on the day..

 

[VOLK]

Ahh k thanks man. Think I'll make the jump straight to openport 2 for changing the values. Have the ISC step tables been defined in the xml for our roms yet? Would love to reduce em as I'm hitting around 2krpm on cold starts and takes forever to taper off, it's a little annoying. Also using the highly scientific BISS screwed all the way in method

I've got the distinct feeling my SAS (know how to set it?) and subsequently TPS are set incorrectly. It's like the idle switch is being closed on really low but not quite closed throttle and the ISC steps to 0 while on the move at low speed in 1st and 2nd. It's pretty rough, hauls me forward in my seat at low speeds like I've stepped off the throttle. ISC steps for a/c is a bit funky too, hangs for a little while at ~1300 then almost like it's discovering the high idle and then immediately drops to normal. Without being able to see the values, I'm really just using guess work. I'll stop talking about it and actually buy the cable one day.

</thread hijack>

 

[BCX]

Ahh k thanks man. Think I'll make the jump straight to openport 2 for changing the values. Have the ISC step tables been defined in the xml for our roms yet? Would love to reduce em as I'm hitting around 2krpm on cold starts and takes forever to taper off, it's a little annoying. Also using the highly scientific BISS screwed all the way in method

there is some idle stuff defined in my defs, but i can confirm there isnt a 3D ISC table like evo/magna roms have.

 

[TME_Steve]

Eddy try moving your TPS so it's reading low to mid 13% values rather than 12's, 12's would make it open the ISCV more combined with you flowing more anyway.... Maybe....

Having problems again Joe?!?!?! Same as before? Can you describe what's happening?

there isn't enough defined yet to reduce idle speed properly, there seem to be overlay maps on the base maps that we do have but you really need both. I have confidence in Bill though

 

[VOLK]

Heya Steve, different problems this round, cold start is now good after you set my TPS but I think the idle switch is closing before I'm completely off the throttle. Driving along in traffic at low speeds, using very little throttle,
it'll start decelerating randomly. Also seems to be more of a dead zone in the pedal before it starts accelerating again. It's nothing major and quite easy to live with just makes low speed stuff in traffic etc. a little bit more of a challenge when it feels like the isc steps to closed randomly. At least that's what I guess is happening.

Will see if I can acquire an openport 2 soon and post logs.