Getting the Porsche running

[Mac-B]

Hi,

got the 911 running, after sorting some stuff out it idles quite good at least.
As I wrote i try to run alpha-N only. I would not need to but the first experiment is to try it out. So please don´t come with "go to MAP", i am willing to get it running with alpha-N. Only if it does not work at all i will switch to MAP.

I fixed the IAC PWM duty cycle to approx. 40% by disabling PID-control and all the rest, think that caused some issues in the first try (as the alpha-N can´t see what the IAC is doing ?)
I also switched off the EGO-control.

Now i get a quite stable idle, the engine runs best with a lambda of 0,9, when i go to 1,0 it gets more unstable.
BUT as soon as i try to accelerate it goes very lean and misfires.
So I raised up the acceleration enrichment to max values. Is this normal for alpha-N only ?
Also after auto-tuning some values around idle and some rev-ups it seems that the VE-table will look like a cliff.
Is this also normal for alpha-N ?

What would be the best way to get it calibrated ? I read somewhere that i should turn off all acceleration enrichments ?
Can you guide my through the first steps and give me the special hints for aplha-N-only ?
Maybe something else is not configured right currently ?
Do you know somebody with alpha-N-only setup where i can look into a config file ?

Here are my log files:
http://www.vems.hu/wiki/index.php?page=MembersPage%2FMacB%2FLogs

Thanks for your help in advance !
Greetings
Marcus

[GintsK]

You are still on 240cc injectors? So start with accordant required fuel....and read validation warnings which most probably are true...

And yes...: start with speed density, get it running, feel more comfortable with your engine and your VEMS. After that you can try out next levels.

Gints

[Mac-B]

Hello,

yes i´m still on the 230cc injectors which should give me a req_fuel of 14,9 when running fully sequential.
As I´m running batch fire, alternating the banks every crank rev the value of 14,9 leads to a way too rich running engine.
So I lowered the value down to this small number. With that the engine idles quite nicely with the correct lambda value.

Regarding the validation error, just saw this now. When the engine is running i have 14V and i tried out deadtimes of 800-1200 for 14V without any major change in running.

Marcus

[GintsK]

For me seems you have to stay at 14.9, because of configured divider =3. It means injector is fired once par cycle anyway.
Now you have too high dead time and 10 times too low required fuel - compensated by misshaped VE table. Once you touches accelerator engine runs deadly lean.

[Mac-B]

The problem is that even if i run a req_fuel of 7.5 i can´t get anywhere near lambda 0,9 or 1 in idle.
Regarding the injectors 0 280 150 158: they are rated at 229cc/min, but i found another spec stating that they deliver 4.72 mg/pulse (ti=2.2ms) and also the deadtimes are specified here:
http://www.the944.com/injector.htm
The mg/pulse are much higher then you would expect from the cc/min
With the spec from the link, what should i configure for req_fuel and for the deadtimes ?
What would be better simplified or traditional config ?

I did a new run yesterday where i ended up with req_fuel of 6.4 and the VE table from around 20 in idle to 155 in WOT. With that i found some nice spots where the car was running quite good, it was possible to make full acceleration with WOT up to the 5.200rpms:
http://www.vems.hu/files/MembersPage/MacB/Logs/130831/v3.3_u008234-2013.08.31-12.20.58.vemslog

[GintsK]

From your link:
With zero lag you would get 6.01mg/2.2ms pulse (from 164g/min). You have just 4.72 So effective injection time from 2.2ms pulse is 1.72ms and lag time is just 0.48ms. We do not knew at what voltage. But use this as starting point around 12...14V. Also Lagtime=topen-tclose=0.40ms Even lower! You previous config had about triple!

In fact you can use twice lower required fuel as calcualted. Then VE table represent doubled Volumetric efficiency. And you get better resolution. Then you can expect max values in table near 200. With calculated - 100. Less for street engines, more for well prepared race ones.

Gints

[Mac-B]

Hi,

now the picture gets clearer. Uploaded my last config and logs:
http://www.vems.hu/filethingie.php?action=list&subdir=MembersPage/MacB/Logs/130904

Got to do some fine tuning now, engine runs quite nice when its warm.

Any other tips ? Do you see any major issues ?

Marcus

[GintsK]

Something is really wrong with lambda signal. Usually it is way smoother.
May be uneven fueling between cylinders? I suggest to check plugs. If colouring differs - then injectors probably gone.

And also TPS signal is noisy for some reason.

Gints.
 

[Mac-B]

For your information: EGO correction and IAC PID are diabled to tune the table.
The log contains some Live VE tuning also.

I´ll do a cleaner log this afternoon.
Nethertheless I´ll check the injectors soon.

[GintsK]

I am not about Lambda value. Your lambda signal is very noisy. TPS signal is noisy.

[Mac-B]

OK. Understood.

[Mac-B]

Short Update:
I´m doing a motor overhauling currently.
The injectors are already tested on ASNU. Some had a bad spray pattern, the rates were OK.
Main reason for the overhaul was oil burning. I´m watching this since the last years. Now that I proved that VEMS works, it is time to do it.
Spark plugs on cyl. 2 and 5 were quite bad (black)

So this might be the reason for the oscillating lambda-signal.
I´ll keep you updated.