05 Centri 3v Log Review & WOT Surging

[ZOOMEEP]

Hi all,

I put an a/a intercooled centri on my 05 GT last year. I've gone to a few dyno shops, but it's tough to find someone who tunes 3V's with HPT.

I'm new to tuning, this mustang being the first car I've tuning. I have read Don Lasota's "THE ULTIMATE FORD HP TUNERS EFI TUNING GUIDE", which was great. I'm looking for some help & suggestions, to go along with my reading and trial and error. Any recommendation on additional PID's to log, histograms set ups, tuning workflow etc. would be hugely appreciated!

Set up info
Vortec V2si
A/A intercooled
Tial Q50 BOV, vented to atmosphere
Blow through MAF, HPX5000
47lb FRPP injectors
400LPH Fuel pump with wiring upgrade
-8an fuel lines, billet fuel rails
CMCP delete plates
AEM wideband, logged via analog input #1
AEM Boost Gauge, logged via analog input #2
MPVI2 w/ pro feature set

Tune Info
-5* timing (for safety while I adjust the MAF curve)
Adaptive WOT - Disabled
STFT - Enabled
LTFT - Enabled
Base Fuel Map .8λ @ WOT
DFCO - disabled

I posted a WOT log, a driving log, and my current tune file. Do they help answer these questions?

1) I use the STFT and EQ err histograms to edit my MAF curve. The MAF voltage rows in the histogram is set up to match the MAF voltage rows in VCM Editor. I copy the EQ err %, expand the lowest recorded value all the way down, and highest recorded value all the way up, then use the "multiply by % - half" command in VCM editor. I've found the EQ err and STFT do not match. STFT seems more accurate at lower loads and MAF voltages, and EQ err seems more accurate during WOT pulls at high load. Should I use STFT to adjust the lower part of the MAF curve and EQ err for the high part?

2) Using the above histograms and workflow, multiplying by % - half after each log, I expected to EQ error and STFT to close in closer to 0 with each log. I've done probably 40-50 pulls, and it doesn't seem to be getting any better. If you look at "R23 First WOT", my EQ err is almost all under 1%, with a high of 2.5%. "R23 2nd WOT" was logged after using log #1's data and multiplying by % - half, but my EQ err is way off, 4-16%. This has been common, I have decent runs and horrible runs, with no pattern I can see. Is there something I am doing wrong here? Should I be applying my correction factors differently?

3) During WOT pulls the car is surging. I noticed in "R23 First WOT" the MAF lb/min data log looked off. Rather than smoothly increasing with revs, it oscillates. I think the surging is due to the "squiggly" MAF reading. For "R23 2nd WOT" I logged TPS voltage and cam angles, thinking airflow over the MAF might be the culprit of the variable reading. TPS voltage was consistent, but cam angle seemed to vary with the MAF reading. Would faulty VCS solenoids alter cam timing enough to effect the MAF reading and cause a surge? Or is the squiggly MAF reading causing the VCT to alter the cam angle? Might it be belt slip at the blower? Any other ideas what it could be?

Long winded, but trying to provide enough details. Any help would be hugely appreciated!

[Thatwhite5.0]

Disable your LTFT and adjust fuel using STFT. Use your WB to adjust fuel for WOT.'
What size is your MAF housing?

Decrease Open Loop Delay Max from 1 second to .1 second. Your WOT throttle surge could be your MAF Transfer or torque/inverse table. Fuel does need work, 3V aren't very picky when it comes to the torque tables. Look at the stock Indicated torque table and keep a similar increase in your table. Use the Inverse Calculator to calculate the inverse table correctly. The calculator can be found in Edit.


Also add Engine Brake Torque and ETC Torque Request to your channels.

[ZOOMEEP]

The MAF is in a 3" tube so I think the ID is around 2.8ish. It's after the intercooler, in a long straight section between an elbow up from the intercooler and the elbow into the manifold. BOV is before the intercooler.

Should I adjust using STFT between idle and 3500RPM? Whatever the MAF volts indicate int hat area. Then do WOT pulls from 3500rpm and above? So basically STFT from the highest reading all the way down, then WB from lowest reading all the way up? I had the MAF transfer smoother before I did WOT adjustments.

I'll make those tune edits and add the PID's and return in a bit. I have no idea what the torque and inverse torque tables do, going to have to read up on that. What do you mean by similar increase? If I adjusted the MAF by 10% do the same for torque tables?

[B E N]

You need to be logging your TPS.

Use LTFT+STFT math to setup your MAF v. You can just take the output number and multiply, pull the data points plot directly out of your tune.

Logging IPC and plotting it on RPM vs air load can help you figure out where your problems are.

Plot wideband on RPM vs air load. Use excel to compare what your getting vs what you have as desired EQ in your tune. Apply those changes. You are lean pretty well everywhere.

Here: 3v layout.Layout.xml

That should get you everything you need. Start with the goal of never seeing anything over 1.00 lambda on your wideband, unless your off throttle.

[ZOOMEEP]

New data logs, R24 cruise and WOT. I Disabled LTFT and changed the OL Delay Max to .1.

Also added TPS, Engine Brake Torque and ETC Torque Request to the PID list.

Ben, the WOT pulls are hovering around .8λ, which I thought was pretty safe for a blown car. Should I richen it? The cruise data logs are all closed loop cruising, part throttle and under 3500 rpm. I have 1λ specified in my base map for those conditions. I took those references from the Lasota book, is that leaner than you usually tune for? Thanks for the histograms, I'll review those tonight. One question what's IPC? I saw "instrument panel cluster" online but I don't think that'll help me here.

The HPX5000 also came with a excel spreadsheet to generate a maf curve, and that curve was very smooth before I startled to applying EQ err and STFT percentages. I'm debating going back to those base values and starting over.

[B E N]

IPC is a torque error calculation (referred to as wheel torque), what the engine "should" be making vs what the pcm calculates its making. It's more of an indicator than anything. I just use it as a quick way to point to an area where I might have an issue in the tune.

What I saw that made me think you need to look a little bit at your lambda trims is the .5-1.0 load range between 500 and 3k rpm. You are at or over 1 in the entire range. Have you scaled out your load past 1.0?

For Lambda target you aren't only concerned with wide open, that's critical for sure but you should be below 1 at all loads. What your target is has to be your decision.

It will help driveability if you blend your fuel table a little. Very little real world effect on economy for a nice margin of safety.

fuel trim.jpg


Primary tip in fueling is governed by LWFM on these cars.

[ZOOMEEP]

Thanks B E N.

I haven't had time to play with it after work but I plan to spend a few hours on it over memorial day weekend. That fuel table looks much smoother, so I'll apply that. I'll also get familiar with the torque tables. My driver demand torque table has some values that are negative, is that right? I can't request negative torque can I?

What do you mean by "scaled out my load past 1.0?". I usually scale my MAF curve all the way to 5V based on last recorded EQ err.

The LWFM table gave me a lot of heartache when I started this build. I was using a stock MAF with a MAFia and to small a BOV. It was causing turbulent flow over the MAF, crazy high readings, and a wicked bad stumble at throttle tip in. The tip is is way better after switching to the HPX and 50mm BOV, but I'm sure I'll be revisiting LWFM again.

[B E N]

Do a compare with your tune to this one: Click the compare button at the top left, open compare file (its not going to be what you need, its a PD supercharger tune) Look at the engine torque/inverse tables, and the Airflow=>speed density tab.ported whipple 3v.hpt Your max air charge is now greater than atmospheric, think of air load as bar of boost.

[RobCat030]

Changed missing parameter, in SCT called "MAP Diff from BP for WOT for TP Model" Description: "This is the delta from barometric pressure for WOT conditions for the throttle position model". I'm not sure this will solve everything but I would think this HAS to be changed.

Air flows from high pressure to low pressure. Throttle inlet pressure is barometric pressure, throttle outlet is a boosted MAP. You need to rationalize this somehow, ETC Vacuum becomes negative. ECM 12006 "Max Aircharge Load" is a correction to standard aircharge in order to balance VE calculation. In other words, modifying engine displacement. You're not getting more air because the size of the engine changed at atmospheric pressure, you're forcing more air in at a higher pressure. The standard aircharge at standard temperature and pressure remains the same.

mitch R23 WOT Tuning (1).hpt

[B E N]

That post is real confusing Rob. Are you looking for a map sensor? The GT's don't have one. These cars can be tuned just fine with the available parameters. Maybe if you describe the function a bit more we can translate it to HPT.

What are you talking about "modifying engine displacement"? You just bring air charge load up, has nothing to do with displacement, displacement is a linear air use gain and blowers don't work like that. The PCM can handle 1.99 air charge load(≈14.7psi boost ASL), after that you start scaling. There is no change in displacement, that would really screw you up with a blower, especially a centrifugal.

[RobCat030]

Sorry, let me try to clarify my point. You're talking about changing Max Aircharge Load, which affects the Speed Density, and load calculation. So we can agree that even without a MAP sensor, Speed Density is relevant, particularly to the throttle.

HP and SCT both state specifically that Max Aircharge Load is not a representation of the maximum VE/load the engine can see. It is a modifier to standard aircharge, or SARCHG. What is SARCHG? It comes directly from the maximum cylinder aircharge for standard temperature and pressure that you define in the "Engine Displacement" scalar. So, when you set this max load to 1.3, what you are actually allowing the PCM to do is multiply your 0.00155lbs of SARCHG by 130% in order to satisfy the Speed Density/load calculation when your blower inevitably forces a MAF reading that is way too high for that engine displacement when NA. Now, does it still work? Sounds like you've made it work, but I still think there's a better place to represent the change you're making when a blower is introduced.

To explain this a different way, the standard aircharge of 0.0155 lbs for a 3V is at atmospheric conditions (actually, it's at 29.95 and 60F). If you modify the standard aircharge, you're saying that the engine can actually flow more than that at those conditions, when in fact you're getting more air because of the change in pressure. The WOT calculation needs to know that MAP can be greater than BARO, otherwise you run the risk of clipping the load and fuel calculation at WOT, even when the Max Calc. MAP is raised. When in doubt, the PCM will assume that the sensor readings are incorrect, not the physics calculation.

A similar metaphor for this would be pulling in injectors that flow 20% more and throwing that error into the MAF curve instead of changing the injector data. Granted, that's not a great comparison, we do have to make some concessions when dealing with boost on an NA PCM.

I don't have a 3V in front of me to work with, but I can say with certainty that the approach to deal with boosted 3Vs is different between users of SCT and HP, and I'm confident that has to do with the available parameters.

[ZOOMEEP]

Thanks guys. I reviewed the PD blower tune, I see what you mean by scaling the load. My loads are scaled out to 1.6 now, and I adjusted the speed density tab from 1.3 up to 1.7.

It looks like my calculated torque table goes higher than my inverse torque. I rescaled the inverse torque table up to 530 ftlbs, it maxed out at 406 before.

Quick question, in VCMedit the inverse rows are in lbft but in the VCMscanner the rows are in air load. How do I use the scan to correct errors if the rows have different units?

I also initially had this tuned with a Diablo MAFia, which required scaling engine displacement, hi/lo injector flow, and breakpoint. That scaling made the car run awful. Transient fueling was off, and it coughed and sputtered under any light throttle load. I'd like to avoid adjusting those values again.

It's running much better with the HPX5000, I'm just new to tuning and figuring this all out. The WOT surge seems to be the last crux, it's just sorting out fueling then timing after that.

[ZOOMEEP]

Scaling the inverse torque table rows seemed to help, there's less WOT surge and the MAF signal is graphing a lot more consistent.

I added 2* of timing back in, and adjusted the MAF transfer with STFT and EQ err.

It seems to be running better, but the IPC torque error is way off, like hundreds of ftlbs off. I need to figure out how to fix that.

[RobCat030]

Posting the current revision of the tune would be helpful.

The IPC error can come from many places, but the typical suspects are the torque tables and the throttle body tables. "Good" operation would be that the IPC error always remains at zero or slightly negative. Majorly negative IPC error won't necessarily cause throttle issues but it would indicate an inconsistency in the tune (for example, your WOT log has a driver demand of 500+ torque while EBT is around 330. Is the engine not making the torque that it should or are you asking for around 200lbs too much torque? Only you would know. With regard to IPC errors in the "NA" parts of your cruise log, I normally say to keep as much stock data as you can in your torque tables, and just populate with new data when you scale into boost.

The main issue is the same as I stated earlier. You have a centrifugal blower that's changing your throttle inlet pressure and messing up the throttle flow calculation. When the throttle opens to its predicted angle, it gets more air load than desired, and more brake tq than desired (compare EBT and ETC Torque Request, for example), it will close the throttle to try and correct and add on IPC error. Typically with forced induction, the throttle will oscillate right at the point where boost begins to build.

The tune needs to eat this error somewhere, what Ben and I were disagreeing about is where is the best place for it to go. No matter where you put it, it's going to be dishonest from a physics standpoint because you cannot accurately explain what is happening to this PCM. Personally, my rules are 1) Tell as minimal lies as possible to the PCM and 2) Stick to putting your error in as few places as possible. The three primary approaches I've seen are raising Driver Demand in problem areas, making torque inverse less than the torque table, or modifying the throttle body tables to thinking that the throttle has a larger effective area. You also have the option of limiting the Driver Demand so that the car only generates boost past the WOT setpoints. If you post your current tune, I can explain the way that I would go about it, but there are many ways to skin this cat, and most important is you find a way that works for you. Hope this helps

[ZOOMEEP]

Thanks for the explanation Robcat. I attached the current tune revision here.

Another weird thing I'm noticing is the PCM is commanding 1λ under every throttle position and RPM except WOT. Even after blending the fuel table, EQ log and Fuel Base table don't match.

Also if DFCO is off will it mess up my fuel trims at lower MAF voltages? The middle and top section of my MAF transfer are close, the fuel trims and EQ err and closer to 0. However, lowest voltages are reading rich. I'm wondering if that's because DFCO isn't cutting fuel under deceleration. Do you usually turn off DFCO for tuning?

[ZOOMEEP]

Another quick note, the oscillation/surge feeling does feel like opening and closing the throttle, however the TPS sensor voltages are consistent. If the PCM was closing the throttle to correct IPC errors would he TPS voltage show that?

[ZOOMEEP]

Would this method work? I understand that torque is calculated, but air load is read from sensors correct?

This looks like you log air load, calculate torque, smooth the table, then take the inverse to populate the inverse table.

https://forum.hptuners.com/showthrea...se-calculation

[RobCat030]

At the end of that thread you can see me alluding to what my thoughts are on the torque calculation. Ford used curve fitting, we should too. Here I used the stock data to extrapolate into boost. All these do is relate torque to load for the PCM. If you wanted to remove a translation layer, they could have made it so that the Driver Demand table is populated with air load values and then you wouldn't need a torque inverse table (I'm oversimplifying, but you get my point). You can request any amount of load between 0 and peak TQ at any accelerator position, that's the beauty of a DBW. I think those saying something about "matching what the engine is doing" are conflating accelerator position and throttle position (of course, the most predictable pedal feel is where air load stays relatively consistent throughout the RPM range).


My approach is to leave the torque be and focus on the throttle body data (ie. dont throw torque errors into both the throttle body and the torque table, it becomes hard to keep track of). Your airflow requests are converted into an effective throttle area which is used to navigate to the Predicted Throttle Angle table. Higher MAF requests with a "broken" (boosted) PR function will request effective areas that exceed the X axis of that table. You can rescale that axis so that all of your boosted throttle angles become associated with those higher effective areas. Set up a histogram to see where IPC errors generate within that table (you should be able to log both effective throttle area and ETC Vacuum), that will give you hints where to adjust.


The Effective Area table is just an inverse of that, the calculator is under the "Edit" tab.

Air load is a derivative of MAF, so it's still technically calculated. The only thing you truly trust as "sensor" data is your MAF volts/period and your wideband voltage, save for any mechanical issues. In general, the fuel base table only becomes active once you've passed the open loop threshold, defined by throttle position. I believe that's true for the 3V as well.

I still disagree about moving Max Aircharge Load, but to each their own.

The attached tune is just for review and to demonstrate my point, and as always, this is just my $0.02.


R26 - Rob Cat.hpt

[B E N]

Still seeing lambda values over 1 in places on your engine speed vs TPS, when filtered to only show positive acceleration. This will cause a surge. Try increasing your MAF failure load by 20% across the board and see how it responds (make sure and write this tune iteratively so you can revert if the results aren't what you want).
WOT surge can also be a mechanical issue, what are your spark plug make/model and gap?

Rob; To answer your question about why there are differences between SCT and HPT, there are differences between 2 tuners on HPT, there are multiple ways to address a problem in the software. My car makes ≈500hp on a stock throttle body, and I haven't ever felt a need to touch throttle body data. Tip in fuel is largely governed by maf failure load, manifold volume, and torque management tables.

My view is change the minimum possible from ford, time after time I see people struggle with throttle body tuning so for my own sake I just don't. You can get one of these to run really well on boost with only about a dozen parameter changes (plus injector data). It doesn't have to be complicated, and if you are new to tuning keeping it simple makes a lot of sense. There is nothing more frustrating than having your brand new supercharger car drive like shit.

[RobCat030]

There is nothing more frustrating than having your brand new supercharger car drive like shit.

We can certainly agree there. Do what you have to do to make the car run right.

To answer your question about why there are differences between SCT and HPT, there are differences between 2 tuners on HPT, there are multiple ways to address a problem in the software.

Yes, sure. Ideally the software of choice has all of the relevant parameters to the issue you're dealing with, so that you can address the problem in the way you see fit as a tuner, and also so that you can see the bigger picture. Don't get me wrong, I'm not here to talk smack about HP or SCT. I'm not a software engineer, but I can't imagine it's easy. But there's a difference between choosing between one of multiple solutions and not even being aware that other solutions exist because they're not in your software. That's frustrating, especially when you're first starting out and aren't really aware of it.

Also, I don't mean to write novels on here, so I do apologize for that. I just wanted to offer an explanation surrounding Max Aircharge Load.