Need input on logs.

[kiddsm95]

Previously my E316G was spooling slow, and it turned out to be lean & rich conditions. After a crap tune on the SAFC2 my car is running better, but not quite there yet. I'm running the factory narrowband O2 sensor and using OBD Gauge with a palm m130 for logging. Pulls were done in 2nd gear because I don't want my car impounded for doing WOT 3rd gear pulls.

Any feedback will be appreciated.

 

[linksys42]

First thing you need to do when tuning is bring the rpm to around 2500 rpm while cruising along then go to WOT (accelerator all the way to the floor the whole time), TPS at 100% from 2500 rpm's all the way to 6500 rpm's or redline. And I know you say you don't want to do the pull in any gear higher than 2nd but your really not going to tune the car very well in 2nd compared to 3rd or 4th. Because 2nd your not going to see as much of a load on the engine as in 3rd or 4th. Also all you need to log is RPM, timing, and front 02 or TPS, since you are trying to tune off of the stock narrowband 02. Logging fewer items will give you a lot more samples per second, which will give you more info to tune by. Go back out and go by what I have said and post another log then I will be able to give ya some advice.

 

[kiddsm95]

I'll go make another pull per the suggestions above. As for the front O2 sensor, OBD Gauge only specifies Ox sensor 1 & 2. I'll assume that the one I need to use is Ox sensor 2 as this give logical numbers as opposed to Ox sensor 1.

Will post back soon.

 

[linksys42]

Here is an example. Don't mind the knock value you wont be able to log it with the logger you are using. Also typically it will be o2 1, o2 2 is usually rear o2 sensor you should have a value on the front o2 sensor of any where between (hunting back and forth for the perfect mixture of 14.7:1 air to fuel) .20 to .80 while at idle or cruising/partial throttle. And in the .90's constant at WOT. Also be careful and watch the timing and 02 sensor to watch for knock and a lean mixture.

 

[kenamond]

First thing is to figure out what your TPS percentage is by flooring as far as it will go and see what the logger says. You don't have to be recording it. It should be in the high 90% range. Same goes for coolant temp and intake air temp, as both of those will affect timing, but won't change much during a pull. You can log them during a pull just to see if they change much, but I only see a degree or two change on each.

Figure out which O2 value is front O2v. That's all you care about when tuning. As was mentioned, it varies between 0 and 1. The ECU only uses it at part throttle, and it should oscillate around 0.5v which is 14.7 AFR (stoich). Under load, it just goes by the fuel map and fuel trims.

The LTFT will give you feedback of how the AFR is doing, but only at part throttle. The ECU knows if it's getting too much or too little fuel than it expected, and compensates via the trims. If your LTFT is off a lot, you're too rich or lean...but this is only at part throttle and on average, so your AFC might be lean in one place but rich in another and it won't tell you that very clearly.

You might try brake boosting in 2nd to slow down the acceleration and get more samples without getting a ticket. It's a bit hard on the brakes though.

Without the ability to log knock, it's kind of tough to tune on a 2g IMO. There are ways of calculating it if you log RPM, timing advance, and airflow, but you have to have the stock timing map and do a bit of math. The Excel spreadsheet of the fuel and timing maps for the 2g are available if you search for them. So is the approach for identifying knock.

The O2v can tell you if you're leaning out, but it's very crude, too, because it's narrowband. If it drops into the 0.8v range, you're lean, but you can also be lean in the low 0.9v range, too. When I replaced my O2 sensor, my O2v went from 0.95v to 0.98v for otherwise identical pulls, so you can't guage too much from those readings. Just try to keep them around 0.98v or so, and that's probably the best you can do without a wideband.

I was screwing around the other day and wrote a program that will plot your timing curve on top of the timing map's "ideal" curve and show the difference in timing advance. The next step is to use the ECU's same logic to back-calculate knock count from the timing offset. It's not something I can pass around, but maybe some day I can put a javascript tool up somewhere that 2g folks could use...just plop your log data (with rpm, timing advance, and airflow) into a window and have it spit out the ideal and logged curves on the same plot...or generate a list of numbers that you can paste into whatever plotting tool you own. Anyway, I can see if I'm in the ballpark with timing or not, and that tells me if I'm knocking or not.

 

[linksys42]

Here is a link that will give you an idea to get you going in the right direction.

http://www.dsmtuners.com/forums/articles-tuning-ecu/58403-definitive-piggyback-tuning-guide.html

 

[Calan]

For one thing, unless your TPS is hosed, you were in and out of the throttle which was causing the car to switch between open and closed loop. And based on how far out your trims look in that log, it would be likely that your car is trying to switch from a rich condition (closed loop) to lean (open loop, depending on your WOT tune) back and forth.

First things first: Get your trims in line, and then do a steady pull at WOT.

 

[kiddsm95]

Ok, I made two runs today. Apparently my TPS is only opening up to 87 WTF, so I'll have to chase down the problem with it. Can anyone comment on the logs with the TPS as is, or is it only possible to speculate what's going on at 100% WOT.

Your thoughts please?

 

[kiddsm95]

Found the throttle problems, so I did a couple more runs today in 3rd gear at WOT. I'm having some hesitation issues around the 1k-3k range, but right at 3k the turbo kicks in hard all the way to redline. Please comment on what kind of adjustments I need to make throughout the powerband. If at all possible please recommend a particular amount of adjustment (ie. +3% at 1k, +1% at 2k, etc...) as I've only got about 4 hours of tuning experience.

I would also like to know what your thoughts are about the last part of the log. I had a CEL come on here and the car started running pretty rough, when I checked the DTC it threw an O2 bank 1 sensor failure. I'm assuming that the ECU tried to compensate the fuel trims and couldn't therefore throwing the CEL?

Your comments are welcomed.
Thanks.

 

[kenamond]

First of all, I don't know if it's just my browser (IE7) or what, but when you post your logs in that format, it makes the windows super-wide so that I have to use the left-right scroll bar. Maybe it's a forum glitch, but if you could put them in as regular text, that might help (and edit all your other posts on this thread to change the log format as well).

Second of all, I have no clue what your O2v numbers mean. I'm used to ranges of 0.0v to 1.0v. So logs usually look like 0.98 instead of 520 or whatever. Anyone else know what these numbers mean? Without knowing what your O2v mean, I have no clue about your lean/rich status.

Finally, floor it starting at 2500rpm and take it full throttle to redline. Your logs don't hit full throttle until late.

 

[linksys42]

Yes it is a forum problem I have posted logs before and it had done the exact same thing. I used the exact same instructions as here in this thread in the tech index http://www.dsmtuners.com/forums/art...st-logs-so-people-can-actually-read-them.html .

I am just guessing here but I bet the number for the o2 just needs the decimal in front and the rear 0 taken off then you have the voltage. Just guessing though, you really need to find out what the value for the 02 is.

As it was said above when doing a pull to tune Hi throttle you should get it in 3rd bring it up or down to 2500 then put it to the floor and don't let up till 6500 or red line unless you see a lean condition or see you timing being pulled hard from knock.

First I would get your STFT's and LTFT's in check as close to 100 as you can while at idle or partial throttle cruising. The best way is to go out on the interstate and take a long cruise adjusting my trims at my set ne points on the AFC. Go completely through this link http://www.dsmtuners.com/forums/articles-tuning-ecu/58403-definitive-piggyback-tuning-guide.html . And post any of the questions you have on the guide.

Also your timing is all over the place. With that 16g you should be seeing full spool by at least 3400 even if you are not going to WOT till 2900 rpms. The timing should steadily climb from the point when you hit full boost. Your timing is being pulled through out the whole rpm band (because of knock). What are you Lo and Hi fuel trim setting all the way across on your AFC for the evo 560cc injectors you have?

2/3/2008 16:07 2964 18 99 360 920
2/3/2008 16:07 3328 16 99 340 940
2/3/2008 16:07 3724 16 99 300 940
2/3/2008 16:07 4104 16 99 260 960
2/3/2008 16:07 4480 15 99 220 980
2/3/2008 16:07 4864 14 99 200 980
2/3/2008 16:07 5264 15 99 180 1000
2/3/2008 16:07 5680 13 99 160 1000
2/3/2008 16:07 6064 14 98 140 1000



Here is an example how timing reacts to knock on a rough tuned 3rd gear pull.

 

[kiddsm95]

I'm set at -10% from 1k to 7k, for Lo & Hi settings. According to the guide, it should be set at -19% for the initial tune, but when I dial in -15% on the SAFC2 the exhaust starts popping indicating lean conditions.

I have gone through the DPTG many, many times and as you can tell I'm still having tuning issues. I'll try to re-tune the Hi settings at each Ne point again to see if that helps.

As for the O2 voltages, OBD Gauge units are in mV so the decimal should be shifted 10X-3. However, I can't figure out why it's reading on sensor 2 instead of sensor 1??? Anyone using OBD Gauge have this problem?

Once I retune the Hi settings I'll do another 3rd gear pull and post the results in a few days.

Linksys42 ~ What logger software are you using, and how'd you log knock on a 2g? Also, how'd you get so many sample rates on the log you posted?

Thanks again for everyone's help.

 

[kenamond]

If O2v is in mV and it's O2 #2 in your logger, you need to add fuel wherever it's below 980 and remove fuel wherever it's at 1000. But you have to log from 2500rpm to redline at *full throttle*. I actually start my logs at 2000rpm. If you're at 800, you need to add less fuel than if you're at 500, but the end result should be some O2 values in the high 900mV range.

So go do a pull from 2500-7k rpm logging rpm, timing advance, O2 #2, and airflow while the gas pedal is completely floored. Post that log. I can give you an idea of knock if you log rpm, timing, and airflow, but if you leave out any of those three, it's impossible.

 

[linksys42]

Was wondering what kind of boost you are seeing? And as kenamond was saying all you need to log are timing, 02 (just the front one which is 02 #2 on your logger), and airflow. Just by logging just those three items your sample rates will be higher (less you log the more info you get). But anyhow by your previous logs you don’t seem to be leaning out as long as o2 # 2 is your front o2 sensor. But once we get a good pull from 2500 till 6500 or red line we should have a better idea. The logger I use is called Evoscan. Its has a baud rate of 15625 that’s why I get so many samples per sec, it also uses mut III protocols to log, I can log IDC, knock, airflow/rev, air volume, external wideband A/F ratio, injector latency, ecu load, octane level that the ecu uses and it also has a hp/tq calculator to see if I am making any progress (helping or hurting) my tune. Here is a link to the logger EvoScan OBDII Mitsubishi MUTII DataLogger Scantool . And also has map tracker and an overlay setting like kenamond was talking about below. Very nice logging software below is a screen shot.

I was screwing around the other day and wrote a program that will plot your timing curve on top of the timing map's "ideal" curve and show the difference in timing advance. The next step is to use the ECU's same logic to back-calculate knock count from the timing offset. It's not something I can pass around, but maybe some day I can put a javascript tool up somewhere that 2g folks could use...just plop your log data (with rpm, timing advance, and airflow) into a window and have it spit out the ideal and logged curves on the same plot...or generate a list of numbers that you can paste into whatever plotting tool you own. Anyway, I can see if I'm in the ballpark with timing or not, and that tells me if I'm knocking or not.

 

[kiddsm95]

Yeah I was reading up on Evoscan the other day and came across your post above. Did they ever release a copy for the 95-96 ECU's? We've got snow and ice here so my next log will be a few days from now after it melts off. I was messing with the car lastnight and discovered a boost leak that was probably affecting my last log, so we should label it as an outlier.

So referring back to my initial post, my fuel trims are all out of whack. It looks like I'll ned to reset my LTFT's and try to get my trims as close to 0 as possbile. Looking at the DPTG can someone explain to me how the addition trick works for 2g's?

"2g: 2g's only have 2 fuel trims, a long term fuel trim (LTFT) and a short term fuel trim (STFT). The STFT varies with the O2 sensor, an the LTFT goes for every rpm range. Since the STFT directly effects the LTFT, then you can just add the two together, and tune from there. For example, if the LTFT is +20%, and the STFT is -5%, you are at approximately +15%. "

Thanks.

 

[linksys42]

Yeah I was reading up on Evoscan the other day and came across your post above. Did they ever release a copy for the 95-96 ECU's? We've got snow and ice here so my next log will be a few days from now after it melts off. I was messing with the car lastnight and discovered a boost leak that was probably affecting my last log, so we should label it as an outlier.

So referring back to my initial post, my fuel trims are all out of whack. It looks like I'll ned to reset my LTFT's and try to get my trims as close to 0 as possbile. Looking at the DPTG can someone explain to me how the addition trick works for 2g's?

"2g: 2g's only have 2 fuel trims, a long term fuel trim (LTFT) and a short term fuel trim (STFT). The STFT varies with the O2 sensor, an the LTFT goes for every rpm range. Since the STFT directly effects the LTFT, then you can just add the two together, and tune from there. For example, if the LTFT is +20%, and the STFT is -5%, you are at approximately +15%. "

Thanks.

Evoscan will work on all 2g's if they have the OBDII diagnostic port. It has always supported all 2g's the reason why people cant get it to work is because they are trying to connect at too high of a baudrate. Only 98 and 99 ecu's that are Mitsu H8/500 ecu's can log data at the baudrate 15625 that evo 8 and 9's log data at.
On 95-97 they have to adjust the baudrate too 1953. That's why they cant get it to work, well that and they prob didn't install the correct ftdi cdm driver and tatrix driver which come with ecu flash all they need to do is download and install ecu flash that comes with the drivers and make sure they have .net framework 2.0 and it will work.
It says that on Hamish's website the link I posted. Below I posted a pic on where to change the baudrate so if some one does a search they will know what to do. But any way back to your fuel trims before you go making any WOT pull's you need to get the LTFT as close to 0 as possible go out on the highway and set your cruise control at one
of you ne points on your afc lets say 2500 and just cruise for awhile watching your LTFT if you LTFT is at lets say 7 that means the ecu is compensating for a lean condition it means that your STFT has been at 7 for an extended length of time so the LTFT has set 7 at that rpm of 2500 until it will see a change in the STFT. In short the STFT directly
affects the LTFT over a period of time. So what we need to do is at the ne point of 2500 on our afc we need to add some fuel typically +1% more fuel on the afc will bring the LTFT down from 7 to 4 over a period of time, but you will have to experiment with that. And it might take a bit for the ecu to change the LTFT after you have changed the
setting on the afc but the STFT will show the change instantly and therefore change the LTFT over time. I hope that explained it clearly for you. But I am going to say this again you need to get your LTFT as close to 0 as you can at you AFC ne points what ever they may be (IE 1000, 2000, 2500, 3000) before you make any more WOT pulls.
Also once you get your 3000 or 3500 rpm ne point as close to 0 as possible you can just take that setting for 3000 or 3500 on you afc and copy it to 4000, 4500, 5000, 5500, 6000, 7000. then take your Lo throttle setting and transfer it to your HI throttle settings. Then you can fine tune your Hi throttle settings. Also you didn't say what boost pressure are you running?

 

[kiddsm95]

Ok your explaination of the fuel trims makes perfect sense. I'll get them adjusted as close to zero as possible. I'm running a dual stage MBC so 1st stage is set to 15 psi which has been the boost press I've been using for the logs above, and the 2nd stage will be set to 20 psi. This brings up another question, E316G with supporting mods on a stock internal 7 bolt @ 20 psi should be fine right?

As for EvoScan, does it require a special cable to attach between the T16-003 Multiplex translator and your laptop, or is that the $79 cable they refer to on their webpage?

Wish I would have know about EvoScan before going the palm route...

 

[linksys42]

Ok your explaination of the fuel trims makes perfect sense. I'll get them adjusted as close to zero as possible. I'm running a dual stage MBC so 1st stage is set to 15 psi which has been the boost press I've been using for the logs above, and the 2nd stage will be set to 20 psi. This brings up another question, E316G with supporting mods on a stock internal 7 bolt @ 20 psi should be fine right?

As for EvoScan, does it require a special cable to attach between the T16-003 Multiplex translator and your laptop, or is that the $79 cable they refer to on their webpage?

Wish I would have know about EvoScan before going the palm route...

Instead of just turning it up to 20 psi from 15. You should get a good tune at 15 psi then slowly bump the boost up in 1-2psi increments along with adding fuel in the afc where
needed in the rpm band. And that way you will be able to find safely what psi you can run. The thing that's going to hold you back is the stock side mount intercooler heat soaking.
But you can tackle that when the time comes either with a front mount or water/meth injection. But yes the stock internal 7 bolt will have no problem with a good tune and a well maintained engine at 20psi.
As for the evoscan usb cable I didn't want to mess with trying to convert the pocket logger serial cable to a high-speed usb so I just bought the tactrix cable, less headaches.

 

[kiddsm95]

Ok so in my last post I mentioned that I discovered a boost leak, which I've recently fixed. I also mentioned in my previous post that I was running 15 psi
in stage 1, which is incorrect as well. My 1st & 2nd stage were completely bottomed out, so I'm really only running about 10 psi. I completely
forgot that I had the boost turned down due to my slow spool condition I was experiencing before I did my fuel upgrades, REF: http://www.dsmtuners.com/forums/newbie-forum/278366-evoiii-16g-slow-spool.html

Could the boost leak have been the culprit as to why my car pulled timing as shown in my previous logs?

I have also been monitoring my LTFT's & STFT's for a few days now and they are 0% & 5% respectively when the car is on cruise control at the Ne points.
So I think I've got these under control.

Tonight I'll be turning the boost up to 15 psi for stage 1 and I'll try to post the new logs by tomorrow morning at the latest.

Thanks.

 

[kiddsm95]

I had extra time over lunch so I did a "Lunch Log".
I logged the RPM, timing, intake flow rate and O2#2
as you guys suggested above. For some reason TPS
wasn't logged, but I can tell you that the pedal was
on the floor. I also don't know why the flow rate only
shows zero lbs/min, obviously not what its supposed
to do. Don't know if this will help Kenamond & Linksys42 out much...

Based off of the O2V I can tell that I need to richen
it up at 2k and lean it out from 3k on. What about 1k,
do I just leave it where it's at? I'm also assuming that
only the Hi throttle tables need to receive the adjustments?

A little help please...

 

[spyderturbo007]

I would love to jump into this thread, but after the first post I started getting sea sick from scrolling around all over the place.

Is it possible for a mod to fix whatever is causing the problem? Thanks!

 

[kenamond]

Okay. Now it's looking a bit better. Don't know what's up with your airflow numbers. I'd sit there at idle and look at the list of things your logger can log and try to find if there's maybe a different airflow variable to log.

So all of the O2v start out good but then go rich. Try leaning it out a bit and get a sense for how much adjustment causes how much O2v change. Try to get the O2v all at 980.

Your timing curve "looks" normal, but it's tough to tell without airflow numbers. The timing will drop rapidly as boost first "hits" and then the timing advance should smoothly increase to redlin...and that's what your log shows. But you're still rich for most of the pull.

And I agree that whatever is happening with the left-right scroll bar on this thread is very annoying. In Netscape, it does't even give a scrollbar, so it's impossible to reply to the thread (I'm on a PC with IE7 right now).

 

[kiddsm95]

I'll lean it out some from 3k to 7k, but I'm
still confused about what to do at 1k. Should
I just add 1% of fuel in here to get the O2V
up to 980?

I agree, the formating is freaking annoying...
this is why I started typing 8 words per line,
then hitting enter.

 

[99gst_racer]

I would love to jump into this thread, but after the first post I started getting sea sick from scrolling around all over the place.

Is it possible for a mod to fix whatever is causing the problem? Thanks!

I'm trying. I fixed post #1, but I can't figure out where the other problem is....


EDIT: Fixed it.

I had to remove the tabled format in post #'s 8, 9, 11, and 20.

 

[99gst_racer]

Yes it is a forum problem I have posted logs before and it had done the exact same thing. I used the exact same instructions as here in this thread in the tech index http://www.dsmtuners.com/forums/art...st-logs-so-people-can-actually-read-them.html .

Actually, those tables seem to conform to the 'size' or width of the screen. And post #1 had a 1024 wide picture, so naturally, the tables that you guys made conformed to that size. Unfortunately, to save the thread, I had to remove those tables. If you re-do them, they should be fine now that the picture size in post #1 was corrected.

Sorry for the inconvenience.