Ok, I took the plunge to SD tuning. It seems to run good except the idle is very rich 11.5 AFR. I tried to make the Base Running Airflow numbers bigger, up to 12.6, that seems to help some, but I was expecting a bigger change.
I have not driven it yet to give it any time. Am I doing the right thing??

 

if you tell RAF that it's flowing more air, it's gonna only dump more fuel (aka rich gone richer). if you want to lean it out, take RAF number down. how are you measuring AFR at idle? does your cam have a decent amount of overlap? is that only cold or warm ECTs?

 

.. or drop your VE in the idle cells.
I pretty much just use RAF to stabilise Idle. Then lean it out with the VE table.

 

VE table is where the fueling accuracy is at in SD. If commanded AFR equals reported (WBO2) AFR, then the Open Loop Fueling table is commanding the richer number you are seeing.

 

It's a good size cam 233/239 600 lift. I am checking with a Wideband at idle, hot engine.
I will work on the VE table, Thanks guys.

 

Larry, if you need any more help, gimme a call, you know the digits. if it goes to voicemail dont bother, i never check it, just keep redialing

 

Does this help?

For VE Tuning with a wideband,

You need to:

- set the Engine/Fuel Control/OL & CL/Closed Loop Closed Loop Enable/ECT vs IAT to 284* across the board. This will put you in open loop along with setting the MAF fail freq to 0.

- on the same page as above, you want to disable LTFT

- also on this same page, in Open Loop/EQ Ratio, set all to 1.00 or I just change the ones from 140* and up to 1.00 (this will command 14.63 or stoich)
This step can be done several ways too. You can leave it as is if you want or change it to 1.13 which will command 13.0 AFR, it doesn't really matter what you are commanding, because the % error will still be the same. After going back to Closed Loop you won’t use this table anymore.

- in Engine/Fuel Control/Fuel Cutoff set DFCO enable temp to 284*

- in Engine/Fuel Control/COT, Lean Cruise set COT to disable

- in Eng. Diag/Airflow set MAF fail freq to 0 (you should also unplug the MAF if you can. If you are driving a 2001 and newer C-5 , the IAT is wired into the same plug, so you can’t unplug the MAF without doing some rewiring. They are seperate in the earlier C-5s)

- Now go log. Make sure you log VE for Actual AFR, Commanded AFR, and AFR Percent Error. I log first and then change the cell hits required to 25 for the % Error histogram.

- Right click/Copy the entire % Error histogram.

- Open Editor, highlight, right-click and use the “paste special/multiply by %” to the Engine/Airflow/General Airflow/Main VE/Primary table. As you get closer to 0, start using “paste special/multiply by –half”.

- Then, if you have a 1997-2000 C-5, copy the even numbered MAP lines to the secondary VE table by using the copy/paste method. Save and "write calibration only" to the PCM. Drive and see where you are.

Wash, Rinse, Repeat until your VE tables are in line. Your AFR % Error histograms should be as close to 0 as possible. It should only take 3-5 runs.

After this, it is time to tune MAF/WOT. Reset the MAF fail back to 14000 (leave everything else as is for now) and modify the MAF Calibration/Airflow vs. Frequency tables the same way you did for the VE tables, except you will use MAF vs Output Freq (Hz) histograms instead of the VE histograms. The MAF Calibration table is just above the VE table(s) on the same page in Editor.

After tuning MAF/WOT, set all of the above tables back normal and you’re done. Time to move on to spark, transmission settings or whatever you decide to do next. Have fun.

 

nope, wrong on multiple counts:

do not EVER set your OLFA to 1.00 across the board. leave it stock and don't touch it until you understand it on a whole new level.

using a regular AFR%Error method will show change in airflow as a change of VE, which is simply not true. unless you account for air density changes, your VE changes will follow the weather, not the breathing capacity of the setup. calculate air density first, normalize it based on that, THEN adjust VE.

copying every other row to convert from full- to half-res creates skewed numbers, because you will use averages for ie. 50-55kPa where you should be using 50-60. so on non-mild setups where efficiencies can change quite dramatically, you will make a mess of your VE. if you're tuning with the half-res table, use a half-res table to scan into, and make changes based on that. you lose resolution, but at least it's an average for a full range not for just some of it.

leave DFCO on, no reason to dump unburned fuel onto hot O2's. you want to filter out deceleration cells for VE tuning anyway, it's not 'steady' airflow.

for MAF tuning, just dump all your logs into excel, create a 'corrected' airflow column where you multiply your error by the airflow you thought you had, and then use that data to create a 3rd order poly on that corrected airflow vs mafhz scale. use the resulting polynomial to calculate proper calibration points. you get brownie points if you also programatically filter out obvious wrong values

yup, everything you know about VE tuning is wrong... ain't science fun?

 

Thanks Guys, I'm in a 99 C5 and I just added a LS2 TB/ FAST90 and removed the MAF.
It actually drives better now than it ever did with the MAF. Now for some tuning fun. I like the idea about disable the LTFT. It seems like they are always overreacting and going from way rich to way lean. I am going to the race track tomorrow for one SCCA practice session. I will log some data and report back.

 

Hmmm, interesting as I have my DFCO disabled completely like everyone else. Maybe deep decel DFCO should be left on just for that point you mention, raw gas on hot O2's? either way, got my deep decel VE table down fairly well, though very accurate in that it is a moving target and can never get it completely correct.

Very interesting regarding VE tuning and air densities. I know I have a very large problem day to day, hour to hour tuning here in NJ as the air seems to change every run out on the road. Do you have any method, esp with HPT of normalizing the change in weather? I always get the same issue of gettin vE correct, but then air changes and my MAF tune is off.

 

I too would like to know how to normalize the values. That may be a helpful spreadsheet if it's a proven-to-work type of thing. Right now, my attempt at dialing in my VE a little closer has me doing two things out of the ordinary:

1) I try to tune at the same time of the day from one day to the next. I know weather variations are hard to get around. But with it running consistently richer in the morning and leaner in the afternoon, I've decided to stick with using afternoon logs only to keep from fighting myself.

2) I'm using filters to try and keep me near the center of the VE cell. In other words, I don't like the data near the edges of the cells because there's more influence from the next cell over due to interpolation. So, I set filters up to show me data from 1150~1250 RPMs, 1550~1650 RPMs, and so on....

As good of an idea as it would be to develop a spreadsheet to normalize for conditional variations, I think the complexity will be too high. We're not only talking changes in atmospheric pressures, humidity, and temperatures from day to day...but, changes in the combinations of IAT's and ECT's as well. IMO, there needs to be a way hold the charge temp calculation to a constant in order to have a consistent VE table. I believe GM's solution to this was to build a multi-million dollar weather generating facility to control these variables....but, I don't have that kind of $bank$.

 

SSpdDmon, you're totally on the track. by seeing the 'hour to hour' differences you are exactly describing the airdensity throwing you off, thinking you have VE differences, while what you really have is an airdensity change. VE is a breathing ability descriptor table, not an airdensity table. So by using filters and times of time you're basically limiting temp ranges basically narrowing down airdensity range. so you're aiming for the same thing i am, i'm just going more methodically about it. unfortunately excel is not exactly suitable for the job. i got it done in MatLab, but i doubt many people here have it available. Next month i'll get it done with probably some custom code but for now it's a proof of concept/engineering sample only.

another very good point you bring up is making sure than averages are gathered truly across the full range covered. that's a much harder problem with the limited data we usually get, so it becomes a math problem. i have found some methods to clean up the data, and to arrive at the averages from the middle of cell, but again, excel wont do it.

i'm not quite at 'proven to work' stage yet, but I got enough math and physics behind it that I'd be suprised if it didn't work. if you want to see the work, i can send you my derivations, email me or something.

 

Actually, I think I've accomplished the data cleanup with a complex filter in EFI Live. After listing the usual lines that eliminate throttle transitions, I basically told it to only accept data where the MAP kPa ends in *0, *1, *2, *4, *5, or *6. Then, after 60 lines of filter code, I realized it would have been easier to tell it to ignore data where the MAP ends in *2, *3, *7, and *8. There's also a couple of lines that say to ignore data above 'x' RPM and below 'y' RPM too, where x would be set to something like 1250~1300 and y would be set to something in the range of 1100~1150.

What does this all mean? Well, the down side is I have to edit my data 1 RPM row at a time and I'll have a lot less data to work with. But, the plus side is the quality of the data should be much better, which means (in theory) I should be able to dial in the VE with better accuracy. Using the same time of day should help too assuming the weather stays relatively consistent.

 

Not wrong, just different from what you think. This tutorial is designed for beginners to help get acclimated to the software. It has been scrutinized by several of the experts at HP Tuners (Bill, EC, Russ, Soundengineer and others) and OKed by them. I know you are an intelligent individual, but you tend to overcomplicate things and are way to ****. I don't mean that in a bad way, but not every one is so worried about that level of perfection as a beginning tuner.


Setting OLFA to 1 is fine. It makes things easier as you command 14.7, not a changing AFR that is hard to follow for a beginner. You can set it back to stock when you are done tuning if you so desire. You don't use this table when you go back to closed loop anyway.

My way is what is taught by the software designer and is the easiest way for the beginner to learn and will give you good results. After someone learns the basics, they can tune any way they see fit.


Sure you can log and make changes to the secondary VE and then copy it to the Primary and smooth, but which way is easier for someone that is still trying to grasp what the word stoichiometric means?


Where is unburnt fuel coming from? Sure there is a small amount of fuel going into the cylinders, but spark is there too. It doesn't cut off the spark and leave the fuel on. Plus, DFCO is set very lenient from the factory, so that a stock vehicle doesn't even enter DFCO unless they are over 2000 rpm.


Are you related to Rube Goldberg?

My MAF and Dynamic Cylinder Air match almost perfectly, which is the basic goal of the MAF tune. You can't tell a guy that just bought HP Tuner to do the above and expect results. My tutorial was designed to help take some of the frustration out of learning curve, and I don't really appreciate you bashing it just because you think you have a better way of doing things, and you need a doctorate in physics to understand. Again, this is the way it is taught by the designer and seller of the software, so it can't be too wrong.


That is your opinion and opinions are like ********, everybody has one and everybody else's stinks.

 

Ok guys help me out here. SD tune, no MAF. I was on the track today and my WB scan is rich below 4000 and lean above 5000. I run this motor to 6500, mods are in sig. That was with a screwed up VE table, I copied the wrong one last night. So tonight, I put in my best VE table, went for a run in the country, and it is a lot better. It breaks the tires loose in 2nd and at 6500 I'm going 130 mph in third so I have to back off, but it has the same pattern, rich below 4000 and lean 5000-6500.
My PE is pretty flat and OLFA flat. I know you are going to say work on the VE table. OK, very basic question, which way do I go with the numbers on the VE, where it is rich <4000, make the numbers lower?? lean >5000 bigger?. I'm don't have much experience with this and I need to get it right before my track day tomorrow.
Thanks

 

A SD tune is essentially the same as a MAF tune, just without the MAF. It's done the same way. If you need more proof, just look at your help files in Scanner or Editor. It's the same as what I wrote, I just simplified it a little. You can tune your car in a few hours.

 

Ok, I read the instructions, I will add to the VE where it is lean. I don't have the luxury of driving and working out all of this tonight, I just have to make changes from looking at the WB during WOT.

 

so what you're saying that it's OK to teach people how to do things wrong, because they're new to it? like it's OK to teach kids that 2*3=5 until they reach 5th grade, at which moment you teach them that it's really 6?

my friend's dad has this saying "there's dumbing things down, and then there's just being dumb..." i'm all for automation and making things easier (that's why i make all my tools public) but i refuse to simply things simply because some people have problems understand it.

no it's not fine to set it to 1.00. the only reason to set it to 1.00 was back in the day when few people had widebands and we had to use narrowbands to get AFR readings, which could only tell you when you were commanding 1.00. the fact that you keep repeating it means you don't understand how does the corrective factor gets calculated. AFRwb/AFRcommanded will get you the right factor no matter what AFR you're commanding. do not dismiss what you don't understand.
in the meantime the 1.0 OLFA will cause problems on cold starts, making people wanna abandon modding in and tuning. it also takes away the 'pick richer' safety scenario when you combine it with the stupidity of disabling PE.

except that it never yields steady results for a lot of people, and the new stuff like E40 computers completely dont operate on the same set of rules. principles are important, once you teach them wrong, they'll refuse to adapt later (oldschool tuners, LS1Edit and IFR tweaking comes to mind...)

again, you're completely not understanding how to use histograms correctly. you can have two histograms, two separate 'views' onto the same set of data. if you're done tuning the secondary and you want to convert the same settings to the primary, just have two histograms, displaying full- and half-res versions of the same thing, and then apply them on the last refinement.
it's not hard to do, you just have to understand the process, and have a good degree of familiarity with the tools provided.

unfortunately i'm not jewish enough

I wasn't bashing, i was trying to explain things for the sake of progress.
doctorate in physics? so far we've used proportions, formulas for density and rate/amount/time relationship. that's all elementary school material. the fact that the process aint easy i think speaks the most for the quality of software we're forced to use. most of stuff should be self-evident and automated. the number of spreadsheets floating around to help out with the process is a testament to how much the tuning tools are lacking. and the fact that they havent been integrated into these tuning tools after years of floating is a strong statement on how much they care to address the needs of the users, the people without which they would not make a dime to start with.


formulas and derivations are not opinions.
science is not a popularity contest.
the fact you cannot understand it all doesn't make me an '**** rube goldberg relative'
so stop spreading wrong info until you can prove it's right.

 

OK, Enough bashing, let's get back to the important stuff, the tune. I went to the race track today for qualifying. Came in third, a 2006 Z06 427 is on the pole, 911
in the front row, then me and another 911. It should be a good race. The 911 that is beside me is only a few 100ths from me, the one in front is .3 sec. OK guys we need a good tune for this race on Sunday.
I added to the VE table last night and tweaked the PE to try to flat line the AFR. Well, it worked. From 3500 to 6500 it's pretty good, just a little rich, I can fix that with PE.
Got ready to drive home and the car would not start. Cleared code102 and it runs.
I tried to uncheck the SES box for 102 yesterday and the car would not start. With my SD tune, should I just set the MAF fail freq to 0, and leave the SES light box turned on on ? Will that do it.

 

I believe (I'm a novice tuner so please no bashing but please do correct me if I'm wrong) that you can turn the MIL light off for that code and then set the MAF fail to very low... I don't know if you can set it to 0 (again just going off of memory).. I think that you have to have some sort of value but it has to be very low so that it won't use the MAF at all because it will fail out early...