95 Cobra + Turbo ... Oh Noes!

mach1 · Post by **mach1** » Thu Dec 15, 2005 10:13 am

Man this car is driving me nuts, I am sure that my own ignorance about the Tweecer is the only problem. Here is whats happening. The car will start fine for the most part and drive pretty good until it reaches around 180F (Goes into Closed loop I imagine) then it just falls on its face. My A/F will go to 8:1 and the car will just load up and die, I have to give it ALOT of throttle to keep it running and it still runs horrible. When I am at WOT the car seems to run ok, 11.8:1 ish AFR which is good for my setup. Anyways I have my tune posted up so hopefully you guys can take a peek and give
me some pointers. I have tried to reseat the tweecer, clean the terminals, I even tried another computer(T4MO). Here is my setup.

J4J1 EEC running a modified J4J1 tune.
90MM LMaf
42LB Green Top Injectors
Base FP of 39psi
T70 Turbo
PLX-M300 Wideband (Connected to passenger o2 bank)
F-Cam
TH-400 Transmission

Here is my tune.
http://www.supraspeed.com/final+1hego.CCF

craig429 · Post by **craig429** » Thu Dec 15, 2005 10:41 am

I would change the injector min pulse width scaler to .0003 Seconds from the .0013Sec that you currently have. This will allow the EEC to command a PW down to .3mS otherwise you will run rich because if the engine only needs a PW of .8mS it will command a 1.3mS pulse and provide too much fuel.

Craig

mach1 · Post by **mach1** » Thu Dec 15, 2005 11:06 am

Cool, I think the issue is a bit deeper than the min PW since the car runs fine in open loop but that needed to be changed anyways. Can you please post/email me your tune since you have a Turbo CBAZA strategy car too.
Thanks, Tyler
morperformance@hotmail.com

Post by **2Shaker** » Thu Dec 15, 2005 12:16 pm

Tyler,

I'm just going to give you a stream of connected thoughts here to maybe give you some ideas where to look. I am not able to look at your tune as I write this so apologies if I have some questions or pointers here that would have been obvious if I examined the tune.

First off, it should go into CL after just a minute or two of running, not when it gets to 180*. It has to be in OL if the AFR is steady rich, especially at the values stated. Need to monitor CL and OL using Calcon.

Assuming the above is correct, (it is in OL) the question to answer is why is it going to OL. Is the MAF curve off enough that EEC cannot get/keep control in CL so it falls out of CL? Stuck injector? Does EEC know it is at 180* and the correct ambient temp? Are the O2's working (no EEC errors)? Is the EEC commanding rich or lean? Could it be that the EEC thinks the car is lean and commanding rich?

Without some logs from CalCon its a guessing game. Do you have the RT version of TwEECer? If you could give us some logs we could help a lot more. :? I would be looking for commanded AFR, movement in the O2 sensor values, the TPS voltage values, the MAF voltages, OL vs CL state, and so on.

craig429 · Post by **craig429** » Thu Dec 15, 2005 3:30 pm

mach1 wrote:Cool, I think the issue is a bit deeper than the min PW since the car runs fine in open loop but that needed to be changed anyways. Can you please post/email me your tune since you have a Turbo CBAZA strategy car too.
Thanks, Tyler
morperformance@hotmail.com

If I remember correctly, mine also ran ok when I had the min injector set wrong. As it warmed up it didn't run as well. The colder engine needs to be richer anyways. My testing was done only at idle when I figured out it needed to be changed so I never got it on the road.

I will send you my latest tune later tonight, they are on my other computer. I have J4J1 and a T4M0 based tunes.
Craig

craig429 · Post by **craig429** » Thu Dec 15, 2005 3:44 pm

Also,
This is not your current problem, but this could cause some hot start problems.

In the Fuel_crank_PW_multiply_ECT function, make your 65534 value equal to the 180 value of .001 instead of .0018 because as the temperature increase above 180, your injectors will provide a little more fuel than less.

I also make the last x value (corresponding to the highest Y cell that is usually a very high non reasonable number) the same value as the prior cell (the highest Y cell with a reasonable value).

mach1 · Post by **mach1** » Thu Dec 15, 2005 4:18 pm

Well I will get some datalogs tonight. One of the things I thought it could have been is that one of my o2's is disconnected and the computer was reading that O2 lean all of the time so it was dumping fuel into the motor, I have 1 o2 sensor and I have the Num of HEGO's set to "1" and I have all of the Injector Output Port variables set to "0". I tried to change them all to "1" and then to "2" but the car wouldnt even run with them on 1 or 2. I have my narrowband output on my plx going to the passenger o2 harness, the driver side harness is just hanging unplugged

db1994 · Post by **db1994** » Thu Dec 15, 2005 5:32 pm

I run a real similar setup on my car. Make sure its not something simple before you go tweaking little things. 8:1 AF ratio is not a little tweak, something is way off.. Are you sure your MAF transfer is right at the low end of the sheet? Are you using the LMAF in Draw through or Blow through?

I have both my o2's unplugged with the harnesses djust hanging and the stock settings for injector output port and number of hego sensors. I simply cut the signal wires for the o2's near the eec, and wired left and right to the single wideband simulator. Works like a champ as long as you ground the wideband box to the same spot as the EEC. One of the side benefits being if you disconnect the tweecer for whatever reason, the car will still see 2 o2 sensors.

My car runs and idles nearly perfect now. 0 Driveability issues. If you want a look at the tune I am using you are welcome. It should be pretty similar to what you would run, no exotic changes to it either really.

I will post a link to it tomorrow..

-Jason

craig429 · Post by **craig429** » Thu Dec 15, 2005 7:22 pm

I had problems with my O2 sensor overheating when I had it in the down pipe about 20" from the turbo. I had to move it to about 36". I don't know how the PLX handles when the O2 sensor overheats, but the Innovative sends an error command. I have been told that when other Wideband O2 sensors overheat, the WB still outputs a reading but it may not be correct.

I thought I would post so you could keep this in the back of your mind if you get some funny readings.

I haven't done any work with the EEC running on 1 O2 sensor. I know of a guy that has one stock O2 sensor and the other O2 sensor location he has a WB set up for narrow band simulator and run that into the EEC input.

mach1 · Post by **mach1** » Fri Dec 16, 2005 1:19 am

Man this thing is making me act like a 3 yr old that just got candy taken away from me. Here is a log I made, it begins with the car cold. At frame ~1625 it goes closed loop, runs pretty good too, all the way through closed loop then the mixture hits the roof a little before 3300 and it kicks back into open loop and the car just falls on its face(Just as 2Shaker said). Man oh man the megasquirt is calling my name!

http://www.supraspeed.com/051215222808.DB

db1994 · Post by **db1994** » Fri Dec 16, 2005 9:51 am

Blow smoke out the tailpipes when it goes rich by any chance?

I fought with my car for a week trying to figure out why it was going rich, and I had an injector that was getting stuck open...

Will the car will run without the tweecer? Like just off of the MAF calibration for the injectors...

-Jason

TripleBlackVert · Post by **TripleBlackVert** » Fri Dec 16, 2005 9:53 am

how about splitting the output of the one o2 you are using so that it provides the same signal for both sides? then the computer would still think it had both o2s.

in theory, it seems like it might work ...

mach1 · Post by **mach1** » Fri Dec 16, 2005 10:13 am

i'll try splitting the o2 sig tonight, yeah it spits out smoke when it extremely rich like that. I checked my rail pressure and it hold solid which tells me its not a stuck injector, also the car will run fine until the computer goes back into open loop then it gets trashed. The car wont run w/o the Tweecer because its a LMAF and its not cal'd for 42's on a EEC-IV

JuiceSC · Post by **JuiceSC** » Fri Dec 16, 2005 11:32 am

What do you see in your datalog as far as maf aiflow/voltage when the car goes pig rich?

I had a similiar issue when I started tweecing, and when the car would flood, the maf reading in calcon was pegged

and stayed pegged till I cycled the key.

mach1 · Post by **mach1** » Fri Dec 16, 2005 12:03 pm

JuiceSC wrote:What do you see in your datalog as far as maf aiflow/voltage when the car goes pig rich?

I had a similiar issue when I started tweecing, and when the car would flood, the maf reading in calcon was pegged and stayed pegged till I cycled the key.

it stays pretty stable for me

JuiceSC · Post by **JuiceSC** » Fri Dec 16, 2005 5:21 pm

I took a look at the log you posted:

1 You need to log lambse to get a better idea of whats going on.

2 Your load% hit 200% several times, need to fix this later.

3 I never saw KAMFR change from 1, I'm thinking the maf is pretty far off.

Post a log with lambse logged when you get a chance.

mach1 · Post by **mach1** » Sun Dec 18, 2005 2:33 pm

Ok guys thanks for all of the help, Craig sent me his tune and it seemed to help. I did some datalogs with his tune and the reason it helped is because the car stayed in closed loop longer and it runs pretty good in closed loop. as soon as the car gets into open loop the A/F ratio hits the ground and the car wont run for crap (pretty much not at all) so my guess is that the MAF transfer is wayyyyy off(just as Juice has suggested), when the car goes into open loop does the EEC pretty much use the maf flow to calc the mixture? ECT and ACT sensors are working, I ran some codes and the only code the computer spit out was MAF voltage not as expected. The maf is "working" because its changing with the incoming air but I think the transfer is way off. Can you guys give me link to a good write up on how to tune the maf transfer curve? From what I have read I need to log my wideband, keep the engine in open loop, and if the car is running at 10:1 at say 2.1 MAF volts then I need to lean it out 30%, I cant beleive the transfer function would be that far off?? The maf is right next to the inlet on the turbo (draw through) the turbo has a 4" inlet and its a 4" maf, theres are 45 degree angle between the maf and the turbo inlet and on the inside of the bend there is about .5" of space between the maf and the turbo and on the outside of the bend theres around 4" of space between the maf and the turbo. Also I wired my wideband into both factory o2 inputs so both banks are reading exactly the same now.

JuiceSC · Post by **JuiceSC** » Sun Dec 18, 2005 4:13 pm

"Can you guys give me link to a good write up on how to tune the maf transfer curve?"

I like to use lambse to help dial in the maf. Lambse with no KAMRF build up is a pretty good indicator of whats going on. It should be between 12~15 depending on driving conditions. If you see lambse over 15 mostly, your tune is RICH and its LEAN if your under 12. But this is just a general guideline. As KAMRFs build up, lambse (commanded afr) stabilizes. Ideally, you should have lambse equal the wideband reading or very close to it. By looking at lambse, I don't have to wait for KAMRFs to build to get the maf close.

Ps: I don't like the fact your getting a maf code, but it maybe just because your transfer is off.

95 Cobra + Turbo ... Oh Noes!

Who is online