Printable View
As the title states, I'm setting up a graph to log Injector Flow Rate Density Multiplier just to do some testing and learning. What value should I be using to log values? Row and axis values are Injector Tip Temp and Fuel Rail Pressure but just want to start off looking at the correct data. Thanks in advance.
Gen V L86 Yukon Denali
I'm in the same situation as you and was thinking that EQ error ratio of commanded lambda vs. actual (if using a wideband) might be a good place to start. Would imagine that you would need to inverse positive/negative values if you were looking to use the histogram data to "paste special - % half" into the table... Certainly not going to profess this is right way, as I'm sure much smarter folks have found the correct way. This is just where I would take an initial SWAG.
I'm pretty sure from the toons I've seen that 99.9some percent of the people out there don't bother with this table - some will change it all to 1's and really screw things up or copy out of other like injector OE cals, but that's about it. I'm probably the only one that puts the time in to adjust it as it does require A LOT of other things to already be dialed in and a particular procedure to mostly be followed - not to mention DAYS of idle-ish logging. Using percent half or whatever will really screw you up - you could use percent and then inverse that and then come up with a math to convert that to a (.000) decimal point - that would probably work best, but still not perfect as some of the temp ranges won't respond to the same mods ;) Then you have to lock in rail pressures, rpms and do cold starts to hot idles after dialing the MAF in doing the same procedure ;)
Good luck - I gave out a lot of info there, but not really I guess - just glad to see some people taking the initiative to do things right.
Appreciate the replies. @GHuggins - definitely helps and good perspective. I'm not really trying to solve for anything just learning so figured I'd tap the well of knowledge.
I've been mulling this over for a few days now (in between the day job). I appreciate all the input and knowledge you provide @GHuggins. Totally understand that MAF would need to be really dialed in as an absolute must for this. For the sake of this post, let's assume that it's within 0.10% accuracy. Is the need to lock in rail pressure and RPM intended to keep as many factors as constant as possible during the process? Rail pressure, and am I correct that steady RPM is really intended to keep injector pulse width as uniform as possible during each cycle? Still seeking that 'ah-ha' moment of why this has to be done so meticulously vs. collecting logs while idling and driving from cold starts and during the warm-up of injector tip temps...
I was playing around with my maths, and confirmed that inverse of fueling error ratio applied percent was way to drastic of a change. Going all the way to the thousandths though seems a bit far to be effective. Maybe I'm going about this the wrong way. I'm currently plotting this math (while running in pure MAF mode, no DFCO, no VVE): 10 * ([50118.238] - [50119.238]) / [50118.238] on my histogram using averages and minimum cell hits of 50+ (currently no filters), and it looks like this may yield a decent value to apply via percent. Is this making sense and am I headed in the right direction?
Looking at this table across several 2016 platforms (5.3L Sierra, 6.2L Sierra, 6.2L Camaro, C7), the truck tables look hacked together vs. the performance cars. The Camaro and C7 table looks more to me like what I would expect to see for a density multiplier applied to flow rate for a given pressure and temperature, with the Camaro table looking the most 'right' to me. The lower temps and pressures seem to have a smaller density multiplier that doesn't entirely make sense to me when I think about the density properties of gasoline in relation to pressure and temp. The circled area just seems strange.. would think it'd be set to "1" to reflect density of gasoline at low temps, even with low pressures.
Attachment 122908
Maybe I'm overthinking it all.
This table isn't necessarily ever going to look smooth. Think about the DI injectors spray patterns/efficiency across a wide range of fuel pressure(0-3000psi). Even though the table is 3 dimensional, try thinking about the logic as the density multiplier is essentially the efficiency of delivering a specified volume of fuel given the rail pressure. The lower the value the less efficient the injector is and the longer it will need to stay open to deliver the fuel volume commanded for that event(event meaning revolution of the camshaft during the intake stroke). THEN, add the injector tip temp as a modifier the the two axis value logic I mentioned above and it will "slide" the efficiency around the 3D graph. Also, keep in mind that GM sometimes puts some bogus numbers in areas of the table that the car will never see. Is this intentional by GM? Is the programmer doing it to create mathematical blocks of data that fit a value they are looking for? (think of old school gaming on NES where the programmers had to do things like this to fit games into a specified data set size). Maybe it is just HP Tuners pulling data and extrapolating values into the table that are not actually populated at all, so rather than to have blank data they guess at it? I'm not saying any of this is correct, but rather just that you are overthinking some of it as well. lol
Anyhow, that is how I personally interpret that table and its data. It is one of the tables that can give you some granular fine tuning for when you swap out to larger aftermarket injectors. I don't believe there is a correct way to dial this table in as it depends on other parts of your tune. That would then determine what fine tuning you want to do here. Expect to spend more time with the car to adjust these values. Probably another reason not many tuners mess with this table, or just set to all 1's at WOT.
Thanks Jason! Good to get the feedback that I was overthinking it. I had a feeling, but coming from Gen3 to GenV tuning has me questioning everything due to the added complexity and interwoven dependencies. Also I'd totally missed the concept that this table is used to accommodate for equipment inefficiencies, rather than being purely representative of fuel behavior. Good call-out and makes sense, total miss on my part.
I haven't thought about NES gaming and the programming gymnastics they did to fit games onto those cartridges in years! In today's world of big data, high speed everything, that's a lost art form for sure. I have every Atari 2600 game on a single SD card instead of a mountain of cartridges :)
These tables won't ever or shouldn't ever wind up flat - should be fairly smooth, but not perfect. AS Jason pointed out the rest of the tune and ENGINE BUILD will highly effect this all the way down to something as simple as a cam install.
This is one I'm working on right now - pretty close to being dialed in - cam, pd blower, heads, lt4 fuel system, headers - I don't personally like them going over 1 as much as this one is, but this one has a pretty bad problem with being pretty lean then super rich run to run - hoping this takes care of it as I'm going in behind several big shops on this one.
Attachment 122937
This is a common one that surprisingly with minor changes works in just about every cam only build
Attachment 122938
You can see how these two differ greatly, but once dialed in fueling stays the same no matter what, which as we all know greatly effects torque and related associated calcs ;)
I do it the way I do it to lock in "constants" - if you figure out how to dial this in driving around - which well I guess you could only log maf data during 50 or 60 degree celcius and dial the maf in only there, then log for corrections using errors from that somehow, but you have to have constants to go by either way.
Just another fyi to help with this - raise your idle rpms up to smooth out airflow and O2 switching then after dialing in the MAF more so at specific temps - (already hinted at this above) - dial your multiplier table back to correct fueling with rail pressures locked. Then personally I interpolate between rows because that already takes a lot of time.
Now is the way I do it perfect - heck no, but it works. I also hinted at the correction terms in my first post - again not perfect and certain builds require double this, but it gets it close (.003 to .005 usually will make a 2 to 3 % correction)
Thanks Greg, that really helps! I was thinking that it's important to pick a point and use that as the 'constant'. Not sure how much logging I can get for data in the 50-60C range, as it's scorching hot right now in DFW and my injector tip temps blow past that range pretty quickly, but my logging does have alot of good data centered around 70-80C, and that's the general temp range where I've gotten my MAF table dialed in. I was noticing some pretty wild swings between logging sessions originally when I wasn't isolating data to include only a certain tip temp range.
I may have to wait for the 'winter' here to fully dial in this table, but at least I think I have enough of a grasp on it to move forward.
I appreciate all the wisdom, guidance, and general willingness to help!
Just use the 70 - doesn't really matter which you pick as long as you have constant you're going by. It may or may not throw some things off, but we're already doing that by not having a "true" lab set MAF curve - i.e - we're already throwing things out of adjustment changing the air model or rather adjusting the air model for the mods and then we're readjusting tip temps to redial fueling back in true to that, so doesn't really matter what you're using for your constant as long as you have one. I'll go ahead and let you know though - you're going to need to find a way to cool the engine down as you'll only be able to adjust this twice a day if you're lucky in those temps. I usually take just over 2 weeks dialing this in remote tuning.
Now, I just need to figure out how to make one of these binary so I can find some missing idle tables. Got one really kicking my butt right now and changing OS's changes how it runs, so..... Plus there's obviously something that caps timing OS to OS that I would really like to find.
Speaking of missing idle tables. I swear there is a table missing that refers to idle learn airflow limits. I've had to tune a few big throttle body cars now by butchering the Virtual Torque and VVE at idle so that it would actually learn not to have negative timing. Then, I have to undo this a few drive cycles later and the car never has another issue unless the idle learn is lost. Otherwise the car will never learn to not have negative timing and will eventually store a code in the ECU that says idle learning event failed. Can you find this table too lol.Quote:
Originally Posted by GHuggins
I think we all need someone who can read binary :D Yes, I've got one absolutely kicking my butt right now - doesn't matter on virtual torque, idle integral, proportional or anything - it surges badly on startups. I can put the exact same tm in another vehicle and it'll idle super smooth with positive torque and timing. Then there are the cars that will only idle at +3 to +6 degrees of timing at idle no matter how you shift the torque model - there are definitely missing tables controlling these things.Quote:
Originally Posted by TriPinTaZ
I can read binary and hex but I've never tried to mess with anything for HPTuners. Maybe I should poke around. Which tool do you use?Quote:
Originally Posted by GHuggins
Well that's just it - I can't - at least not yet - work in progress still. That's why I say I need to find someone who can :D
What I have found through talking to several people and companies is that I can convert a hpt tuner file to bin with special software - use that to open in tuner pro - find what tables are defined there - modify and install back into hpt. Some people have even stated that you can reopen a bin file after background items are modified back into hpt without having to use predefined parameters - haven't gotten the full gist of this yet. I've used software like this in the past to get around tuner locks, so I know it should be capable. Just not sure how many different programs I'm going to have to use to accomplish what I'm after. Personally I think tunerpro and hpt should have worked out a better deal when user defined was introduced to be able to open full files in their respective programs then figured out a way to not have to worry about license money being lost.