Good day all,
I've been doing some very cautious tweaking of the fuel dynamics tables in my tune lately, to good effect. I've sort of been shooting in the dark, without really understanding what the tables are actually doing. I was hoping someone here has a good understanding and can enlighten me. I have several questions about the tables, and the general strategy here:
1. B3406 Impact factor -- this has been the table I've mostly been playing with. I understand that this is a measure of the fuel that stays on the walls of the intake and doesnt enter the combustion chamber, on a per temperature and MAP basis. I have heard that the PCM takes the start point and the end point of an airflow change, and its the difference between these two points on this table that determines the extra fuel injected or held back. Is that right? Is it to simplistic of an explanation?
2. B3401 evaporation time -- I'm a little confused about this one, is it "this is a measure of how long the excess fuel in the intake takes to enter the combustion chamber"? if so, lower numbers in this table should mean that the PCM expects the excess fuel to enter the combustion chamber faster? So if this table is too low at one point, then it will result in a lean spike, and then a rich tail off, because the pcm pulls too much fuel at the beginning of the event, and too little fuel at the end? Conversely, if the table is too high at one point, the pcm will pull too little fuel, and then too much, causing a rich spike and then a lean tail?
3. B3426 initial decay multiplier for throttle opening compensation fuel mass -- I've heard this called "stomp compensation", but I honestly havent the foggiest what it does. It also says in my XDF that it is only used at engine startup, which makes little sense to me, because people also talk about using only this table to tune out fuel dynamics problems.
4. B3428 Gain multiplier for previous throttle opening fuel mass -- I know even less about this table. It appears to be a rate at which B3426 is ramped out? I cant see any indication about how this is applied, what sort of time frame its applied over, etc.
5. I've done a lot of logging and tuning the VE table and the MAF table, and I'm fairly confident that they are at least pretty close, and I dont generally get the "so lean its off the charts" spike when getting on the throttle, but I usually get a big rich spike when I snap the throttle shut, that slowly tapers off. I've heard that there are tables in newer PCMs that deal with the asymmetry of the fuel dynamics, how does 12212156 deal with the "excess" fuel being negative? Is my VE table still just too rich in those areas, or is it dynamics?
Most of my information comes from here: https://forum.efilive.com/showthread.ph ... e-throttle, which appears to have some conflicting information, for example post 3 contradicts my thoughts on B3401.
12212156 Fuel dynamics questions
Re: 12212156 Fuel dynamics questions
The rule is
fuel evaporates rate is based on temperature of fuel, temperature of intake valve, and vacuum.
The higher temperature and higher vacuum[lower MAP] the faster the fuel evaporates.
The tip-in lean spike is associated with MAF signal filtering[ there should be some scalars based on different conditions].
The rich spike on throttle shut can be leaned out with DFCO tables.
EOIT target is also very important factor on modified engine.
First tune these out, and as a last resort start playing with fuel dynamics tables. Most of them use very complex math models and without heavy modded engine[intake,heads, valves and so on] it is better not to touch them.
To figure how they work you will need some RAM logging of the output scalars, and how they change the PW based on different conditions.
fuel evaporates rate is based on temperature of fuel, temperature of intake valve, and vacuum.
The higher temperature and higher vacuum[lower MAP] the faster the fuel evaporates.
The tip-in lean spike is associated with MAF signal filtering[ there should be some scalars based on different conditions].
The rich spike on throttle shut can be leaned out with DFCO tables.
EOIT target is also very important factor on modified engine.
First tune these out, and as a last resort start playing with fuel dynamics tables. Most of them use very complex math models and without heavy modded engine[intake,heads, valves and so on] it is better not to touch them.
To figure how they work you will need some RAM logging of the output scalars, and how they change the PW based on different conditions.
Re: 12212156 Fuel dynamics questions
I should probably have listed more information about my setup. It is not just a stock engine. I've converted an l31 Vortec 5.7 to LS1 26lb injectors and this intake: https://www.holley.com/products/engine/ ... ts/300-263. I made an adapter plate to adapt the stock throttle body and EGR to that intake, because I'm cheap and I live in Canada. I'm reconsidering keeping the EGR now, but I probably still will since I went through all the trouble.
The MAF setup has not changed appreciably. It is still in the same place and using the same pre-MAF tubing and post-MAF tubing. The only thing that has changed before the intake manifold is the elbow into the throttle body. I 3d printed a more gently curving one because the stock one would not have fit. From my testing it has not changed the MAF curve very much if at all. That said, I did do the 411 swap after the intake swap so I don't actually know if the correct settings are there.
I originally did all of this because I got so frustrated with the stock injection setup and manifold. The injectors (even the mpfi ones) are trash. I had a stock poppet injector setup go bad so I replaced it, and then immediately started having problems with the new setup. I took everything back apart again and replaced the sticking mpfi injector and then after I put it all back together again it started leaking coolant, so I took it all back apart and did the lower intake gaskets. I noticed a lot of pitting on the lower intake where it faces the coolant passages and on the gasket surfaces around that, so I cleaned it and tried to get all of the rotten metal out and jb welded it nice and flat. It was basically perfect looking, jbweld was nice and secure, all was well. Put it all back together with new gaskets and everything nice, and then 3 days later big coolant leak. Rip it all back down and all of the jbweld was gross and bubbled. I must not have gotten it clean enough or something. At this point the stock coolant quick connect broke as well and I just said I wasn't going to deal with this garbage setup anymore. I even tried to find a non rotten stock lower and it was just not happening.
I dont really want more power out of this, the stock engine was grunty enough for me. I want it to not be constantly requiring me to tear it apart. So far the new manifold has not leaked at all. I'm actually really happy with that part of the setup. I'm fairly sure that since the new manifold has quite a lot less air volume and runner length, that I've lost some lowend, which is fine, it doesnt really feel much different to me, but for the same reason, I'm reasonably sure that the dynamics have changed.
So to respond to your statements:
1. Where do I find the MAF signal filtering settings? How do I know if and how they should change?
2. I got this PCM from Lextech with a basemap on it, and the DFCO has a big delay before it kicks in (like multiple seconds at low RPM). Should it? I would think that getting the rest of the settings right should fix the rich problem, without needing to resort to something that seems like a bandaid. I could be wrong though and GM could have intended it to work that way.
3. I haven't changed the cam, so I don't know how much the EOIT would change, assuming I got the injector settings right. How would you change it, based on my above description? I assumed that it should be the same as the vortec 5.7 stock, because (I assumed) its more about when the valve is open than the particular injector.
4. I actually really would like to log these things and figure the math out, do you know where these RAM parameters are? Is there a listing of them that I haven't found?
The MAF setup has not changed appreciably. It is still in the same place and using the same pre-MAF tubing and post-MAF tubing. The only thing that has changed before the intake manifold is the elbow into the throttle body. I 3d printed a more gently curving one because the stock one would not have fit. From my testing it has not changed the MAF curve very much if at all. That said, I did do the 411 swap after the intake swap so I don't actually know if the correct settings are there.
I originally did all of this because I got so frustrated with the stock injection setup and manifold. The injectors (even the mpfi ones) are trash. I had a stock poppet injector setup go bad so I replaced it, and then immediately started having problems with the new setup. I took everything back apart again and replaced the sticking mpfi injector and then after I put it all back together again it started leaking coolant, so I took it all back apart and did the lower intake gaskets. I noticed a lot of pitting on the lower intake where it faces the coolant passages and on the gasket surfaces around that, so I cleaned it and tried to get all of the rotten metal out and jb welded it nice and flat. It was basically perfect looking, jbweld was nice and secure, all was well. Put it all back together with new gaskets and everything nice, and then 3 days later big coolant leak. Rip it all back down and all of the jbweld was gross and bubbled. I must not have gotten it clean enough or something. At this point the stock coolant quick connect broke as well and I just said I wasn't going to deal with this garbage setup anymore. I even tried to find a non rotten stock lower and it was just not happening.
I dont really want more power out of this, the stock engine was grunty enough for me. I want it to not be constantly requiring me to tear it apart. So far the new manifold has not leaked at all. I'm actually really happy with that part of the setup. I'm fairly sure that since the new manifold has quite a lot less air volume and runner length, that I've lost some lowend, which is fine, it doesnt really feel much different to me, but for the same reason, I'm reasonably sure that the dynamics have changed.
So to respond to your statements:
1. Where do I find the MAF signal filtering settings? How do I know if and how they should change?
2. I got this PCM from Lextech with a basemap on it, and the DFCO has a big delay before it kicks in (like multiple seconds at low RPM). Should it? I would think that getting the rest of the settings right should fix the rich problem, without needing to resort to something that seems like a bandaid. I could be wrong though and GM could have intended it to work that way.
3. I haven't changed the cam, so I don't know how much the EOIT would change, assuming I got the injector settings right. How would you change it, based on my above description? I assumed that it should be the same as the vortec 5.7 stock, because (I assumed) its more about when the valve is open than the particular injector.
4. I actually really would like to log these things and figure the math out, do you know where these RAM parameters are? Is there a listing of them that I haven't found?
Re: 12212156 Fuel dynamics questions
That cleans most of the unknowns.
As a start we need to know fuel pressure, fuel system type[returnless or vacuum referenced]
Get accurate injector data from another bin file that use them, in case they are OEM ones.
For fuel dynamics tables, there is 2006 mexican truck l31 engine that runs ls1 pcm. We can port some of the dynamics from there, and fix them if needed.
As a start we need to know fuel pressure, fuel system type[returnless or vacuum referenced]
Get accurate injector data from another bin file that use them, in case they are OEM ones.
For fuel dynamics tables, there is 2006 mexican truck l31 engine that runs ls1 pcm. We can port some of the dynamics from there, and fix them if needed.
Re: 12212156 Fuel dynamics questions
Fuel pressure is 4 bar (4 bar bosch FPR in a Radium FPR housing, I have a pressure gauge permanently installed into the radium housing). I actually did have a dying fuel pump as well (who doesn't, in these trucks?) but that's sorted as well now. It is technically a returnless system, but the FPR is mounted in the engine bay. I intended to make it a return system but I couldnt get any fittings that were low profile enough to miss the distributor, because I used taller injectors, it basically perfectly lined up with the port on the fuel rail that I needed to use for a return. I just capped the return and ran the system that way, with the "return" from the FPR instead. If I could find a non-crab-type more traditional distributor cap I could convert it but I dont think it would make much difference. This setup is specifically called out in the Radium documentation as being a thing you can do.
Not Vac referenced, and the injector data is scaled correctly by MAP value instead of the same across the board. I had a hell of a time finding data for the 26lb injectors, see this post: viewtopic.php?p=119743#p119743 I can vac reference in theory, but I found it didn't make a difference to my setup as long as the tables were set up correctly. I'm not planning on adding boost or anything, and haven't noticed it getting leaner than it should at higher loads.
I don't know anything about the mexican truck engines, did they run the same intake as the l31 in my truck? If so, I would expect their dynamics to be the same as what I started with.
Now that I think about it, when I was doing VE logging I had the MAF failed and it would still do the lean spike on throttle and then the rich spike off throttle. I would think that any MAF signal filtering would not apply there. Am I wrong?
Also I really appreciate the help with this and am willing to try things out to get it sorted, but I'm also quite interested in the theory behind it, and actually looking at the math. If you (or anyone else) have time, could you go into a little more detail on my previous questions? It might seem like I ask a lot of questions and some of them might be nonsensical, but I cant really learn anything without knowing that.
Not Vac referenced, and the injector data is scaled correctly by MAP value instead of the same across the board. I had a hell of a time finding data for the 26lb injectors, see this post: viewtopic.php?p=119743#p119743 I can vac reference in theory, but I found it didn't make a difference to my setup as long as the tables were set up correctly. I'm not planning on adding boost or anything, and haven't noticed it getting leaner than it should at higher loads.
I don't know anything about the mexican truck engines, did they run the same intake as the l31 in my truck? If so, I would expect their dynamics to be the same as what I started with.
Now that I think about it, when I was doing VE logging I had the MAF failed and it would still do the lean spike on throttle and then the rich spike off throttle. I would think that any MAF signal filtering would not apply there. Am I wrong?
Also I really appreciate the help with this and am willing to try things out to get it sorted, but I'm also quite interested in the theory behind it, and actually looking at the math. If you (or anyone else) have time, could you go into a little more detail on my previous questions? It might seem like I ask a lot of questions and some of them might be nonsensical, but I cant really learn anything without knowing that.
Re: 12212156 Fuel dynamics questions
the l31 uses this manifold 17113201 . Not quite the same as yours.
Did you copy the short pulse offset tables too.
So far I managed to find that dynamic fuel kicks in when there is engine steady flag set. In transients the fuel dynamics logic might be disabled.
The maf scalar might have evolved to dynamic air coefficients with different zones, much more complex than earlier versions.
Pcm always blends data from maf and VE and than uses the dynAIr coefficients. The basic principle is that it polish the abrupt changes in airflow to a much tamed curve. removing all kind of abrupt spikes on rapid airflow changes, On different conditions it adds or remove airflow faster or slower based on throttle delta changes, usually the tip-in and lean-out are associated with this. I have seen in RAM on older lt1 platform how this works. The unfiltered MAF airflow and the tamed filtered MAF airflow.
To make a better understanding how it works.
if airflow changes from 5 to 10 the pcm don`t go from 5 to 10 in airflow model, it goes from 5 ,6 ,7 ,8 ,9,10 in 10ms interval, or it could go from 5, 7.5, 10 in 5ms interval, based on some prediction how fast the airflow change will reach the cylinder.
I will try to dig out some l31 files for comparison.
Did you copy the short pulse offset tables too.
So far I managed to find that dynamic fuel kicks in when there is engine steady flag set. In transients the fuel dynamics logic might be disabled.
The maf scalar might have evolved to dynamic air coefficients with different zones, much more complex than earlier versions.
Pcm always blends data from maf and VE and than uses the dynAIr coefficients. The basic principle is that it polish the abrupt changes in airflow to a much tamed curve. removing all kind of abrupt spikes on rapid airflow changes, On different conditions it adds or remove airflow faster or slower based on throttle delta changes, usually the tip-in and lean-out are associated with this. I have seen in RAM on older lt1 platform how this works. The unfiltered MAF airflow and the tamed filtered MAF airflow.
To make a better understanding how it works.
if airflow changes from 5 to 10 the pcm don`t go from 5 to 10 in airflow model, it goes from 5 ,6 ,7 ,8 ,9,10 in 10ms interval, or it could go from 5, 7.5, 10 in 5ms interval, based on some prediction how fast the airflow change will reach the cylinder.
I will try to dig out some l31 files for comparison.
Re: 12212156 Fuel dynamics questions
I did copy all of the injector tables, including the small pulse tables. The small pulse tables for the poppet injectors are insane negative values.
I see, so maybe the MAF value doesnt follow the actual airflow as closely sometimes. It makes sense to me that it could be doing some smoothing on the data I suppose. The only thing I can see is called MAF Airmass Filtering Coefficient, which is just a scalar value of 10.2%.
I'll post my current bin file and xdf, maybe theres something obvious Im overlooking. Just need to get to my tuning computer.
I see, so maybe the MAF value doesnt follow the actual airflow as closely sometimes. It makes sense to me that it could be doing some smoothing on the data I suppose. The only thing I can see is called MAF Airmass Filtering Coefficient, which is just a scalar value of 10.2%.
I'll post my current bin file and xdf, maybe theres something obvious Im overlooking. Just need to get to my tuning computer.
Re: 12212156 Fuel dynamics questions
Here is the bin file and xdf I've been using.
- Attachments
-
- truck tune.zip
- (365.9 KiB) Downloaded 102 times
Re: 12212156 Fuel dynamics questions
hmm, looking further into this, I think I got the small pulse adjustment from another car. Probably one that runs the bigger ls1 injector (I bought these before I knew that there were two types and of course I end up with the more obscure ones). I cant find the small pulse adjustment table in the 99 camaro bin in UniversalPatcher. It looks like I may have used the small pulse threshold from the bigger injector too, as mine is 4.0274, and the 99 camaro one is 4.0677. Its a one percent difference though, and as far as I can tell it just limits where the PCM indexes into the small pulse table. That table is all 0's near the 4 msec mark anyway, so I don't think it really makes any difference.
Re: 12212156 Fuel dynamics questions
looks like the 12212156 "Min deliverable injector PW for fuel dynamics compensation" scalar _could_ be the same as the 9373372 "Min_Pulse_Width" scalar in universalpatcher? the "Min_Pulse_Width" table description says "This is the minimum pulsewidth the pcm will command for the fuelling portion of the final pulsewidth equation". If thats true, then mine is too low at 1.0 msec vs 1.4 in the camaro tune. A 40% difference in the minimum fuel it will try to deliver might explain the richness off throttle I'm seeing. I've also found a couple of other minimum pulse width discrepancies, so I'm going to play with that a bit.