volvo 5cyl turbo idle issues, at wits end, need some help!

[mdlimy]

The car is a 95 volvo 850 turbo, has an 04/05 2.5t motor, stock bosch 330cc injectors, vems v3 running sequential coil on plug ignition and sequential injection.

From the start i have been having trouble trying to get the car to idle correctly, i am not using the iacv(last resort i will put it back on) i managed to get the car to run and idle very well with the megasquirt ms2extra without an iacv, so i know its not impossible.

i can send a data log/config file to anyone interested, or maybe i can upload it(ill try after posting this)

I cant seem to get a steady idle at around 1k rpm, it always wants to hunt up and down(lean right?), so i richen it up until it stop hunting but it only stop hunting when very overfueled, like .70 lambda, so rich that the wideband reads lean. I have tryed changing spark advance, alsorts of different combinations of fuel ve and spark ve, cannot get a smooth idle for the life of me.
The car seems to drive great, but as soon as i let off the throttle in neutral the rpms fall down low and jump back up and hunt up and down, alot of the time it will hunt down so low that it just dies.

To be honest i am a bit lost using vemstune, the injector settings are different to what i had in tunerstudio with megasquirt, so im not sure how to set them up properly, injector effective rampup?

ANY help is very very very appreciated!


EDIT: here is a datalog,ill make another tommorow and see if i can get some more info on it.
https://sites.google.com/site/mdlimy/my-upload/v3.3_n003369-2011.04.08-19.49.51.vemslog?attredirects=0&d=1

Thanks
Robbie

[sink]

I woun't be very helpful to You but I'm also having the exact same problem    Like You said the car woun't idle AT ALL. I, myself have already checked everything that I could think of - fuel pressure, injectors, ECU configuration and still nothing. Many are stunned by this problem.
What I've found on my car that seems very odd is that the injection time is very very short. I use 630cc/min injectors and the injection time is just 1.0ms - that should be impossible.

How's Your injection time? Are WE really having the exact problem?

[mattias]

Robbie: You MUST flatten out your VE table.  And set your pressure based overrun fuelcut a tad lower, normal idle vacuum must not trigger it, it's impossible to miss that the pulsewidth drop to 0 and the lambda totally lean out repeatedly in the log.

I would advice you to begin with using only 3, at most 4, pressure points in the VE table -  between your normal idle pressure and 100 kPa.
Resolution is not everything and you have a pretty un-eventful engine VE-wise.
Get the VE tuning sorted out (it's not even close) and only add pressure rows if you think you need it - most likely not.

To get a more solid idle, keep the same advance at all load cells around idle. 15 degrees would be nice.. NOT like this :
http://www.savarturbo.se/~mattias/motor/vems_data/vemstune-dev/ve-not-in-tune-robbie.png
(Linux running 'awesome' window manager and VemsTune using wine)

[rob@vems.co.uk]

I've never had a problem with getting a car to idle right without using the IACV, in fact I don't believe that the IACV should be used for anything other than increasing RPM for cold start and to cope with high idle loads such as alternator (heated windows, lights, heater fan etc) and air con.  Well balanced torquey engines often dont need IACVs for anything other than cold start.
Really clean idle can be acheived setting up fuel breakpoints 200rpm either side of the target idle rpm (if the idle target is 850 then: 650 850 1050) and flattening out the fuelling.
Only when its steady and holding target lambda should you turn on the wideband feedback.

[mattias]

sink: Your problem is a bit different, can't really say how.

While you should be using 4.5 ms req-fuel according to calculations on a 2200 cc 5 cyl engine with 630 cc/min @ 3 bar injectors, you're down at 3 ms which means 50% larger injectors..

[sink]

Yes I already understood that to, just that at first it seemed to be the same but I didn't have the possibility to check out His conf with my work computer - no Vemstune.

[mdlimy]

thanks for the killer fast replys Mattias and Rob!

I did notice the duty cycle dropping to zero and i knew that was causing the hunting problems but i forgot to mention it, i jsut couldnt find out what was causing that problem, you were correct on the overrun fuel cut, i set that lower and flattened out the ve table and started tuning the idle and i managed to get it to idle at what i think is a pretty solid smooth 750rpm, it blurts a little every now and then but very small and im not going to worry about it till i get some new injectors. I had to set the timing at idle at 0deg, is this ok? its the only way i could get it to run smooth, any higher and it got lumpy.

So now that it idles well my only issue is when i free rev or push in the clutch and let the rpms drop they fall down below 750 down to about 500 and then come back up to a bit over 750 then settle on 750 and it idles fine, i found that if i gave it more fuel around the 500 ve cells where it falls to it would keep it from completly dying, but i would like to eliminate this from happening at all, is it possible to have the rpms fall and land perfect at 750 and stay?

Any other input on my current VE tables would be appreciated, i didn't have a whole lot of time to play with it tonight.
I know every tuner has his own preference but what do you recommend setting the rpm/load intervals at? boost comes on around 3k rpm.

I appreciate the help a ton, its amazing how two posts gave me the answer to fix my car and i could have it idling wonderfully in less than 30 minutes 

Here is a little log from tonight
https://sites.google.com/site/mdlimy/my-upload/v3.3_n003369-2011.04.12-22.48.57.vemslog?attredirects=0&d=1


Robbie

[mattias]

Don't fix a problem with your overrun fuelcut with your VE table. Use either of the pressure or rpm based version, at the moment your engine gets above 20 kPa at 800 rpm with a completely dry engine,  which is not soon enough to catch it.
Actually you should disable overrun fuelcut while doing basic tuning, if you did you would have caught this problem yourself. KISS = Keep it simple stupid.

If you want to keep a 750 rpm idle, set a column just below at 600 rpm with slightly more advance to let it naturally catch itself. Later on you can activate ignition based idle control to improve this a lot.
Many modern engines can tolerate a low advance at idle, it also helps lighting up the cat for emissions.
One or several injectors can have problems at low pulsewidth too, causing idle problems. You can check the columns gauge, power profile and try to learn if a cylinder is weaker than others.

Use the "Tools -> VE tune by statistics" to get suggestions on where you should be going, use it repeatedly and use EGO correction as long as the engine seems to hit lambda target. You can analyse offline and save the config, later while online opening a dialog of the VE table and pressing the import button to import only the VE table from earlier (or just drag'n'drop that config file over the dialog).
If you want any input, overall you should drop a lot of VE% on load.

Try putting in 5th gear at about 700-800 rpm and use full throttle up until you build boost, adjust that part of the map too - it avoids un-necessary acceleration enrichment later because you are often there in shorter gears but not long enough to tune it.

[mdlimy]

Thanks mattias. Not sure I understand your first paragraph, care to elaborate a bit for me?

I will turn off the over fuel cut, I under stand that, but I set it low enough that its not effecting me at idle currently, so that is alright. Once I did that I just adjusted the ve cells around idle area to get my labda correct, am I doing this wrong?

I copied over my ign map and fuel map from tunerstudio that I had tuned very well on my megasquirt, I rescaled the entire fuel map so that I could keep the correct 9.8ms required fuel, it is a much smoother map, I will still need to tune obviously but this should be a good starting ground correct?

I have my effective rampup set at close to 600, is this ok?

Thanks
Robbie

[mattias]

With your overrun fuelcut settings, the engine had no fuel when settling down to idle rpm and the intake is by that time completely dry. It has no chance to catch itself if the rpm is falling and it has not received any fuel. I can't make this any clearer, you can see it for yourself in the datalog and by testing it in the car it's easy to notice the difference. You were trying to compensate the lack of fuel while entering idle by enriching the lower rpm that it falls through to.. just plain wrong.

Just set the pressure based overrun fuelcut to 0. Only use the rpm based one, but turn that one off too while doing basic tuning.

Your ignition map should be ok, but the VE map from your old ECU will not work out of the box.
With VEMS you have a true VE table, which does not incorporate the mixture you will be running at each load cell, that is the role of the lambda target table. That means your VE table has too high values where the engine previously ran rich, causing it to run super-rich now, that extra % added before is not needed with VEMS, as the enrichment is done through direct calculation using the value in the lambda target table.

I would start out with effective rampup = 0 to begin with. Assume that open time is as slow as closing time. Only if you seem to get a slight non-linear behaviour at idle should you add any effective rampup, I think 200-300 is the max needed in your case.

[mdlimy]

Crystal clear now! Can't believe I wasn't thinking about the higher vaccume on decel and falling revs, ill kill the overrun for now and swap to rpm when its all dialed in.
I think I'm good for now, ill give this all a shot tonight and get a couple hours of street tuning in an see where that puts me.

Thanks a ton Mattias!

Robbie

[Kenny Watson]

sink: Your problem is a bit different, can't really say how.

While you should be using 4.5 ms req-fuel according to calculations on a 2200 cc 5 cyl engine with 630 cc/min @ 3 bar injectors, you're down at 3 ms which means 50% larger injectors..

My guess is the injector rampup (deadtime) is set incorrectly. My engine was horrible to tune at low loads until I got this set half-decently.

[mattias]

Kenny: This is a bit more complex than that, I'm pretty sure something else is up. The req-fuel really tells me something is way off along with strangely short pulsewidth at idle for this engine and injector combination. It's an old board with low voltage flyback, this will be changed and tested.