?? about CLI and PowerVision

[07heri]

This should be an easy one for some of you guys but I really don't know the answer.  Tuning (actually just helping) an 08 FXSTC with the PV.  Most of the CLI's are below 5% but a few are like 107 108 and a few down below 94 after numerous runs.  What would happen if we just lowered or raised those cells by the % amount the CLI's are showing they're adding or subtracting?  He plans to run open loop is why I'm asking.  If the CLI's are showing an addition or subtraction isn't that saying that the VE's are off by that amount?  Or does this interpolation factor have more to do with it.  We was bs'ing over a beer and this came up and I really didn't have a concrete answer.  My feelings are, in closed loop to just leave the VE's alone and ride it.  But, if someone was going to run an open loop tune would adjusting VE's based on the CLI % be a way to get things closer without burning up time and gas?     

[joe_lyons]

You would prob be fine just running it or you can change the VE in those areas by that amount and be fine also.  Your close enough that it just depends on how lazy you are.  You can run a basic AT and accept its changes also.

[strokerjlk]

not sure if I understand the queston 

tune it in closed loop 14.6 in the afr table.
autotune
"export learned" until you dont see it changing ve's much anymore.
using those ve's
get map .....to your computer
take the afr table to whatever afr you want to run in open loop.
reload that map (send map) in to the vision then flash it to the ECM.

[07heri]

not sure if I understand the queston 

tune it in closed loop 14.6 in the afr table.
autotune
"export learned" until you dont see it changing ve's much anymore.
using those ve's
get map .....to your computer
take the afr table to whatever afr you want to run in open loop.
reload that map (send map) in to the vision then flash it to the ECM.

using the log tuner software versus just letting the unit make the change.  I like seeing whats happening in the cells and just let the software make the new map.  Using the software we can see exactly how many hits he got in every cell and if he's not working a certain area long enough.  Either way the CLI question is still the same.  When he does a datalog and looks at the numbers some of the CLI's are either high or low.

Maybe i didn't ask the question the right way.  I know we won't get it 100% tuning on the street and in closed loop it will compensate for the small variances.  After the last new map was made we did a datalog run.  A few of the CLI's are showing 107 108 93 94.  Are we understanding this correctly to mean the VE's are off by 7% 8% in the areas that the CLI's are showing high or low corrections?  Is a CLI of 107 telling us the ECM is adding 7% fuel based on what the VE is?  That's how we're interpreting that CLI reading.  Are we correct in that assumption? 

If that's wrong please explain if you would.  If were correct, couldn't we just add 7% to the corresponding cell to lower the correction the ECM is making?  Hope that makes more sense.   

[hdmanillac]

Hello,

It's normal to see such values for CLI when trims have just been reset for cal reflash. AFF (adaptative control) need around 20 minutes to start integrate, and when AFF are set, CLI are back to 100% +/- 3%.

[07heri]

Hello,

It's normal to see such values for CLI when trims have just been reset for cal reflash. AFF (adaptative control) need around 20 minutes to start integrate, and when AFF are set, CLI are back to 100% +/- 3%.

Wants to run open loop.  AFF are turned off.  Just asking if the VE's can be adjusted based on the CLI%, nothing more.  Scenario:  New flash, trims reset, AFF off, datalog, CLI = 108.  If the ECM is adding 8% can we add 8% to the VE's.  If he wanted to run closed loop I'd run it like it is.  Seeing as how he wants to run open loop I'm just curious if using the CLI% is a way to manually adjust the VE's so it's as close as possible in open loop.

[hdmanillac]

Ok, I understand. I've tried this few month ago. Not exactly the way you suggest. I worked with AFF that are readable with PV. My aim was to get all AFF around 100% +/-3%. But finally I could not obtain a good result.

Maybe your suggestion is better. You should try !

[joe_lyons]

Yes if you see a 108 then you can add 8% to that spot and if you see a 95 you can take 5% away from that spot.

[07heri]

Ok, I understand. I've tried this few month ago. Not exactly the way you suggest. I worked with AFF that are readable with PV. My aim was to get all AFF around 100% +/-3%. But finally I could not obtain a good result.

Maybe your suggestion is better. You should try !

The reason I was looking the CLI's was because I thought those readings meant the integrators were adding or subtracting to get the AFR to 14.6.  I just wasn't, and still not, sure how the surrounding cells played into how the ECM came up with that CLI number.  I may try 1/2 the CLI number and then datalog it and see if it messes up the cells surrounding the area I'm trying to correct.

[ToBeFrank]

It's really best to let the PV's software algorithm handle this. You have a couple problems doing it by hand:

1) A sample is never from just one cell
2) The CLI you're getting is probably "off" from the area it's logged at. This is due to the way the J1850 bikes do their logging.

1) is rectified in the PV's tuning algorithm. 2) is rectified by using many samples to tune a cell, which, again, would be done by the tuning algorithm. Picking out one or two samples here and there and making corrections won't work nearly as well.

[delta]

Last year when I first started tuning with the PV, I had similar questions. This was my post and the reply from Jamie on another forum.

Q: "I've been monitoring my AFF gauges after loading a new tune. I have been observing swings from 98 to 102 in the closed loop areas, which with my limited knowledge I think is pretty good. That's the equivalent of + or - 2% right? Now onto my question: when I get a reading of 98, is that telling me that the VE of that particular cell is a tad to large; and that, 98% of the fuel required by that cell is actually being delivered?

I'm wondering if I can do a datalog showing the AFF and maybe CLI, could I then use this info to better smooth the VEs of adjacent cells? Or, should I be happy with what I have."


A: "If the AFF and CLI is within a couple % your tune is right where it should be. A value of .98 tells us the ECM has pulled 2% fuel to match the closed loop target, the closed loop integrator (CLI) is a short term/real time fuel trim and the adaptive fuel factory (AFF or AFV) is a long term stored trim, however varies by resolution and block learn area. While it appears you are well within reasonable tune, if you want to work with the tune a bit you can do some logs and use the Log Tuner software to apply these normal corrections to your VE tables directly."

Thought this might be useful to your discussions.

[07heri]

It's really best to let the PV's software algorithm handle this. You have a couple problems doing it by hand:

1) A sample is never from just one cell
2) The CLI you're getting is probably "off" from the area it's logged at. This is due to the way the J1850 bikes do their logging.

1) is rectified in the PV's tuning algorithm. 2) is rectified by using many samples to tune a cell, which, again, would be done by the tuning algorithm. Picking out one or two samples here and there and making corrections won't work nearly as well.

Yes, that was pretty much my concern...how a manual change would affect the surrounding cells.  Also was thinking that just because the CLI indicated high at one cell location wasn't relevant to just THAT cell if the surrounding cells are used to compute the AFR. 

I'm gonna go out this afternoon and just adjust out half ie;  4% in the area thats saying 108 and see what happens.  If I do that and then datalog again and it didn't throw off the surrounding areas I would assume all is good. 

It's just 3 or 4 areas that won't get to within 5% on the CLI's.  It's probably not that far off to really worry much about as it sits.  But if a slight manual tweak gets ALL the CLI's under a 5% swing I'm assuming it will be better.

[07heri]

Seems to have done the trick.  Right or wrong, the CLI numbers came down and didn't seem to affect any surrounding areas.  I'm now wondering if this might be a way to fine tune the final VE's when there's just 3 or 4 areas that have high or low CLI numbers showing on a datalog.  I'm interested to hear from the folks that know the inner workings of these tuning products.  Granted, it may have turned out different if the % changes were alot but in this case we're talking only 3 or 4%.   

[ToBeFrank]

Seems to have done the trick.  Right or wrong, the CLI numbers came down and didn't seem to affect any surrounding areas.  I'm now wondering if this might be a way to fine tune the final VE's when there's just 3 or 4 areas that have high or low CLI numbers showing on a datalog.  I'm interested to hear from the folks that know the inner workings of these tuning products.  Granted, it may have turned out different if the % changes were alot but in this case we're talking only 3 or 4%.

Interesting. The PV software should have got it, then. Maybe it was a transient area or not enough samples? No way to know without looking at their algorithm.

[hdmanillac]

Using CLI mean you can only adjust VE in CL.

Using AFF works also in OL. But AFF table does not correspond to VE table.

[07heri]

Using CLI mean you can only adjust VE in CL.

Using AFF works also in OL. But AFF table does not correspond to VE table.

Shouldn't have anything to with either or.  After 4 runs had 3 or 4 areas that the cli's were 107 108 109.  The rest was all 1 or 2%.  So In the areas that were, for example 108, I added 4% to the VE's.  Made 2 data log runs and now those areas are at 103 and 104.  It was just an experiment but all cli's are within 5% now.   Best assumption on my part is the VE's were off a tad and the ECM was making up for it.  Added a little to the VE's and CLI's are lower. 

[joe_lyons]

You would prob be fine just running it or you can change the VE in those areas by that amount and be fine also.  Your close enough that it just depends on how lazy you are.  You can run a basic AT and accept its changes also.

Holy monkeys I was right.
But now if you take the square root of the clb and then divide that by pie and then square it by string theory you will have something so confusing that you will just give up.  Or just do it the simple way that you did and get the results that you need.