Another idleing engine problem

[jannoke]

3000GT, non-turbo, DOHC, 1995

What are the problems and what I have done so far:
(all degrees in Celsius.)
1) If ICV is connected, it will usually start then air is moderately cool (15-18 degrees) without any apparent problems.

2) If engine runs for about a minute and is shut off then it will not start or will start after few times of cranking (continous cranking will not help, need to crank few seconds, wait crank again).

3) Pushing gas pedal while cranking makes it worse - if engine even starts to allow to run, pushing the gas pedal will kill it.

4) In cold weather (5-10 degrees) it will not start at all or takes a lot of cranking to do so.

5) Ofter then I stop cranking fuel ignites in the excaust. (So I have fuel)

6) If engine gets hot (20 degrees room temperature, 2x fans have been run) it will start to skip randomly. For 0.2-1 seconds every 20-30 seconds. It usually lasts until pause is big enough so it can't recover back the engine speed and it stalls.

7) Sometimes idle speed is ~1400rpm and stays there from cold to hot (20....100 degrees).

8) If driven then there is no apparent problem with engine on higher RPM's (1500+). I have driven it only few hundred meters near garage (since it's ripped apart at the moment). Engine has power and everything seems OK.


What i have done:

1) I have dissasembled the ICV. It's in good condition. It has a decent 29 ohm resistance on all of the coils. There is no wear marks or burn marks on any of the coils.
2) I have made a led indicator which consists of 4 leds, that is connected to the 4 outputs from ECU that will drive the ICV motor. ECU seems to control the outputs just fine (if he wants to)
3) I have removed the ICV from the engine. And have put my hand instead of the ICV to imitate ICV.
I have monitored the ICV on the same time:
*) If ingnition is on then ICV will not move.
*) If engine is cranked ICV will start to open the valve to about halfway.
*) If engine runs and i let more air into the engine (by using my hand as valve) then it will compensate somewhat on ICV. If i let less air in there then it will try to open the valve. It seems that it works, but seems "lazy".
*) If engine is stopped and going from ING=> ACC then it will close the ICV valve all the way.
4) I have removed the motor end of the ICV valve, so I can adjust the valve manually. It works. CUrrently i have set it to some medium position and it always starts, but runs badly for few seconds because i have set it to run about 900rpm warm , sob obvioulsy it's a little hard to work with that same air cap then engine is cold.

So ICV valve seems to open and close just fine. And ECU can control it also. There seems to be something else that messes up the ICV controlling, so that it's probably too closed (open?) for the engine to work. And it seems that the problem i'm looking for is not just something that is "broke" , but it's deviation from good working curve.

I see from service manual that middle windings (the "negative" power side of the ICV motor. The positive side comes from ECU) comes from Camshaft sensor for one coil and crankshaft sensor from other coil. Can I measure these sensors somehow?
I have amateur experience for bigger car engines like this ( i have rebuilt small 125cc motorcycle engine). I have very good skills on electronics and schematics.

I have access to Bosch OBD hardware and OBD shows no errors.

So..first post in here..what do you guys think?

[AlbeUrDemise]

All this ICV info. Have you considered the MAF- (Mass Air Flow sensor)? Un-plugg it from the intake during idle and see if you still get this problem. My buddy's MAF was acting up and it did what yours is doing. If It straightens out you'll need a new MAF.

[jannoke]

Will it work at all if MAF is disconnected? Actually my friend gave also hind about MAF sensor, but i was wondering how I could check it, because i know it shows on the graph if air volume changes. So it kinda works, but how to check what are the correct values is a different thing. As said - I do have computer hooked up and i can get the data. Maybe datasheet have some values for MAF sensor?

[AlbeUrDemise]

You can in hook the MAF at idle and it should straighten out. It will idle a bit higher but nothing serious. You shouldn't drive with it unplugged though. It will cause the motor to die often, and you have to keep it going with rpms constantly, I would at least check though. Could be simple fix. Just start the car normally and let it warm up if you can. Once it's starts idling funny unplugged the MAF and see if the idle stabilizes. Like I said rpms might increase a bit. If it does plugg the MAF back in and see if the idle returns to retarded. lol

[AlbeUrDemise]

You should be able to. Check values based on it being unplugged and plugged in. Then search the variable data to see what the standard values are.

[jannoke]

So what I have learned:
* Car will not run without MAF connected. As soon as i disconnect the MAF engine dies in few seconds or makes 2-3 swings reving up and down until it goes so down that it stalls.
* Car will not start without MAF. Seems like every cylinder fires just once and then it stalls. If i keep on cranking then this continues - they fire once, then 1 second pause and then they fire again, then 1 second pause. In a loop.

* If car is hot (90C coolant) and MAF IS connected then it will actually act the same at the beginning - If I start cranking it will instantly fire all cylinders then for one second it will just crank and then it will fire again and engine start to run stable.

Seems like ECU has some kind programmed mode to burn fuel at the start of cranking and then switched over to sensors to take over, but seems like something is not giving the correct information so it stalls.


* If engine is cold (15 degrees celsius) then it will start up instantly and stay working (no pause after initial firing). Like nothing is ever wrong with the engine. And then I let it run for 20 seconds. Turn off and crank again in 20 seconds and it cranks 5-6 seconds...until cylinders start to work one by one. It starts up and runs stable.

I don't believe it ICV since it moves just fine (unless it's just slow for some reason (lost some of the magnetism?)). And MAF is not too cheap to get in this side of the world for just a hunch.

[CoreyB]

Its the engine coolant temp sensor, its faulty.

[jannoke]

Its the engine coolant temp sensor, its faulty.

Nice try, but allready tried that: ECU reports temperature correctly atleast by looking at cold car (cooland temp = ambient temp). Also made a test using adjustable resistor instead of the temperature sensor:
* I set the engine temp the same as ambient temp with cold engine.
* I increased the sensor temp reading with 10 degree step from 10 to 80c on a cold engine and it din't make give any improvement. Also it didn't see any improvement then it was hot.
* If I disconnected the resistor or sensor then it went to factory programmed default mode which I think is something around 80c and fan is constantly on. If I recall then it would act the same way - it would not start at all or start poorly but once engine runs it's all stable.

ofcourse if i sat the temp sensor way off the actual temp then it was catastrophically worse. But if I were close enough of real temp then it seemed to act as if real temp sensor is connected.

[jannoke]

Does anyone have readout how barometric pressure sensor should change. I currently have it fixed 100.6KPA which is 1atm (14.5PSI for you US guys) as is should, but it's not even budging if I hit the throttle. I have never been seen it to anything other than 100.6KPA. I do have cold air intake , but still...
Could anyone take a look at theirs - should it be so rock stable ?

[jannoke]

Barometric pressure sensor seems to work. Wrapped plastic around intake and it shows vacoom then revving.

I was messing around with the engine to see what could be the problem. And for some reason i tryied disconnecting the maf again and what do you know - it stayed running without any noticeable problems. Hitting the gas would increase rpm with no problem. This is the first time engine actually stayed on (i have tryed diconnecting the maf atleast 6-7 different nights then i have been working on the engine). So something is definately randomly faulty.

Oxygen sensor is totally faulty. It seems to work on somewhere aorund 1000-1500rpm. Anything over or under that is totaly crap. So I disconnected it. Will need to wait for next time (probalby tuesday) to try the engine running cold. One theory would be that oxygen sensor is giving bad data. Since oxygen sensor actually need to start working then it reaches 600F then it could be that it is allready giving out bad data then running cold and messes up the ecu.