Help with VE cleanup

[cobaltssoverbooster]

I have been working to smooth out a ve table and i'm not having any luck.
99% stock 6.0 with the 1% being a magnaflow replacement muffler that is all.
I default to sd mode, leave maf plugged in for iat, i have an innovate mtx-l (properly offset to match scanner to gauge).
If i use the gauge, my corrections using afr error results after the any sensor based changes are just horrendous. i can be running within 2% then when i apply that correction i'm suddenly off 10%.
If i use the Trims method by enabling stft in open loop then i get the values to come down and i can verify it with my gauge but it puts spikes and ridges all over the ve table. i'm currently running in maf mode because i couldn't figure out what i was doing wrong in the sd ve to get so many ridges.

I have heard talks of transient dx/dt filtering in the scanner but i cant figure out how to use it and the forum doesn't say much about setup. The vcm help file is useless, basically just lets you know the software has the feature.
Also heard talks about innovate sensors needing signal speeds reduced through LM Programmer software. Does changing the port speed through the sensor software really help that much? I never reduce log rates when using them with megasquirt or haltech, but then again those are better units than my stock sloth ecu.

My vehicle is an 04 yukon denali fixed awd with the lq9 (non-egr, and stock return-less fuel system) I do not have access to an awd loaded dyno in my area. i try to drive as smooth as possible without jerking the throttle. my best guess is i am still having issues filtering out transient data to give me steady state correctable data. Any suggestions to filtering transients out?
I attached both my SD file and my MAF file. I'm not exactly happy with the wave i have going on in my maf calibration either. I understand its not desired but its getting me the afr targets to get by until i figure out a new move to try.

Thanks,
Alex

[dermotw]

The MAF curve should really be as smooth as the standard one, and that will take some effort...
I hope you are using "multiply by 50%" for the VE...

[cobaltssoverbooster]

I always use 50% correction factor. I found that always gave me the best results when I did ecotec stuff so I've continued to do so.
I set up an excel file that removed 50% for me because I thought the editor functions we're corrupted but I got the same results using excel as I did with the editor functions. Still swings hard per correction using afr error.

When I tune with megasquirt I tune the ve in and then add 10% fuel to see if the offsets are correct on the injectors. I haven't done that with this because its all stock and I figured that would be the most accurate data. Could be off now that the injectors have 150k miles on them. No system leaks and fuel pressure is regulated properly to 5800 rpm (that's as high as I let it go)

[dermotw]

I use Innovate btw and never needed to reduce its speed. You do need to filter out stuff you don't want biasing the results, obviously (overrun, idle maybe...).

I never got/get swings like that. Possibly something isn't setup or is left on??

Hope yr XL is doing what HPT does.. i.e. you get 1% error in some cell in the graph, "multiply by 50%" is 0.5% error, and so it changes the cell by .5%...

Heres the VE error from your own log, no peaks/troughs in the errors, looks pretty usual for a non-tuned set of values...?? So I'm not sure why u have 15% peaks etc. - if you applied the table below to a std map then there wouldn't really be any?


Capture.JPG

Oh yes, I'm sure you know, but you need to collect a reasonable sample, say 20 hits/cell, perhaps more with VE than with MAF, some people say 60, but I usually use between 10 and 20...

[dermotw]

As a test, I took your VE data and dumped into a std VE table (from another vehicle, lol). I get this (using copy/paste, "multiply by half");

Attachment 70596

As you see, no big changes caused by that...VE got slightly more 'lumpy', thats all..

[cobaltssoverbooster]

i will have to flash in my ve file again and try setting cell counters up.
Ill try filtering some data ranges as well. tps vs rpm based stuff.

my negative spikes seem to all be throttle related such as a lifting after a short pull and it goes rich on the plate close. i really wish i could filter based on tps rate of change.

[dermotw]

You can make filters like this (not a particularly good example, it for something else) just to give idea;

Capture.JPG

Cell hits is there too, set it to 10 or 20 for example. This example filters on the FTC (fuel trim cell) and'd with the average change of throttle in last 100mS (I think!, I haven't used it for many months!). But you see the way..

[dermotw]

Ha I got interested and went and looked... I checked, this worked ok;

Capture.JPG

Its only a test of course. It finds any (manual) throttle change of over 40% in the PREVIOUS 500mS from the current point. If you put -500 it would find ahead (only in a log, obviously).

[cobaltssoverbooster]

Ill test this out and report back this week after i get some data on it. thank you

[CP_CUSTOMS]

so when you guys multiply the VE by 50% are you doing the whole table or just portion

[cobaltssoverbooster]

We are only multiplying the ve table by half of it's logged error. If you use a copy and paste method this half error will only apply to the cells that have recorded values in the scanner software. We do not directly multiply the ve by 50%

[mightyquickz28]

What version of HPT are you using? One release I downloaded provided absolutely crazy results when I logged the wide band/MAP. I went back to a prior version for awhile till a newer release came out. All us well now.

[cobaltssoverbooster]

Version 3.4.84

[cobaltssoverbooster]

I updated the software to 3.6 because of the ecu recovery options listed by hpt.
To shorten the story i lost my configuration of my wideband which isnt hard to setup but i need to do it again.
I will be running tests to see if the new software calibrates correctly and if it remains the same as before then i will be reviewing the log data to see if i have injector fueling errors.
Currently testing in maf mode because its the easiest option to adjust fueling fast.

Id like to do some tps slope filter testing as well to limit transient conditions to make the tuning cycle more like solid state simulation.

[cobaltssoverbooster]

ok so i have returned to a smooth curve on my maf to see if i am having injector issues.
im running the current version of the software as of 7/19/17 and it is working very nice.
im just going to put my file and recording here as a base so as i post more stuff other people can see what kind of progress was made.
Fuel Verification Test File HPT.hpt
HPT Base Fuel Testing Layout.Layout.xml
Fuel Verification Test_Base 1.hpl

[cobaltssoverbooster]

correction batch 1. will get longer log tomorrow assuming my interface doesn't error out like it did today.
Fuel Test Correction_1A.hpl
Fuel Test Correction_1B.hpl
Fuel Test Correction 1.hpt

[cobaltssoverbooster]

Had an issue with the car trying to not idle down properly. It was deemed a dangerous scenario and the RAF was revisited before continuing.
Some data will be thrown off but thats life.
Fuel Test Correction 2.hpt
Test Correction 2A.hpl
Test Correction 2B.hpl

[cobaltssoverbooster]

I don't have a log file or tune to post for this update but it should make sense without it.

I recently went on a few trips around the local area to gather some data. The first trip out i notice the Voltage gauge swinging from 12-16 Volts. HAHA S#$t the voltage regulator died. Thanks to Powermaster i now have a new 180 amp alternator which was technically needed due to the installed stereo equipment dropping system voltage when used. (When i logged the last tunes for correction data the stereo was disabled to prevent voltage issues in case that crossed your mind) So after getting running again with solid voltage i make the drive to Sonoma for an awesome weekend NHRA event. Watch some wicked cars shred some asphalt only to come back to a car that has no power at all. I tried jumping it with no success so as usual i figure the fried voltage regulator killed the battery which is what some of you may guess as well. Well in order for that to be true you would have to know all the wiring was legit in the truck and it apparently was not. The main battery cable was not corrected properly to an aftermarket buss plate bolted to the side post of the battery. Turns out the battery was chilling at 12.7 volts completely disconnected without being recharged. The main system harness lost proper connection and cause low voltage to every vehicle system including VATS which shut the fuel pump off when trying to jump it. So everything is running perfect and the voltage issues gave me this idea that when an injector is flow tested it has voltage offsets specific to its tested pressure which if also tested at running voltage can be called your zero coefficient test. So knowing the system operates at 14 volts basically all the time thanks to the regulator, i went and disabled the voltage offset table by setting it to null value from 13.5 Volts and above in all pressure zones. This means the injector flow is unmodified with exception to anything that falls under the short pulse adder. I do a quick calibration of the maf and the results are, for me, quite astonishing. With the maf tuned in this new mode i see the least amount of afr swing i have ever seen in any of my vehicles. The wideband reports back a max rich of 14.4 and a max lean of 15.2. When the modifier was left stock the wideband reported max rich of 13.7 and max lean of 16.5 which is a monster spread and hard to adjust on. It seems for the application i have, the injectors no longer flow the way GM has expected them to per the operating voltage.
The next files i will upload will show how off my base injector data is by separating it into logs which can prove fuel error by monitoring the trims over a cell count average of at least 1000 hits per read. This discovery, which is new for me, is getting me excited to tune again as i have now taken my AWD 6.0 from the bone stock 13.5 mpg to 16.8 mpg average city (out here average speed is 35-45 mph round trip for a 30 minute drive).

To those who have offered help i appreciate everything you have done. Now that mechanical issues are resolved i can hopefully share some legitimate data.
p.s. if you have a modified electrical system in a vehicle from 2000-2006 gm made a secondary battery bracket that fits on the passenger side of the Tahoe, Yukon (including Denali), and Silverado chassis. Highly recommending installing this second battery for vehicles that tax the electrical system harder than designed by the factory.

[cobaltssoverbooster]

i didnt mean to keep everyone waiting but i got distracted by working on a few local race cars and straightening some stuff out on my own door slammer project.

anyways i figured out what kind of issues i was having and sorted them all out. car is running the best i have ever had it.

so heres what was broken: 1) voltage regulator was swinging hard 2) the audio connections at the battery caused loose terminal problems 3) the audio system would dip voltage when running at a low setting.
the repairs i made fixed all of these running issues, and those are: 1) new powermaster 180 amp alternator 2) second battery installed with 4 gauge wire 3) 4 gauge charge wire to replace stock 8 gauge charge wire to stock battery at engine harness 4) repair and modify battery and audio terminals to prevent connection issues.

so after these repairs fixed the running issues i noticed the wideband still fluctuated pretty horrendously so i set out to fix it. first i recognized that at the operating charge voltage the injector should be running true flow. the voltage offset should be used for drops in voltage when the charge system has failed. First change was to default the injector voltage offset to "0" where the alternator range operated. so from 13.2-max voltage i set the multiplier to "0".
immediately after this i noticed the wideband stop dancing as much. with it now somewhat consistent i tuned in the maf sensor to as close to 0% trim as possible. the car ran so smooth but i noticed it was still trimming irregularly when decelerating and accelerating when on hills. so i set out to correct that issue as well. To fix this trim issue i recorded fuel trims against the base injector flow and made sure those reported close to 0% trim error. then i added 10% to the entire maf table and recorded the error again against the base injector flow. anything that was not reporting back close to a 10% trim was modified by hand to 1) bring the value closer to 10% return and 2) keep the injector base flow linear.
Turns out the factory injectors i had no longer flowed properly with 150k miles on them. on average they dropped roughly 1-1.5 lb.hr in flow. after this change things really started settling out and the wide open afr changes were very accurate; to the point i could make small trim changes with using the multiply by "full" percent function.

the stock injector data was close but to prevent wandering i would try disabling the voltage offset in the alternator charging zone to minimize corrections on the base flow. this steadied the fuel trims out the most.
results of the testing:
i have achieved 16 mpg on a stock 6.0 with just a magnaflow muffler swap. this vehicle still has the factory awd and 3.73 gears but sit on 20" wheels with 285/75r20 michelin tires. Not the best setup for getting gas mileage but i made it work. The original mpg when i picked the vehicle up was 12.7 mpg in the same conditions. i took this tank to the race track and it runs 26* advance on pump 87 octane fuel without detonating (verified with haltech knock ears). The pass was shamefully slow; a whopping 16.4 sec at 84 mph, 2.4xx 60 foot. Afr was 12.3 steady through every gear(glancing at gauge). My track doesnt allow open laptops on the track so i wasnt able to log since i had a std interface module instead of my shop pro module. I dont have a track time for the stock setup so im not sure how much i gained on the track but knowing it ran 10.3 afr on the tune that was in it stock and felt super slow. i would bet the track time would have been a high 19 or 20 second run.

Here is what the file has evolved into. i have not revisited ve since this testing. currently still set to maf only tune.