scheides

iTrader: (10)

Acknowledged, I'll try it as well. Haven't fully reviewed logs from initial change yet so I'll do a side-by-side compare.

03whitegsr

iTrader: (8)

Initial drive felt good, no data logging though. I've been getting some roughness when the motor was cold below about 2500 that I had ascoiated with just warm-up issues as the car has never droven well when cold. Seemed to be gone this morning with this change implemented. Warm up still isn't perfect, but this seems to have had a marked improvment.

Can't comment on any other conditions though as I hadn't noticed the issue anywhere else.

If what you guys experienced was the same, only at all times, I can understand why people were switching back to MAF. It was horrible for driveability, but only happened for the first couple minutes on cold starts for me so I hadn't worried too much about it.

Accel enrichment seems to be back to normal. I was getting random results with the first swap where some tip-in events would go lean and others rich. Now it dips slightly rich in the few situations I intentionally tested it out.

Evo_Jay

iTrader: (17)

Why swap 05s to 03-04 computer?

9417 will work on a 05 computer.

scheides

iTrader: (10)

^Because of this jitter problem. Those ecu swapping days are behind us now I think

That's just crazy JB.

Evo_Jay

iTrader: (17)

I understand he swapped ECU so he could run 9417 so he could eliminate the jitter problem.

What Im saying is he could have used an 05 ECU and loaded 9417 on it.

mrfred

iTrader: (50)

Doing all this has made me take a closer look at how sync accel works. Turns out that there are two other tables vs master load that play a role in sync accel enrichment. Their roles are somewhat indirect in that they represent threshold values that DeltaMasterLoad must exceed before it can be used to scale the sync accel contribution. This could be useful for tuning SD. I'll add further info to my fuel pulse width post in a few days.

l2r99gst

iTrader: (2)

Awesome updates, mrfred. I think people thought I was nuts when I first mentioned this issue back in June or so. But, since I had the logs to back it up, at least there was some proof that there was an issue.

When I initially tested the first fix, I was so happy to see the jitter gone and the car feel smooth on SD again. Now, we can all reap the benefits of an SD setup. I, for one, have been waiting to install my 4" Dejon intake and Amsoil filter, but just couldn't do it with he IPW issue.

So, a big thanks to you for finding the solution to the problem and 03whitegsr for starting a new thread to renew the effort. Everyone's help is truly appreciated, as always.


Eric

clockworktoy

Tried this out this morning and the car drove smoothly without any of the jitters. Car still drives horribly when cold but once warm and cruising everything seems great now.

JohnBradley

iTrader: (30)

I didnt read enough to know that it would work on the 05 as well. I have ALOT of 03-04 stuff here because of AEM being the only game for speed density for so long. Its an easy swap for me to get it to work. I save the 05 stuff for 2G Eclipse swaps anyway since I need 9653 on those. I suppose that also works on 03-04 though? I blink on the wrong day and get behind the times

aaron

Slo_crx1

iTrader: (9)

I tried out the first fix this evening before I read the update, and even then it's a night and day difference. I really didn't notice too much leaning out, and with my luck how it is somehow my Evoscan reset itself and I forgot where I put the activation code, so no log for me tonight.

Also the first fix seemed to have fixed my cold idle and driving issues, it ran smooth as silk as soon as I started it up. Tomorrow I'll grab a few logs on my drive to work to see if I pick up what mrfred mentioned.

03whitegsr

iTrader: (8)

I noticed some rough post start up behavior on one of 3 cold/warm starts today. Every start, the car had been sitting 4+ hours in like 30 degree or cooler weather. I would imagine it was pretty close to fully cold on each start.

It's reduced the issue, but it faintly remains. Could be accel related though as it's only rough on light acceleration below 2000 RPM. AFR goes lean.

Thanks for the exceptional work Mrfred.

scheides

iTrader: (10)

Ok, so the initial change seemed to make a vast improvement in cruising smoothness. I didn't get many logs from it but I'm considering switching back and getting some, and here's why: the second change (reverting the initial one of course) resulted in a nearly equally smooth car, but while the engine was cold, I felt that the cruise IPW dropout issue had returned. When fully warmed/hot, maybe its just me but I could swear that it was still there, although drastically reduced. Does this make any sense? Basic WOT fueling checks out with both, and fuel trims were unaffected with the initial change. I'll check fuel trims on the second one and report back, along with some more logs.

Is it possible there is a true fix combining these two values somehow? Maybe the 205 gets bumped to 230 and the 112 drops to 60?

mrfred

iTrader: (50)

You are welcome to use either fix if you are not having any driveability issues. Here's a rundown of the situation:

For uncompensated loads of < 70, the FPW calculation subroutine wants to use the following formula to scale the base fuel pulse contribution to the total fuel pulse:

BFPW scalar = (MasterLoadW +/- C*DeltaMasterLoadW)/2048

where "+" is used when the load is increasing the "-" is used when the load is decreasing. The "W" means the values are scaled by the fuel pulse width warmup up compensation table, aka MAFMULTWARMUP.

When uncompensated load is >= 70, then the code wants to use this formula for the base fuel pulse scalar:

BFPW scalar = MasterLoadW/2048

For reasons that aren't clear to me yet but are surely in the details of how the master load is calculated in low load conditions, a car in MAF control seems to do ok with this scenario, but a car with the SD patch does not like the DeltaMasterLoadW contribution.

Here is what the first fix does:

1) It forces the code to use MasterLoadW/2048 at all times for the base fuel pulse width scalar.

2) It changes the method for calculating the min load change needed to induce a SyncLoadAccel contribution to the total fuel pulse width. Compared to the factory setup, at low loads, the min load change required is increased, and at high loads, the min load change required is decreased. Thus the car may have a little more hesitation during load change at low loads (below 160 load), and may be a little richer during load change at high loads (above 160 load).

The second fix just forces the use of MasterLoadW/2048 at all times.

Without knowing how much SyncLoadAccel contributes to the total fuel pulse width, the second fix seems the safer route to go. However, if you are happy with AFRs during load transitions when using the first fix, then its ok to use.

What ultimately may be the best solution could be the use of the second fix in combination with access to the tables that control the min load change required to induce a SyncLoadAccel enrichment contribution. These min load change values for SyncLoadAccel contributions may be causing the jittery response that you feel with the second fix. I plan to post those tables in my advanced fuel control options thread later this week.

03whitegsr

iTrader: (8)

Pure genius, thanks for everything you do.

scheides

iTrader: (10)

Funny you should mention the hesitation at low-load for the initial change, I thought the car felt a bit lazy but it was also so smooth I was in temporary heaven! I hadn't logged WOT much but it does all add up, thank you for sharing the calculations and what each one does.

Right now I think the best bang/buck is def the second fix, the jitter is isolated to only right around 2500rpm (almost exactly) and it can only be felt at highway speeds...who cruises at 2500rpm in top gear anyways? Lame! On top of that, it is SO small it is hard to tell unless you felt the previous jitter. Too bad these variables couldn't be set dynamically somehow based on load or TPS or something (hint hint tephra?).

I'll watch for your other updates but for now I think I'll stick with fix #2. I'm in the middle of an all-nighter for work and drove on fix#2 down to our datacenter and flashed fix#1 on before heading in. I'll try it for the drive home and see if I can try and articulate the differences.