First logs..Not sure if they are good or bad

[Ridgeracer]

IF by resent, you mean just that...resent. Then this does nothing for resetting the stored learned values. You have to reset or clear the RAM for learning to be back at 0.

 

Disconnect battery and reflash both base and realtime? If so i'll do it later today.

[Infamous1]

Disconnect battery and reflash both base and realtime? If so i'll do it later today.

Disconnect the battery and you will only need to reload the realtime map. Or you could just reset the ECU via the RomRaider logger and reload the realtime map.

[edmundu]

^^What he said......Or if you are reloading via the AP, you can use that to reset the ECU, or if via ST, it can do it as well.

 

Pulling the battery also loses your radio presets, along with the clock.

[SeeeeeYa]

Not being the most politically correct person here I'm going stick in my $0.02.

 

One, I would NEVER accept 3% LTFT, plus or minus.... but especially MINUS!!, from my own amateur tuning on my car. I value my engine too much, not to mention performance.

 

Two, all the mumbo jumbo about fudging this and that, and altering things in non-standard ways that no one agrees with or really understands is pure hackery... IMO.

 

Three, I've seen some of this coming, and commented about it in some forgotten thread in the past. I also got the same response as the far more learned members have in this thread. The member who was the object of my referenced comments has since vanished from the forums.... and I would never wish to dis a valued vendor, but I will quote a post that gives a clue;

 

I think he is TDC Mike.

 

[LittleBlueGT]

Not being the most politically correct person here I'm going stick in my $0.02.

 

One, I would NEVER accept 3% LTFT, plus or minus.... but especially MINUS!!, from my own amateur tuning on my car. I value my engine too much, not to mention performance.

 

Two, all the mumbo jumbo about fudging this and that, and altering things in non-standard ways that no one agrees with or really understands is pure hackery... IMO.

 

Three, I've seen some of this coming, and commented about it in some forgotten thread in the past. I also got the same response as the far more learned members have in this thread. The member who was the object of my referenced comments has since vanished from the forums.... and I would never wish to dis a valued vendor, but I will quote a post that gives a clue;

 

While I agree with what you said, I don't think 3% is all that crazy. While I strongly disagree with skewing the injector scaler and MAF scaler, and everything that goes along with it.

 

I highly doubt that most tuners on a single tune (like "professional tuners") get better then that.

 

My trims (except for A) are constantly within less then 1%. I check them every time I use ST, which is about 3 times a week. I have been doing this for about a year.

 

FWIW when I drove to BC and Oregon all my trims changed, of course they all went in the same direction, and all by about the same amount, I believe it was about 3% off for B,C & D. So the trims were doing what they were suppose to do, change with the fuel, thus changing things a bit at WOT, but one would theorize that that would be correct given the different gas.

[Legacy_GT_Pilot]

How do you determine the correct injector latency? I have PE850s, and have tried both the PE suggested values, and the values suggested by Cobb for these injectors and my fuel trims are still a little more erratic than what I'd like.

 

And how do you log latency? I see voltage as an option in RomRaider Data Logger, but not latency?

 

Dave

 

Assuming you don't have a massive boost leak and that the logs were done at Patm = 14.7, then your MAF scale is approximately 27% off. Multiply your MAF and load by 1.27 to get a more accurate MAF and load value. I don't think you have a boost leak because your car is accelerating correctly based upon your "adjusted" MAF.

 

In order to make your fueling work, it looks like the injectors may have been scaled to 650 cc/min.

 

After discussing modified injectors with some of the STi guys, I think I know where the problem is. Modified injectors DO NOT have stock latencies even though there is a HUGE myth that they do have stock latencies. The only way to make the injectors idle properly with stock latencies is to either change the injector scalar or screw around with the MAF. I know that some of the STi guys scale their modified injectors at 650 cc/min and use stock latencies.

 

The problem with using MAF or scalar to adjust for incorrect latency is that it won't work at low load and high rpm.

 

In the low load and low rpm range, the ecu is in closed loop so any errors in fueling will be adjusted by the ecu.

 

Above 4000 rpm, the ecu is almost always in open loop so your fueling must be correct.

 

At 4000-4800 rpm and load between 0.6 and 0.85 g/rev, a latency that is too low will cause the fueling to run too lean. I think that's why you are getting -7 FLKC in this range.

 

Try logging battery voltage and latency.

[mickeyd2005]

How do you determine the correct injector latency? I have PE850s, and have tried both the PE suggested values, and the values suggested by Cobb for these injectors and my fuel trims are still a little more erratic than what I'd like.

 

And how do you log latency? I see voltage as an option in RomRaider Data Logger, but not latency?

 

Dave

 

You need to get the latest defs. However, it won't help you if you are self tuning. All that will tell you is the value that the tuner typed in except it will be interpolated for the actual battery voltage. If you're the tuner... then it's no use.

 

Do NOT use the Cobb values for latency. I didn't realize people were using this spreadsheet that Cobb published. I know for a fact that the 816 cc modified injectors do not use stock latencies. I have tuned them myself and Justin 05 STi on romraider also proved that they do not have stock latency.

 

I helped setup bugblatterbeast's method for calculating latency in romraider. However, it's turned out to be too difficult to use for most people primarily because they don't bother to tune the MAF PRIOR to tuning injectors.

 

The easiest way to set up latency and scalar is to log fuel trims at high and low low load using your STOCK injectors. Then install new injectors and then do the same log and compare fuel trims. It's VERY easy.

 

However, I have found that you can not use idle fuel trims for the low load due to the erratic nature at idle. Log trims at walking speed in a parking lot in 1st gear.

 

A friend is going to install PE850s soon. He just sent me logs of his stock injectors. We'll find out what the latency is.

 

Keep in mind that the scalar for all injectors vary a little bit. For example, DW650 injectors can actually be anywhere from DW640 to DW660. DW bins their injectors and sends out matched sets.

[Infamous1]

Any success with the RomRaider Latency and scaler tool? I will be tuning a 20g setup with all the bells and whistles here shortly. I plan on going with the stock latencies and the flow chart numbers for scaler (Modified stockers), then smooth it out with the tool.

[Legacy_GT_Pilot]

If you don't mind, please send me the latency values and the scaler you figure out with your friend. Hopefully it's close to where I am currently. If not, I will try it and log to see if it helps at all.

 

Dave

 

A friend is going to install PE850s soon. He just sent me logs of his stock injectors. We'll find out what the latency is.

 

Keep in mind that the scalar for all injectors vary a little bit. For example, DW650 injectors can actually be anywhere from DW640 to DW660. DW bins their injectors and sends out matched sets.

[Legacy_GT_Pilot]

The Cobb spreadsheet provides the "advertised" values and their "tested" values for PE850's. I've tried both, and just checked my current tune and I have them set at the tested values.

 

6.00 9.00 11.50 14.00 16.00

Advertised 7.46 2.49 1.17 0.85 0.55

Tested 7.66 2.69 1.47 1.15 0.85

 

Another member who has the same injectors PM'd me his latency settings, and he used the tested values as well.

 

Dave

[SeeeeeYa]

See, isn't it nice to have a Tuning Forum where all this good tuning information is available all the time to everyone.......

[Infamous1]

Haha take a bad situtation and turn it good! Tuning forum?

[Tuning Alliance]

Regardless of how anything is scaled, when any car learns negative fuel trim its not good at all.

 

The high load column is around 40 g/s, so in that area (30-40 g/s), even a perfectly scaled maf can work against you. It actually makes sense to calibrate the maf slightly leaner 1-2% in that area, so if the car learns anything it will learn to add fuel.

 

Using 650-700cc scaling and near stock latency does not create any lean spots that I have seen. PE injectors will require a much higher latency value! The issue typically is at idle where the car is most sensitive to injector dead time. These cars usually only see low load 4800rpm when the engine is decelerating in over-run.

 

Learning will always work against you! Reset the Ecu with the Accessport, and resend map.

[mickeyd2005]

Using 650-700cc scaling and near stock latency does not create any lean spots that I have seen. PE injectors will require a much higher latency value! The issue typically is at idle where the car is most sensitive to injector dead time. These cars usually only see low load 4800rpm when the engine is decelerating in over-run.

 

This is not true. You will get a lean spot when the ecu transitions from CL to OL in the low load area at 4000 rpm. That's where OP got the -7 FLKC.

 

The rpm and load where the knock occurred is 4000 to 4800 rpm and somewhere between 0.60 to 0.85 g/rev. That's not over-run. I drive in that region all the time.

 

Using a too low value for the injector scalar also causes the load calculation to be off which causes the ecu to use a higher timing than it should in the stock region.

[mickeyd2005]

The Cobb spreadsheet provides the "advertised" values and their "tested" values for PE850's. I've tried both, and just checked my current tune and I have them set at the tested values.

 

6.00 9.00 11.50 14.00 16.00

Advertised 7.46 2.49 1.17 0.85 0.55

Tested 7.66 2.69 1.47 1.15 0.85

 

Another member who has the same injectors PM'd me his latency settings, and he used the tested values as well.

 

Dave

 

I haven't looked at the Cobb spreadsheet enough to know where these numbers come from.

 

However, keep in mind that romraider displays latency differently for 32 bit ecus. It was changed several months ago. merchgod discovered an error in the conversion between ecu value -> physical value for the 32 bit ecus.

 

I know that when I used bugblatterbeast's latency calculation method that I could not match the old romraider defs. After merchgod changed it, I was able to make it work.

 

We are also able to confirm the new defs are correct because the new 32 bit ecus use top feed injectors. So, now we can compare the old 16 bit latencies to the 32 bit latencies.

 

Anyway, just something to be aware of when comparing values in a map and values you might have obtained from a test.

[Ridgeracer]

Finally got new logs and actually reset and reloaded the maps on the ECU. I think they look alot better now. Could it have been just bad gas??

 

http://www.box.net/shared/lyjcn7tkc0

[Infamous1]

No, you still are knocking. How many miles after the reset? You already have some learned knock. Oh FWIW the learning view is more useful after several days of driving.

[mickeyd2005]

ridgeracer, have you read my posts?

 

Your knock is primarily in the partial throttle range. Taking learning logs every 1.5 minutes is not going to tell you anything.

 

You need to drive in all ranges. I would drive primarily in the low load/high rpm range.

 

Try doing a log in 1st gear from start to redline at low throttle.

 

Besides that, you knocked immediately in your log even with zero learning.

[Ridgeracer]

Still haven't had the chance to do that higher maf log yet. I'll do some part throttle high rpm logs tomorrow. Still trying to learn how to read/understand what I'm logging. Should I give it a couple of drive cycles before I take a learning log?

[Ridgeracer]

There is no doubt that OP will eventually make more power when his boost control issues are resolved.

 

However, his current problem is the partial throttle knocking and that has nothing to do with the wastegate. The WOT logs actually don't have any knocking in them. The FLKC is driving down the timing. There's no way to know where the FLKC is causing the knock without resetting the ecu and doing partial throttle logs.

 

The large AFR learning values indicates a poorly scaled MAF. That combined with the calculations of MAF based upon MAP indicates that his MAF and fueling issues need to be resolved first. I normally allow AFR learning C to vary by +-3 and AFR Learning D +-2.

 

It's common to have some negative FLKC. However, -7 and -2.8 degrees is too much and FLKC in the high load range is not good. It causes driveability issues.

 

Sorry man. I missed this post

[SeeeeeYa]

In short, and after all this it should be obvious, go to someone who knows how to tune and replace the POS "tune" you now have before you damage something.

[mickeyd2005]

This thread has been bumped a lot so it may seem like there are big problems with OPs tune; however, it really is typical. Lots of people email or PM me professionally tuned logs. I would rate this one as a typical slightly below average tune.

 

There's light knocking up top (even with zero AFR learning) and there's knocking (some severe) in the partial throttle range.

 

Believe it or not... that's typical of a protune. Most protuners don't check the ENTIRE map. There's just not enough time on the dyno. That's why I advocate that everyone learn to log and read logs.

 

I still disagree about using the fake injector scaling. It's clear now that this tune does use stock latency and a scalar between 650 and 700. That's what the tuner implied in his previous post. Using fake injector scaling causes problems in the partial throttle area UNLESS it is carefully dialed out of all the load based maps. It doesn't look like it's been dialed out.

[Tuning Alliance]

Pulling 1.5 degrees up top, probably holding boost better with the new wastegate arm. It needs some minor tweaks, you guys are a bit over kill here.

 

I will test the low load latency you speak of on the dyno the next chance I get. I don't get anyone reporting a hesitation there. I can't edit that map anyway!

[mickeyd2005]

Pulling 1.5 degrees up top, probably holding boost better with the new wastegate arm. It needs some minor tweaks, you guys are a bit over kill here.

 

I will test the low load latency you speak of on the dyno the next chance I get. I don't get anyone reporting a hesitation there. I can't edit that map anyway!

 

What do you mean? I can edit it.

[mickeyd2005]

Pulling 1.5 degrees up top, probably holding boost better with the new wastegate arm. It needs some minor tweaks, you guys are a bit over kill here.

 

I will test the low load latency you speak of on the dyno the next chance I get. I don't get anyone reporting a hesitation there. I can't edit that map anyway!

 

You don't need a dyno. The front O2 sensor can be utilized when load < 1 g/s.

 

Just have OP datalog from idle to redline in 1st gear at very low throttle.

 

If you search on romraider for Justin 05 STi, you will see what happened when he tried to use stock latency for his modified sidefeed injectors. I warned him that stock latency wasn't going to work.

 

You can see from the following graph that IPW and latency has NOTHING to do with RPM or IDLE. It only has to do with LOAD. So, fixing latency at idle using MAF will not correct it at high rpm and low load.

 

http://i141.photobucket.com/albums/r69/mickeyd2005/untitled-18.jpg

 

BTW, I agree that this tune is not atypical. I've seen other tuners do the same thing with the same results. However, if you keep bumping this thread with incorrect information then I am going to keep responding.