Dodge Challenger Forum banner

1 - 6 of 6 Posts

·
Registered
Joined
·
131 Posts
Discussion Starter #1
I’ve been wanting to tune fueling since getting hpt and unlocked pcm, finally had time today to get some data and make some changes for part throttle fuel, from idle to 3k I’m now seeing Ltft at +- 2 and short +- 8-10 while logging and making changes I had dfco setpoints maxed to keep dfco out of the way and wot setting maxed as well to keep it in closed loop with rear O2 set at 376 to prevent rear O2 from mucking with data .

Anyone have a good way to log and adjust trims from 3k up in CLOSED loop and keep it out of WOT ?
 

·
Registered
Joined
·
2,221 Posts
I’ve been wanting to tune fueling since getting hpt and unlocked pcm, finally had time today to get some data and make some changes for part throttle fuel, from idle to 3k I’m now seeing Ltft at +- 2 and short +- 8-10 while logging and making changes I had dfco setpoints maxed to keep dfco out of the way and wot setting maxed as well to keep it in closed loop with rear O2 set at 376 to prevent rear O2 from mucking with data .

Anyone have a good way to log and adjust trims from 3k up in CLOSED loop and keep it out of WOT ?
Right off the bat I admit I'm not a tuner but LTFT readings varying between +/-2 (%?) and STFT readings varying between +/- 8 and +/- 10 doesn't seem all that right to me.

From observation over the years -- since '96 when OBD2 showed up in cars -- long term fuel trims should be pretty stable. Course they won't be with the short term fuel trims bouncing around. My info is that +/- 10% or close to it is CEL territory.

Not sure what you mean exactly by "rear O2 set at 376 to prevent rear O2 from mucking with data". The rear O2 sensor readings want to remain above 0.6V and higher which means there is a lack of oxygen in the exhaust gases. This signals the engine controller the excess oxygen it "added" to the exhaust -- via slightly leaning the mixture (and you can see this as the short term fuel trims change slightly for instance at idle about every second as the engine controller adds just a tiny bit of extra fuel then takes it away) -- is being consumed in the converter which is a good sign the converter is working properly. If by "376" you mean "0.376" volts this is low and can be interpreted by the engine controller as too much oxygen in the exhaust. This might explain why the short term fuel trims are swinging so widely. The engine controller is trying to get the desired reading from the 2nd set of O2 sensors.

The engine controller also wants to see the voltage levels from the 2nd set of O2 sensors vary slightly -- but staying "high" -- in reaction to, in "sync" with, the amount of oxygen that is being fed to the converters.

Have you considered having the car professionally tuned? Say by someone who has a 4 gas exhaust analyzer and can monitor air/fuel ratios and other telemetry in real time and adjust/tune this to be correct through all RPM ranges and including WOT where correct fueling is while critical at all other times super critical under WOT.
 

·
Registered
Joined
·
178 Posts
Right off the bat I admit I'm not a tuner but LTFT readings varying between +/-2 (%?) and STFT readings varying between +/- 8 and +/- 10 doesn't seem all that right to me.

From observation over the years -- since '96 when OBD2 showed up in cars -- long term fuel trims should be pretty stable. Course they won't be with the short term fuel trims bouncing around. My info is that +/- 10% or close to it is CEL territory.

Not sure what you mean exactly by "rear O2 set at 376 to prevent rear O2 from mucking with data". The rear O2 sensor readings want to remain above 0.6V and higher which means there is a lack of oxygen in the exhaust gases. This signals the engine controller the excess oxygen it "added" to the exhaust -- via slightly leaning the mixture (and you can see this as the short term fuel trims change slightly for instance at idle about every second as the engine controller adds just a tiny bit of extra fuel then takes it away) -- is being consumed in the converter which is a good sign the converter is working properly. If by "376" you mean "0.376" volts this is low and can be interpreted by the engine controller as too much oxygen in the exhaust. This might explain why the short term fuel trims are swinging so widely. The engine controller is trying to get the desired reading from the 2nd set of O2 sensors.

The engine controller also wants to see the voltage levels from the 2nd set of O2 sensors vary slightly -- but staying "high" -- in reaction to, in "sync" with, the amount of oxygen that is being fed to the converters.

Have you considered having the car professionally tuned? Say by someone who has a 4 gas exhaust analyzer and can monitor air/fuel ratios and other telemetry in real time and adjust/tune this to be correct through all RPM ranges and including WOT where correct fueling is while critical at all other times super critical under WOT.
the cell is thrown when STFT exceed -25 or +25%, in the meantime, they can bounce up and down, nothing wrong with what he has.
The rear 02's are supposed to simply monitor that the fronts are doing their job. we disabled the rear for example in GM stuff, so the o2 test doesn't run, causing a lean/rich wave as the 02 test is running (this is how the rear 02's can test if front o2's are working), don't know if this is the same in our case, but i notice some resembleance in the OP's approach back when i did tune my GM vehicles :), the 376 is F, operating temp for more likely the rear o2's to be enabled/disabled.

OP
are you disabling NN to tune your VE ?
it seems in this platform we have use LTFT +STFT to do so at least for PT, you'll probably need a dyno to properly map VE past 3000rpm in part throttle.

Why are you tuning VE ? what mods do you have ?

I have mufflers and a K&N filter, and i haven't have a need to dial the VE, my fuel trims are -2/+2%, i wouldn't bother with those numbers, maybe if i had something bigger than 5%, but in this platform, the neural network stuff does a great job compensating for minor stuff.

Changes to your VE tables with NN enabled won't do anything for fueling, as the NN seems to be ruling everything. you can still tune fueling with NN enabled, but you have to make changes using injector offset. or you can disabled NN, and tune the VE tables,some say it's more complicated this way, i haven't tried eirther way yet as i have no need for it at the moment.

What method are you using ?

I'm also learning from the pros at dodge tuning in HPT, this stuff is way different than what i know from GM.

In any case, take a look at this LTFT's at steday throttle on the fwy, i haven't touched my VE at all.

ltft_70mph by Wrangler RSM, on Flickr
 

·
Registered
Joined
·
131 Posts
Discussion Starter #4
Thanks for the comments !
First let me say being new to tuning I am being very cautious and painstakingly conservative in what i am doing to learn, I will be going to a local professional tuner for dyno tuning as soon as my insane schedule allows it!

What i am doing now is learning as much as i can by reading researching and then with experiments, I am still running factory tune and my methodology is to log data, make a new tune file, with adjustments then duplicate my data log ( i have a 20 mile route i found locally that is low traffic and can have a repeatable run for comparing like data under the same circumstances)

mods are few right now with catch can, CAI, and exhaust.

Now the fun part !!

I want to UNDERSTAND how our PCM's work both NN on and NN off what is common to both and what is not so in regards to tuning NN on I have logged stft/ltft and adjusted InjPw tables and saw a dramatic reduction in trims , more stable O2 readings, and less knock retard. saved that tune file and put stock tune back in.

Now to the VE tables ,, there is some info on the VE tables in HPT forums and some here as well, but i cannot find anything definitive on they are implemented by the PCM, i've read they are largely ignored with NN on, and only "active" when NN is off, and i've read they are the basis for EVERYTHING as well.
So in my quest for understanding I set up a VE histogram and watched what the VE really was and compared it to the stock tune file, logged VE was significantly different from the tables in the stock tune.

I wanted to see if modifying the VE table actually impacted NN operation and fuel trims so i disabled NN, logged trims ( same route i established as mentioned above) made adjustments to VE table via the multiply by %-half function smoothed VE table saved, and wrote to PCM. as soon as the write was completed i ran the exact same route / log and saw a reduction / stabilization of fuel trims, I next enabled NN and with no other changes ran the same route and logged the same data, expecting to see trims return to what i captured before tweaking the VE table, however this is not what I saw in the data, the log with VE ON had virtually identical trims ( within 1%) of the log i ran with NN off, and both had significantly less fluctuations than the stock tune file log.

this is why I am very confused at how these operate and am running the stock factory tune at the moment .

I will writing my tune file with adjusted InjPw tables and running NN on until i can get to the dyno and pro tuner but really want to understand how this VE table really works
 

·
Registered
Joined
·
2,192 Posts
Since you are close to stock, leave NN on. If you install a cam, etc... then turn NN off and tune the VE tables.

You need to be logging STLT vs Fuel Mass to dial in the injector pulse width. I attempt to maintain around +/- 1.

Once you get the trims dialed in, then you can adjust PE (if you have access to a wideband).

Set Closed Loop Coolant Temp to 140* and set Closed Loop Enable ECT rear to 140* as well.

You want the rear o2's to participate in fueling. Even with my build I have both of my rears turned on. I simply disable the rear o2 codes.

I'm still learning with NN off, but I've gotten pretty good at tuning with NN on.
 

Attachments

·
Registered
Joined
·
131 Posts
Discussion Starter #6
a new observation regarding ANN tuning and VE tables,, after logging STFT/LTFT to InjPw and making changes to get correct fueling in ANN, i logged several logs ranging from quick 2k_6K runs, long (30-45 min) "normal driving " , logs under heavy load ie low rpm up hill etc... and had consistent LTFT's of 0 with a very few 1's, STFT's hanging zero most of the time with max values of +-3. Conducting an experiment I logged VE bank 1 and 2 and then applied the changes found in the logs to the actual VE tables in my tune. changed NOTHING else, ran several logs as mentioned above and saw massive changes in fueling ( STFT/LTFT) vs InjPw, STFT's went through the roof, LTFTs went up to as much as 5 in spots, tweaked my InjPW tvs fuel mass table and its inverse and then ran the runs again, with the changed VE tables, and bam trims back to near zero, both short long and the STFT/LTFT math histogram, alternatively if i leave my InjPw table alone now and put the VE tables back to stock values the trims are cutting massive amounts of fuel.

Any thoughts on this as supposedly the VE table has no effect on ANN and from what i have heard

The Neural Network takes in 4 inputs, Intake Cam, Exhaust Cam, RPM, and PRatio, and spits out Bank 1 and Bank 2 VE using a neural network, which isn't as readily obvious as to how it works. as a table. Its kind of like a 5 dimensional curve fit, so its not easy to manipulate.
 
1 - 6 of 6 Posts
Top