VVE Tuning Oddities

[LaTechGTO]

I went out today on a rolling dyno after making some changes (thanks jsllc!), and I noticed few oddities while logging and adjusting the tune. Does anyone have any ideas?

Part throttle is LEAPS AND BOUNDS better!!! During WOT pulls, I noticed the following:

1) Air Calc Mode will sometimes switch sporadically from "Normal" to "Hi Speed" to "Unknown" somewhere around .88g Cylinder Mass (~7,800Hz; MAF Hz is still recording despite being failed). Almost like clock at .88g on every log from today.

2) Again, although the MAF is failed, the Hz is still recording. This isn't that unusual as long as it doesn't take the readings into account for fueling, but during a few runs (not all) the MAF Airflow would calculate until WOT was achieved then would go to 0lb/h. Very sporadic when it would just start producing a mass flowrate.

3) During the same runs as mentioned in 2, something weird would occur...

1st pass where it was noticable - "WOT4_weird". It occurred just shy of 4,800 RPMs (~8,000 Hz) then the cylinder air mass would drop to 0g. This would result in a spike of advance then followed by knock. Almost felt like the engine just shut down for a split second then would kick right back on.

2nd pass (WOT5_weird), which was the following WOT pass ~3min later. It occurred right at 4,900 RPMs (~ 8,360 Hz), and the cylinder mass dropped to 0 shortly afterwards.

After this happened twice, I decided to take it to the house.

I am puzzled, but it's almost like it's referencing my something else (linked to the Air Calc Mode issue?), which is leaning it out severely. I still don't feel like everything is failed correctly while I am trying to tune the VVE. The car is definitely responding to the adjustments I am making so there is a silver lining to all of these trials.

Again, I am open to any ideas or suggestions!

V6.3_VVE_MAFFAIL .hpt
WOT3.hpl
WOT4_weird.hpl
WOT5_weird.hpl
Main.Channels.xml

Thanks!

[jsllc]

Need you to all all of the channels in the list attached or copy into your Documents\HP Tuners\VCM Scanner\Channel Configs directory and load it, then just remove the ones your ECM does not support. Then log the issue. You just do not have enough data to tell you what is wrong.

[jsllc]

So took a look at your VVE. There are serious issues. Why is your MAF all Zeros after 3000 hz????? This is why MAF flow reads zero. Your VVE is not being calculated right. Not a good tune and should not be used especially at WOT. Not trying to rain on your parade but do not want to see this go boom either.

On airflow. E38/E67 use dynamic 100% of the time. VVE / SD / MAF and Predictive Coefficients are used as the basis for calculating dynamic with a formula we are not allowed to see or modify. Under certain RPMs the VVE or SD (depending on OS patches) is heavily biased. Over certain RPMs the MAF is heavily biased. We get to edit this change point. It is never 100% either way. MAF if mechanically designed and tuned correctly is always more accurate than SD or VVE. SD and VVE are not the same. VVE is virtual based on 30 cells of formulas. The cell sizes are adjustable and we can edit the coefficients used to calc these values as well as how it transitions between them. They are infinity scalable and close to actual. They will never be as close as MAF or SD if SD has enough cell resolution and tuned in all cells. SD is more accurate than VVE on a given tuned cell but not scalable and not as accurate overall because not all cells can be tuned and we are never given enough cells.

[OneTwo]

..........SD and VVE are not the same. VVE is virtual based on 30 cells of formulas. The cell sizes are adjustable and we can edit the coefficients used to calc these values as well as how it transitions between them. They are infinity scalable and close to actual. They will never be as close as MAF or SD if SD has enough cell resolution and tuned in all cells. SD is more accurate than VVE on a given tuned cell but not scalable and not as accurate overall because not all cells can be tuned and we are never given enough cells.

Could you elaborate on this please? For an e67, the VVE table is created by the formulas whose coefficients are changed on the tab that is named SPEED DENSITY under the heading of VE COEFFICIENTS. Are you saying that that is not what these are? Or, are there more SD settings somewhere else? Or, is all of it mislabeled?

TIA

[jsllc]

There are OSs that have patches that are true SD tables and some older ECMS have it as well. In an e67 VVE is the replacement and there are no patches. It is under the SD tab in HPT but it is not SD.

[OneTwo]

Thanks for the clarification.

[LaTechGTO]

So took a look at your VVE. There are serious issues. Why is you MAF all Zeros after 3000 hz????? This is why MAF flow reads zero. Your VVE is not being calculated right. Not a good tune and should not be used especially at WOT. Not trying to rain on your parade but do not want to see this go boom either.

On airflow. E38/E67 use dynamic 100% of the time. VVE / SD / MAF and Predictive Coefficients are used as the basis for calculating dynamic with a formula we are not allowed to see or modify. Under certain RPMs the VVE or SD (depending on OS patches) is heavily biased. Over certain RPMs the MAF is heavily biased. We get to edit this change point. It is never 100% either way. MAF if mechanically designed and tuned correctly is always more accurate than SD or VVE. SD and VVE are not the same. VVE is virtual based on 30 cells of formulas. The cell sizes are adjustable and we can edit the coefficients used to calc these values as well as how it transitions between them. They are infinity scalable and close to actual. They will never be as close as MAF or SD if SD has enough cell resolution and tuned in all cells. SD is more accurate than VVE on a given tuned cell but not scalable and not as accurate overall because not all cells can be tuned and we are never given enough cells.

I set those fields above 3,000Hz to 0 because I was wanting to verify the MAF was failed - the region around 3,000Hz is achieved at part throttle conditions. See my other thread you replied to here. I listed this as one of the items I changed to force OL"SD" in the first post.

It didn't fumble above 3,000 Hz so I thought everything was setup correctly to tune VVE; same methodology as when tuning older generation cars - fail MAF and tune VE. I find it very odd that this issue didn't rear it's ugly head until the 4th WOT pull and somewhere around 4,800-4,900 RPMs and 8,000+Hz.

So if the e38 uses dynamic all of the time, how do you go about properly tuning the VVE? I thought for sure it would be as straight forward as failing the MAF, setting the Dynamic Airflow Enable to a higher-than-achievable RPM, and disabling related fueling adjustments (COT, DFCO, etc.) then using the EqErr% to adjust the VVE tables. Maybe I didn't tune the MAF correctly either...? My WOT fueling seems to be pretty close though.

I realize the MAF is expect to be more accurate, but I believe there is a benefit to having an equally accurate VVE. Maybe I'm wrong? I obviously don't know enough to stand by that 100%.

There are OSs that have patches that are true SD tables and some older ECMS have it as well. In an e67 VVE is the replacement and there are no patches. It is under the SD tab in HPT but it is not SD.

This is definitely confusing/misleading in the software, but I can see that there probably isn't a better place to store this information and it sort of correlates.

[jsllc]

You do not remove the settings from MAF to tune VVE just like you do not zero the VVE to tune MAF. Your VVE is way off in its set of coefficients as well. Should not be going negative at higher MAP levels, Should not have big jumps between areas. Here is your VVE extended out using what you have supplied.
12-24-2017 9-06-02 AM.png

[LaTechGTO]

You do not remove the settings from MAF to tune VVE just like you do not zero the VVE to tune MAF. You VVE is way off in its set of coefficients as well. Should not be going negative at higher MAP levels, Should not have big jumps between areas. Here is your VVE extended out using what you have supplied.
12-24-2017 9-06-02 AM.png

I was just using "0" values above 3,000Hz as a check step to ensure the MAF was failed. I reckon I need to break old habits where if you failed the MAF, no MAF readings were taken into account for fueling. As I'm sure you are aware, previous PCMs referred to this as SD and allowed you to properly tune the VE. The only way to properly tune SD was with the MAF completely failed. Apparently the e38 acts much differently.

Just to clarify - are you saying that even though I raised the Dynamic Airflow High RPM Disable to 8,000 RPMs, the MAF data is still utilized? The definition for ECM 3003 states "Above this RPM use filtered MAF airmass for airmass prediction calculations.", which leads me to believe that it wouldn't use any MAF data below this RPM threshold. This is quite confusing.

To the contrary, is VVE fueling information being applied while tuning MAF with the Dynamic Airflow High RPM Disable set to~10 RPMs?

I am having a hard time seeing how you can accurately adjust an item (i.e. VVE) with so many variables (i.e. MAF based fueling) going on in the background without applying infinite iterations.

Just so you're aware, I promise I am trying really hard to get my head wrapped around all of this. Admittedly so, this is only through my individual testing and reading MOUNDS of information across the interwebz. The latter causes great difficulty because it is seemingly impossible to decipher between truth and unintentional BS.

My apologies, but I am not sure how to reproduce that coefficient table you provided. I thought my VVE looked acceptable, but apparently the coefficients are junk. I understand the basic concept of the coefficients, and I realize they should not only be logical (RPM vs MAP airflow/fueling relationships) but they should also provide smooth transitions (it appears I have failed miserably here).

VVE.jpg

Thank you, again, for taking the time to help out!

[jsllc]

Filtered MAF means Sample rate of the MAP sensor. We do not know the formula but an example would be 9 samples from Volumetric Efficiency Airflow plus 1 sample from MAF divided by 10. The ECM always uses all sources but weights the primary source heavily. This is why I keep saying VVE is NOT SD and it is not used in the same way. It is nothing more than a backup designed for smoothing MAF at lower rpms and for getting you home during MAF fail. I will make people howl here but, anyone using it as primary is looking for an issue and will never have a car that runs as well.

To see how your entered coefficients look, you have to extend them. HPT does a very good job by allowing you to define the rows and columns. I use these.

400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 6400 6600 6800 7000
15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 131 135 139 143 147 151 155 159 163 167 171 175 179 183 187 191 195 199 203 207 211 215 219 223 227 231 235 239 243 247 251 255

Your car will not hit these but you can see the trend created by your entries. I also use formulas based on accurate MAF readings to build a base VVE the is within 5% that I modify from there.

You also need to take into account injector tip temp when doing this. If you do not I can show you where this can very 20% from one time to the next. Make sure to log all of the PIDs that I have in the channel list attached in this forum. I believe what you have done will work just it has some trends that are not so accurate. Just remember On airflow. E38/E67 uses dynamic 100% of the time. VVE / SD / MAF and Predictive Coefficients are used as the basis for calculating dynamic with a formula we are not allowed to see or modify. This is scaled based on a formula that is neither MAF nor VVE as both run out of range but dynamic keeps going based on trend created. Infinitely scalable. Example attached shows both MAF (655) and VVE (512) flat line. Dynamic keeps going and AFR is stable. Tables run out of resolution Formulas scale.


12-24-2017 12-00-48 PM.png

[LaTechGTO]

Filtered MAF means [FONT=&]Sample rate of the MAP sensor. We do not know the formula but an example would be 9 samples from Volumetric Efficiency Airflow plus 1 sample from MAF divided by 10. The ECM always uses all sources but weights the primary source heavily. This is why I keep saying VVE is NOT SD and it is not used in the same way. It is nothing more than a backup designed for smoothing MAF at lower rpms and for getting you home during MAF fail. I will make people howl here but, anyone using it as primary is looking for an issue and will never have a car the runs as well.

So if I am understanding this correctly, the Dynamic Air Flow RPM is nothing more than basically a bias between MAF and MAP. If this is true, then I would set the RPM Disable and Re-enable to 8,192 and 8,191, respectively, in order to properly tune the VVE (similar to how it was set before). Correct?

I can see where you can arrive at a VVE tune this way, but it still seems very iterative and could possibly require some MAF retuning once the VVE is complete.

To see how your entered coefficients look, you have to extend them. HPT does a very good job by allowing you to define the rows and columns. I use these.

400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 6400 6600 6800 7000
15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 131 135 139 143 147 151 155 159 163 167 171 175 179 183 187 191 195 199 203 207 211 215 219 223 227 231 235 239 243 247 251 255

Your car will not hit these but you can see the trend created by your entries. I also use formulas based on accurate MAF readings to build a base VVE the is within 5% that I modify from there.

*Scratch my previous comment - figured it out.* It is really weird that going to a 2 bar view will create a totally different look.

I'm fairly confident in my MAF tune as it was much more straight forward than my piss poor attempt at tuning the VVE. Makes part of me want to stick to MAF only, but I am very hardheaded and can't stand the feeling of defeat. How do you convert the MAF calibration to VVE?

You also need to take into account injector tip temp when doing this. If you do not I can show you where this can very 20% from one time to the next. Make sure to log all of the PIDs that I have in the channel list attached in this forum. I believe what you have done will work just it has some trends that are not so accurate. Just remember On airflow. E38/E67 uses dynamic 100% of the time. VVE / SD / MAF and Predictive Coefficients are used as the basis for calculating dynamic with a formula we are not allowed to see or modify. This is scaled based on a formula that is neither MAF nor VVE as both run out of range but dynamic keeps going based on trend created. Infinitely scalable. Example attached shows both MAF (655) and VVE (512) flat line. Dynamic keeps going and AFR is stable. Tables run out of resolution Formulas scale.

12-24-2017 12-00-48 PM.png

Thanks for the heads-up! I have actually been watching it like a hawk after reading numerous threads where people were having issues with IVT and Inj Temp Gain tables doing screwy things to the fueling. I think my OL and CL specific gain tables are acceptable (not perfect by any stretch of the imagination), but I suspect [12303] will need to be adjusted some eventually as summer returns. With my car being a basic bolt-on car with e85, it doesn't seem to be as susceptible to fueling swings although I have identified a few areas of opportunity.

That view of the log and statement "Tables run out of resolution Formulas scale." is very thought provoking and really drives the importance of getting everything dialed in correctly. Thanks for sharing.

I am still perplexed as to why this progressively worsened with each passing WOT pull yesterday.

*Scratch this too... think I found out.*

[jsllc]

I have always failed the MAF to tune VVE. Dynamic is more than MAF vs MAP. It is a formula that uses them plus the Predictive Coefficients. This formula scales well beyond the MAF or VVE airflows. It is what keeps the car going regardless of the sensors. The reason you will never get a tune to work by tuning MAF and VVE only is that you adjusting 2 of 3 legs. We have no tools for the Predictive. Many just zero them to make it closer. As for Injector Tip temp, unless you have it right you will chase you tail forever on fueling. I make sure of 113, 122, 131 and then filter my MAF and VVE to only used those temps for info.

Your VVE was pretty good just had the wrong trend as it scaled. Formulas are funny the at way. A little off at the start can lead to going way off in the wrong direction in the end. The e38/e67 are well thought out. They were the beginning of the end for tables as we think of them. Beyond these they change it all and base everything on TQ modelling. Much more difficult to master. If you really want to see a mess open up a tune for a 2017 F150 Shelby Truck. It is SD only with 168 air tables and 56 spark tables that have to be reverse calculated based on inverse TQ tables from driver demand TQ and trans TQ management. It sucks.

[LaTechGTO]

I have always failed the MAF to tune VVE. Dynamic is more than MAF vs MAP. It is a formula that uses them plus the Predictive Coefficients. This formula scales well beyond the MAF or VVE airflows. It is what keeps the car going regardless of the sensors. The reason you will never get a tune to work by tuning MAF and VVE only is that you adjusting 2 of 3 legs. We have no tools for the Predictive. Many just zero them to make it closer. As for Injector Tip temp, unless you have it right you will chase you tail forever on fueling. I make sure of 113, 122, 131 and then filter my MAF and VVE to only used those temps for info.

Understood. I just wanted to make sure I was supposed to fail the MAF and shift the Dynamic Airflow High RPM settings at the same time like I was doing before. Good idea to filter based only on that range of injector tip temp.

Your VVE was pretty good just had the wrong trend as it scaled. Formulas are funny the at way. A little off at the start can lead to going way off in the wrong direction in the end. The e38/e67 are well thought out. They were the beginning of the end for tables as we think of them. Beyond these they change it all and base everything on TQ modelling. Much more difficult to master. If you really want to see a mess open up a tune for a 2017 F150 Shelby Truck. It is SD only with 168 air tables and 56 spark tables that have to be reverse calculated based on inverse TQ tables from driver demand TQ and trans TQ management. It sucks.

Thanks. It is just tough to form a really representative profile that far in the kPa range with only a NA setup. It would be nice to convert my MAF curve into some sort of VVE profile haha. Just a matter of time before I fall behind even farther. I haven't been tinkering with any of the new ECMs, only the older PCMs. Even those I have taken quite a bit of time off. I'm trying to play catch-up.

Thanks again for all of your help!

[jsllc]

Make sure you log all of the PIDs in the list i gave you. Baro, IAT2 SC inlet will not be valid as these are on a e67 SC car. Rest should be.

[LaTechGTO]

Make sure you log all of the PIDs in the list i gave you. Baro, IAT2 SC inlet will not be valid as these are on a e67 SC car. Rest should be.

Alright, I finally got a chance to capture some datalogs yesterday and again this morning...

First off, I was an idiot and forgot to hit save after altering the O2 Readiness ECT threshold to force OL. I didn't notice it until after the second log. Some of the data can be used though.

Also, something else worth mentioning - I noticed this last time I went from a MAF only tune to failing the MAF and forgetting to change the same table to force OL. Bank 2 shows to be SUPER lean. The STFT will act right after the next restart after it's been driven for a while or if you flash it again. So, here we go:

V7.0 Normal.hpt - The tune I forgot to force OL
Log_1.hpl - Drive to the gas station with partial e85 and 93oct (e10)
Log_2.hpl - Had to fill up with 93 since the e85 pumps weren't working. Mixed driving

V7.1_VVE.hpt - Tune used to force OL

Log_3_OL.hpl - Mixed driving
Log_4_OL.hpl - Mixed driving (shorter)
Log_5_OL.hpl - Mixed driving (shorter)

Log_6_OL.hpl - Short drive this morning.
Log_7_OL.hpl - Return trip home.

Man, I am really digging some of the channels you setup. I can definitely see how some can be reaaaaaly handy to know. I also added my WB to the channel. ZL1-Fuel.Channels.xml

There have been some big weather/temperature swings lately. Really nice to see my %Err staying within a reasonable amount for the cells that are accurately represented with decent cell hits during steady throttle.

Thanks again!

[jsllc]

Here is a graph of your EQ error filtered for decel on Log_3_OL. Shows the error % to apply. Red add(lean) Green remove(rich).

Filter used : [50090.156.slope(3000)]<1 AND ([50119.234]<16) AND [6215.242]<135 AND [6215.242]>68 AND ([6310]=0 OR [6310]=1 OR [6310]=2 OR [6310]=4 OR [6310]=6 OR [6310]=7 OR [6310]=8 OR [6310]=9 OR [6310]=10 OR [6310]=11 OR [6310]=12 OR [6310]=14)

Math Formula used: (100*(([50119.238]*1)-[68.238])/([50119.238]*1))*.75
Column : 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 6400 6600 6800
Row: 15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 131 135 139 143 147 151 155 159 163 167 171 175 179 183 187 191 195 199 203 207 211 215 219 223 227 231 235 239 243 247 251 255
12-31-2017 11-47-24 AM.png

[LaTechGTO]

Here is a graph of your EQ error filtered for decel on Log_3_OL. Shows the error % to apply. Red add(lean) Green remove(rich).

Filter used : [50090.156.slope(3000)]<1 AND ([50119.234]<16) AND [6215.242]<135 AND [6215.242]>68 AND ([6310]=0 OR [6310]=1 OR [6310]=2 OR [6310]=4 OR [6310]=6 OR [6310]=7 OR [6310]=8 OR [6310]=9 OR [6310]=10 OR [6310]=11 OR [6310]=12 OR [6310]=14)

Math Formula used: (100*(([50119.238]*1)-[68.238])/([50119.238]*1))*.75
Column : 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 6400 6600 6800
Row: 15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 131 135 139 143 147 151 155 159 163 167 171 175 179 183 187 191 195 199 203 207 211 215 219 223 227 231 235 239 243 247 251 255
12-31-2017 11-47-24 AM.png

Thank you for taking the time to look at the logs and for the feedback!

I edited my table to match the format you displayed (kPa, RPMs, and decimal count) and attached an image of the data I would have normally used to apply (multiply by 1/2 %) to my VVE table, using the basic filters I compiled a while back. Obviously omitting the 600 RPM column. I recreated the image you provided by applying the filters and math function... after I bumped the cell hits to 25. I will have to do some thinking about your math function as I just use the built-in math for EQ Err%.

My filters are much more archaic than the ones you have listed haha - VERY clever filtering methods you have there. Thanks for sharing! I will have to do some digging on Fuel Cells 3, 5, and 13 to see why they are omitted; I originally thought they were the 3 in the higher RPMs, but the graphs I found online don't reflect that idea. I really need to do more research on the "slope" filters as well. I have an idea built in my mind of what it does, but I will confirm with more research. I hope you don't mind me borrowing those ideas and placing my own twist on them

The last two items...

1) A part of me is becoming leery of making any changes with small swings between logs. I can't seem to nail down what is contributing to the changes (<4%), but my perfectionist-type nature won't just give it a rest. At what %Err do most "professional" calibrators stop?

2) I still wonder what causes the huge STFT swing I have experienced a few times when forgetting to completely fail CL and only failing the MAF.

Happy New Years!

VVE Err.png

[jsllc]

The last two items...

1) A part of me is becoming leery of making any changes with small swings between logs. I can't seem to nail down what is contributing to the changes (<4%), but my perfectionist-type nature won't just give it a rest. At what %Err do most "professional" calibrators stop?

2) I still wonder what causes the huge STFT swing I have experienced a few times when forgetting to completely fail CL and only failing the MAF.

So when you say +-4% are you talking about VVE?? if so remember that the display is for you but really there are no tables. Just 30 areas of formulas for a given range. If you think about them they are nothing more than 30 areas of trends. Trend lines always lop off the highs and lows and create a general curve. This why VVE is used for backup. It is a close enough is good enough technology. it si like the mini-spare for a car. Designed to get you there when needed. Your goal should be to make it close as possible with no lean spots. It is not like SD and tables where you can control each cell. It will never take your corrections and calc an exact replica.

FTrim are different by ECM and OS version, every car type is different. Look in your logs to see which ones your car uses when. https://www.ls2.com/forums/showthrea...-by-MaP-vs-RPM Example for LS1. Here is HPT for LS2 I believe https://www.hptuners.com/help/vcm_ed...dvanced_ve.htm You just need to filter on the ones you want.

On the last part do not over think it. There is 75% of the ECM we do not get to see so works if done right, bad if not.

[jsllc]

I edited my table to match the format you displayed (kPa, RPMs, and decimal count) and attached an image of the data I would have normally used to apply (multiply by 1/2 %) to my VVE table, using the basic filters I compiled a while back. Obviously omitting the 600 RPM column. I recreated the image you provided by applying the filters and math function... after I bumped the cell hits to 25. I will have to do some thinking about your math function as I just use the built-in math for EQ Err%.

My formula is used for many different WBs so I have to adjust the error based on the WB hence the *1 which could change to *1.03 or *.97 depending on the WB that the car has. How I test is using stoic. Your NBs should deliver a stoic value if there are no transitions (idle). Your WB may say that average is .988 lambda. So I may say the WB has a 1/.988 error or *1.0121. If it says 1.015 lambda I would use 1/1.015 or *0.985. Keep in mind when you are logging the more math the harder it is on the computer and most systems do multiplication faster than division. Sometimes I ignore the error if I believe the WB. The last part I have found the using the just the Error% can overshoot so I only use 75% of the error.

Hope this helps.

[LaTechGTO]

So when you say +-4% are you talking about VVE?? if so remember that the display is for you but really here are no tables. Just 30 areas of formulas for a given range. If you think about them they are nothing more than 30 areas of trends. Trend lines always lop off the highs and lows and create a general curve. This why VVE is used for backup. It is a close enough is good enough technology. it si like the mini-spare for a car. Designed to get you there when needed. Your goal should be to make it close as possible with no lean spots. It is not like SD and tables where you can control each cell. It will never take your corrections and calc an exact replica.

Yeah, I am talking about VVE in this situation. I recall the previous discussion, but I was wondering when "enough is enough". When tuning MAF, most people have a particular threshold where they call it good (if I have a representative curve and have a pretty narrow window of %Err swings), and I was wondering where someone might consider the VVE acceptable. Like I said previously, I am sure everyone has their own personal threshold. I can't seem to get enough repeatable data to form a 1bar formula where it also appears representative on a 2 bar scaling. Maybe that is the threshold?

FTrim are different by ECM and OS version, every car type is different. Look in your logs to see which ones your car uses when. https://www.ls2.com/forums/showthrea...-by-MaP-vs-RPM Example for LS1. Here is HPT for LS2 I believe https://www.hptuners.com/help/vcm_ed...dvanced_ve.htm You just need to filter on the ones you want.

The e38 FTC tables I found didn't correlate to the portions of the table I expected when I saw your filtering. Maybe I didn't find accurate tables...? I will search again once I get some free time.

On the last part do not over think it. There is 75% of the ECM we do not get to see so works if done right, bad if not.

Yeah, I can definitely let that go haha. Curiosity gets the best of me sometimes.

My formula is used for many different WBs so I have to adjust the error based on the WB hence the *1 which could change to *1.03 or *.97 depending on the WB that the car has. How I test is using stoic. Your NBs should deliver a stoic value if there are no transitions (idle). Your WB may say that average is .988 lambda. So I may say the WB has a 1/.988 error or *1.0121. If it says 1.015 lambda I would use 1/1.015 or *0.985. Keep in mind when you are logging the more math the harder it is on the computer and most systems do multiplication faster than division. Sometimes I ignore the error if I believe the WB. The last part I have found the using the just the Error% can overshoot so I only use 75% of the error.

Hope this helps.

Ahhhh that makes complete sense. That is an excellent methodology to perform a "correction test". Thanks for being willing to share all of this information!