Mitsubishi Lancer Evolution Forum banner

21 - 40 of 56 Posts

·
Banned
Joined
·
4,212 Posts
I'd never rely on rear O2 logging - it always drifts with temperature. Serial logging is the way to go :thumbup:
 

·
Registered
Joined
·
68 Posts
Discussion Starter #23 (Edited)
Alright guys... I'm starting to go nuts here. I've spent countless hours now trying to get my MAP sensor dialed in, and am struggling. It seems I drive around logging part-throttle and full throttle pulls, and keep making adjustments to the MAP Calc tables, and end up making no progress. Sure, it might change for the better at 3,000RPMs 6.8psi, but by 3,300RPMs 7.2 psi, my MAPCalcs and MAFCalcs are way off from each other. What is an acceptable gap between the two?

Should I just start over with Cobb's Stage 2 or 3 MAP sensor calc tables, and just adjust VE? ...What am I doing wrong here, or why is this being so difficult?! Never had this much trouble on my 3D SD setup IX.

To clarify this, after I calibrated my MAF, I re-enabled it and have been logging MAF Calc and MAP Calc and seeing the difference in percentage at any given RPM vs. Boost and making adjustments on the 3x MAP Calc tables. I noticed, if I disable my MAF and run strictly on the MAP Sensor, the tune is way off! Should I just run on MAP only and log RPM vs Load and Target AFR vs Actual AFR and just adjust the VE table instead of the 3x MAP Calc tables?
 

·
Registered
Joined
·
1,416 Posts
What intake do you have? Does it have a shield around the filter? I know I may be having issues with my AGP CAI


Sent from AutoGuide.com Free App
 

·
Registered
Joined
·
2,413 Posts
bumping this for me later.
 

·
[Administrator]
Joined
·
3,246 Posts
No it's actually the best way, but for the sake of time and efficiency many of us use a large amount of narrowband O2 sensor feedback data to make the same adjustments.

Just make sure it's resulting in a smooth MAF curve. If it's no longer smooth and parabolic, you're probably going to start tuning in circles, so to speak. Often folks bake in adjustments to the MAF curve that ought be applied to both MAF and MAP calculations. This is why tools like Tuner Helper are so nice as they take into account the fact that the MAF curve should remain just that - a curve.
This reminds me... I should add a function to Tuner Helper to use AFRMAP and AFR to calibrate the MAF and MAP. Won't work on Cobb logs (unless they've included AFRMAP) but still.
 

·
Registered
Joined
·
2,413 Posts
Here's a basic flow of my fuel model calibration process for a stock camshaft car:

1. Use well known or "close enough" fuel injector and MAF calibrations. This includes low pulsewidth adders, fuel injector scaling, injector lag time/latency, and MAF scaling.
2. Disable MAF calculations and rely on the factory MAP calibration to find out how far off your injector scaling is at idle and cruise. Adjust injector calibration accordingly.
3. Re-enable MAF calculation and disable MAP. Log several drive cycles worth of MAF Voltage, LTFT in use, and STFT and run it through Tuner Helper (thanks ripnet) to adjust your MAF curve. I run through this twice. Make sure to start with a sane MAF calibration that is either mathematically derived or taken from another car with the same MAF tube.
4. Disable MAF calculation and enable MAP calculation only. Use ripnet's tool to adjust idle and cruise the same way as for MAF calcs.
5. Set up your high octane AFR table to your desired targets. Use a conservative timing map and re-enable MAF. Start making partial and full throttle pulls and log MAF/MAP calcs, AFR target, and wideband AFR. Use this data to bring MAP calcs in the high load range closer to MAF calcs (always trust MAF calculation over everything else by default so long as you have no boost leaks).
6. After converging MAF and MAP calcs, start to use the delta between target and actual AFR to adjust the fuel calibration table ("VE" table in COBB) all around. The goal is to get MAF and MAP as accurate and tightly grouped as possible and then use the fuel calibration table to take care of any further inaccuracies in your fueling model. It affects all injection so it is not a VE table.


Then start tuning everything else ;)
This reminds me... I should add a function to Tuner Helper to use AFRMAP and AFR to calibrate the MAF and MAP. Won't work on Cobb logs (unless they've included AFRMAP) but still.
Is that the same step you use Rip?

I want to start doing this, or at least understand it.

Is this the correct summary?

  1. Plug in the rough numbers for injector & MAF
  2. Increase MAF to max value to disable it and have MAP only. Then follow Clipse's thread to get fuel scalings closer.
  3. Bring the MAF values back to where they were and max the values on the MAP tables to run MAF only. Log the car (assuming best drive would be city, and a cycle is 4-6min?). Then use those logs to run through Tuner Helper.
  4. Not really sure what this means.
  5. Since we're trying to get AFRMAP == AFR set your targets in the map. Use the data to adjust MAP to closely align (or match) MAF.
  6. Fix the difference between AFRMAP & AFR with fuel calibration table.
 

·
[Administrator]
Joined
·
3,246 Posts
Is that the same step you use Rip?

I want to start doing this, or at least understand it.

Is this the correct summary?

  1. Plug in the rough numbers for injector & MAF
  2. Increase MAF to max value to disable it and have MAP only. Then follow Clipse's thread to get fuel scalings closer.
  3. Bring the MAF values back to where they were and max the values on the MAP tables to run MAF only. Log the car (assuming best drive would be city, and a cycle is 4-6min?). Then use those logs to run through Tuner Helper.
  4. Not really sure what this means.
  5. Since we're trying to get AFRMAP == AFR set your targets in the map. Use the data to adjust MAP to closely align (or match) MAF.
  6. Fix the difference between AFRMAP & AFR with fuel calibration table.
So there's two different (maybe more) ways to accomplish this. Open or Closed loop. Currently, Tuner Helper is setup for Closed Loop, meaning it works on the STFT+LTFT fuel trim. Any values in the HOFM (AFRMAP) are ignored and irrelevant for this process.

In a nutshell: TH looks at your logs and averages the fuel trim (STFT + LTFT) for each MAF voltage (or RPM/pressure cell for MAP tuning) and uses that average to determine the new number for that cell. For example, if at 2.6volts MAF, your combined fuel trims are averaging -6.25%, then it will multiply your current table value at 2.6volts by 0.9375. There's a few other calculations that aren't important, just know that they're there... like smoothing, running averages, zero IDC trimming, etc .

For this to work, the system you are working on (MAF/MAP) needs to be lower than the other. In the case of MAP, simply set your MAF table to it's highest value. For MAF, you'll need to set the MAP tables higher, but if you max them the car probably wont start. Multiply by 1.08 or something, just enough to get MAPcalc higher than MAFcalc. Alternatively, Cobb has the ability to shut off one system with a few checkboxes.

What I will add to TH soon, is Open Loop. This will ignore STFT+LTFT and use your wideband recorded value, compared against AFRMAP. TH will calculate the percentage difference, and adjust your map accordingly.
 

·
Registered
Joined
·
86 Posts
Here's a basic flow of my fuel model calibration process for a stock camshaft car:

Have been reading through this. Awesome information. This part I'm wondering though, what is done differently with cars with cams? I'm currently in the process of dialing in injectors and everything, and my car has MX1 cams, mild cams with a good lope to them at idle. So what do you guys do differently? Or are these not really big enough to matter?

Thanks
 

·
Registered
Joined
·
2,413 Posts
So there's two different (maybe more) ways to accomplish this. Open or Closed loop. Currently, Tuner Helper is setup for Closed Loop, meaning it works on the STFT+LTFT fuel trim. Any values in the HOFM (AFRMAP) are ignored and irrelevant for this process.

In a nutshell: TH looks at your logs and averages the fuel trim (STFT + LTFT) for each MAF voltage (or RPM/pressure cell for MAP tuning) and uses that average to determine the new number for that cell. For example, if at 2.6volts MAF, your combined fuel trims are averaging -6.25%, then it will multiply your current table value at 2.6volts by 0.9375. There's a few other calculations that aren't important, just know that they're there... like smoothing, running averages, zero IDC trimming, etc .

For this to work, the system you are working on (MAF/MAP) needs to be lower than the other. In the case of MAP, simply set your MAF table to it's highest value. For MAF, you'll need to set the MAP tables higher, but if you max them the car probably wont start. Multiply by 1.08 or something, just enough to get MAPcalc higher than MAFcalc. Alternatively, Cobb has the ability to shut off one system with a few checkboxes.

What I will add to TH soon, is Open Loop. This will ignore STFT+LTFT and use your wideband recorded value, compared against AFRMAP. TH will calculate the percentage difference, and adjust your map accordingly.

If you do this part, how will accel enrichment/decel enleanment get filtered out? Could we log those tables and see if they being 'used' and use that to filter it out?
 

·
Registered
Joined
·
2,798 Posts
What I will add to TH soon, is Open Loop. This will ignore STFT+LTFT and use your wideband recorded value, compared against AFRMAP. TH will calculate the percentage difference, and adjust your map accordingly.
How do people do this now? Use the evo scan average tracer and just make adjustments by hand to the MAP tables to achieve desired AFR?
 

·
Registered
Joined
·
122 Posts
On my end I created a tool to do it but I sent a PM to ripnet maybe there's a way to merge a few thing or at least use the idea already done. That would save him time, coding isn't my job and I'm not good with user interface and error handling so what I did would never be usable by other people ;)

Let's see maybe with his touch it can happen!
 

·
[Administrator]
Joined
·
3,246 Posts
How do people do this now? Use the evo scan average tracer and just make adjustments by hand to the MAP tables to achieve desired AFR?
Don't make sudden throttle movements. :) Otherwise, IDC=0 (or IPW=0) (decel) is filtered out.
 

·
[Administrator]
Joined
·
3,246 Posts
On my end I created a tool to do it but I sent a PM to ripnet maybe there's a way to merge a few thing or at least use the idea already done. That would save him time, coding isn't my job and I'm not good with user interface and error handling so what I did would never be usable by other people ;)

Let's see maybe with his touch it can happen!
Replied :D
 
21 - 40 of 56 Posts
Top