can i get a better rundown of ltft

[ashford]

it is not very descriptive in the tooltips, other post i have found. essentially i want a rundown of the logic of how it works and some of the options, some things are obvious some aren't

1. flowpath. so far it is just my speculation on how it works, o2 correction gets added to a ram table per correction softness( numbers mean what?) at this point i don't know if this table is incorperated into the ve, treated like an o2 correction initial table or if anything is done with this table. then after time has been met or %age met it is writen to flash( ve table or is ther a seperate ltft table, is corrections applied to or overwritten). if there is a ltft table in flash what is done with it is it incorperated into fueling or is it just something to look at.

2 buttons need better defining, read table to ram( if it is in flash shouldn't it be in ram as well). burn table to flash- burn what table to what flash, apply ltft(ram or flash) to ve, or update ltft ram to ltft flash

[jsmcortina]

1. flowpath. so far it is just my speculation on how it works, o2 correction gets added to a ram table per correction softness( numbers mean what?) at this point i don't know if this table is incorperated into the ve, treated like an o2 correction initial table or if anything is done with this table.

The LTFT correction is multiplied in like other corrections.

then after time has been met or %age met it is writen to flash( ve table or is ther a seperate ltft table, is corrections applied to or overwritten). if there is a ltft table in flash what is done with it is it incorperated into fueling or is it just something to look at.

The LTFT correction table stands alone. No changes are ever made to the actual VE table - that's what AMC did and we consider it to be dangerous - a defective O2 sensor could wipe out your VE table.

There are two copies of the LTFT table in flash - so that if the power is removed during an autonomous "burn" the firmware will have a backup table to use.

2 buttons need better defining, read table to ram( if it is in flash shouldn't it be in ram as well).

RAM and flash are different things. There's no way to apply a tool-tip to the buttons at present or I would have done.

burn table to flash- burn what table to what flash, apply ltft(ram or flash) to ve, or update ltft ram to ltft flash

Update LTFT RAM table to flash table (i.e. make it permanent.)

I hope that helps.

James

[ashford]

about 80% there, i think it might be easier if i wrote a rough draft and errors are pointed out, simply put i don't know what i don't know. if i find time in the next few days ill give a go at writing a rough draft.

[nathaninwa]

I'll follow along with this thread. LTFT is something I have wanted to explore.

[ashford]

overview

ltft is a table derived from o2 correction over time and incorporated into the fueling, essentialy it is a learned table in % of o2 correction without changing the ve table in case of an 02 sensor failure.(limits are unclear is it linked to o2 authority or can it keep going, also if it is linked o2 authority do the 2 stack up for example if o2 authority is 5% and ltft reaches 5% and o2 correction reaches 5% does this yield 10% total?)

there are 2 flash tables one is a backup in case of a power off while burning in which case it will use the good one(not sure how it detects if it is bad or if it copys from good to bad during boot up etc). this table in percent is added to fueling, basically overlayed over the ve table( all times or only when o2 correction is active)

there is also a ram table very similar to how the ve table is handled, if you are tuning ve it will go to ram and be used but is not permanent till it is "burned" this is what is updated continuously by sampling o2 correction, note that in order for it to be updated o2 correction has to be occuring. it is then burned to flash by time interval(if total change is met) or user burn(ts or connected switch) .

## big question here is the stored ltft automatically loaded into ram and used as it is updated(like tuning ve) or only used what is in flash then updated when burned.

settings

Enable experimental long temf trim- turns it on or off

Sample interval- how often o2 correction is sampled( not sure if it is that simple, average of a second, or if in a bin for a second then sampled)

Correction softness- how easily the talbe is affected( not sure what the numbers mean, i assume similar to the smoothing factors)

Write method- how learned ram values are burned
timed- time in minutes till burned( not sure if burned at every interval, but suspect that both time and total change must be met before burned, also if interval passes and total change not exceeded does it burn as soon as total change is exceeded or wait anotherr interval)
button- uses an input that the driver can press to burn to flash

burn interval- time between automatic burns

Led output- used to select a port to connect a light to to inform driver when total change threashold has been reached

total change before burning- the sum of all corrections to be met before light is illuminated or automatic burn allowed

display current % change in sensor 15- a means to observe ltft change realtime in ts( not sure if this is current value or sum of all ltft from flash.)

read table 1 or 2 to ram- still not clear on this by my logic flash shoud be loaded to ram at startup or is it just a means to see the table)

burn table to flash- burn ltft ram table to flash table

zero tables- wipes ram and flash


a few questions
is the table that comes up in ts the ram or the flash?
is there a means to apply ltft to the ve table?

[jsmcortina]

## big question here is the stored ltft automatically loaded into ram and used as it is updated(like tuning ve) or only used what is in flash then updated when burned.

Loaded on boot, then used.

is the table that comes up in ts the ram or the flash?

RAM

is there a means to apply ltft to the ve table?

Not currently, that would be a TS feature.

James

[jsmcortina]

PS.
Did you read through all of the tooltips and the Help menu text?

James

[ashford]

PS.
Did you read through all of the tooltips and the Help menu text?

James

read tooltips multiple times, i didnt think to check the help menu since i thought it was the same as the tool tips. that explains the buttons.

[ap67et10]

I was attempting to play with LTFT today, but everytime I tried to open the table, it would shut down CAN coms to GPIO. I have MS3 running 2v 4.6L mod motor and GPIO connected through CAN controlling the 4r75w Ford transmission in my 64 F100 truck.

I was having issues with the speed sensor to GPIO originally that was causing TS to give the prompt that it "could not communicate with CAN 1 controller and it was going offline to keep MS3 communicating". I got that issue all sorted out and there are no problems with that anymore, but that exact prompt displayed and GPIO went offline every time I tried to open up the LTFT table. I could open the settings menu with no issue, but if I tried to open the table it would shut down. I can't imagine this is suppose to happen, but I don't know anything about it (which is why I was going to play with it a little).


Andrew

[jsmcortina]

For your CAN passthrough issues, please start a new topic.

Post MSQs, and a copy of your TunerStudioAppDebug.txt file after your have enabled "Comms Log Debug" and then replicated the problem. You will likely need to zip it.

James

[ap67et10]

no problem, I will try to do that later today.

Andrew

[nismoautoxr]

question about "Total change before burning (%)"

Is this the percentage of cells on the table that saw a change or the percentage of change in any of all cells on the table that saw a change ?

I read ashfords "correct me if Im wrong " summary above and his take is that its the sum of all corrections rather than a %age of cells changed . I didnt see him get corrected but Id like a yay or nay on that . If that is true then that is why I have seen no burns . Mine is set at the default 100.

[nismoautoxr]

I see now why I got no reply ....the tool tip clearly says the sum of all changes DUH!!! Sorry.

Is sensor15 a normally logged item when datalogging? I cant seem to find it in MLV. I got it on tunerstudio as a gauge .

[nismoautoxr]

I get the feeling that my posts are for the most part being ignored LOL. Maybe its paranoia.

I notice that the lower numerically I set the correction softness ...the more aggressive the correction gets . This is opposite of the averaging in lag factors . I also have observed that the corrections as seen in sensor 15 do not make changes at steady state ...regardless of what EGO correction is at the time . The only time the corrections go active are going from redlight to redlight in traffic where there is alot of variations in actual afrs. On a typical drive from work or to work I might see .2% with 99.9% of my drive being in 6th gear at a steady RPM and throttle . I have to go through 5 city blocks worth of gear changes and from that point to my destination ...sensor 15 goes from .2 to 5.5% . Im not sure how to take that from an analysis standpoint because the release of information on how this works has been so sparse that its almost as if its being kept a secret ...Im just trying to contribute but my questions are not being replied to and Im not sure why .

BTW whenever I open the LTFT table1 ...9 times out of 10 ....tuner studio goes offline for about 15 seconds then comes back. Might correlate to the can comms issue mentioned earlier by AP67et10

[jsmcortina]

It isn't a secret, just complicated to explain. I hadn't noticed this post either.

In this case, sensor15 isn't logged by default. You can enable it by editing your "mainController.ini" file:
Change:
entry = sensor15, { stringValue(sensor15Alias) }, float, "%.1f", { sensor15_source }
To:
entry = sensor15, { stringValue(sensor15Alias) }, float, "%.1f";, { sensor15_source }

James

[nismoautoxr]

will do ...and thanks! For the last 6 months or so I dont mess with the car much at all besides driving it to and fro every day and track days . It has finally become fairly well sorted .

[Chevelle]

The LTFT correction table stands alone. No changes are ever made to the actual VE table - that's what AMC did and we consider it to be dangerous - a defective O2 sensor could wipe out your VE table.

This is fascinating to me and seems like a pretty neat feature. My question would be how is the LTFT table less "dangerous" than burning to the VE table in the event of an 02 failure? I would think that the LTFT table could lean the mixture out enough to be just as dangerous under load as the VE table alone getting re-flashed. If the danger lays in the fact that you could be left stranded from the VE table being wiped out wouldn't you still have that problem with the LTFT table? Would there be a way to "reset" the LTFT table in the event of an o2 failure without a laptop? Like on an OEM ECU you can disconnect the battery to reset it to the factory table. With an o2 failure using AMC you would have to use your laptop to re-flash the VE table to how it was the last time you had your laptop hooked up then turn off 02 correction, wouldn't you still have to do that same thing with the LTFT table?

[jsmcortina]

The difference is that the LTT table has a maximum +/- correction. When AMC changes the VE table, there's no point of reference so it can keep changing and changing and changing.

James

[Chevelle]

Obviously this is an old thread and I've done a bunch of searching and read multiple MS3 manuals, help tips, etc, etc and have not yet found the answers to these questions so I apologize if this has been discussed previously. This seemed like the most relevant thread to post these questions.

What does correction softness mean exactly? It's been mentioned previously in this thread and elsewhere that the LTFT table is safer than the old AMC because the LTFT table has a +/- % limit. I haven't yet found a way to change this % limit in the LTFT table. If I set correction softness for say 20 does that mean the table can change by 95% whereas a setting of 30 will only allow 85%? Or is the there already a set maximum limit of like 20% which is what 10 would be so then a setting of 50 would allow a total percentage change of about 10% and a setting of 100 would be closer to 1%? Or am I totally way off of what this "correction softness" means altogether - and if so what does it actually mean?

Assuming this question isn't answered with the answer for the above question... Is there a fixed +/- correction limit for the ltft table? If so, what is the limit?

Can the LTFT table be set to change only at a higher temperature than EGO correction is set for?
Example: EGO Control is set to correct for 02 above 120 degrees and LTFT table is changed only after temperature is above 180 degrees.

It was mentioned earlier in this thread that the updated LTFT table is only used after the controller is rebooted. Can the controller be configured to read the LTFT table in real time as it is updated in RAM or is it only possible for the updated LTFT table to be used after the key is cycled?